2008-11-01
[00:52:00.0000] <yecril71>
The example with musical footnotes is a wordplay.

[00:53:00.0000] <yecril71>
The word "footnote" meanins that the note should be put at the bottom of the text.

[00:53:01.0000] <yecril71>
It means that.

[01:07:00.0000] <yecril71>
It has a semantic meaning, 

[01:07:01.0000] <yecril71>
although its name clearly contains a presentational part.

[01:07:02.0000] <yecril71>
It is because footnotes are always placed after the text,

[01:08:00.0000] <yecril71>
otherwise they make no sense.

[01:08:01.0000] <yecril71>
A heading is another example of this shift from semantic to presentation.

[01:08:02.0000] <yecril71>
... shift from presentation to semantics, of course.

[01:09:00.0000] <yecril71>
However, the word "footnote" meaning a musical note that has its flag up is purely presentational.

[01:10:00.0000] <yecril71>
Setting the class of such a symbol to "footnote" is a misunderstanding.

[01:11:00.0000] <yecril71>
It should be CLASS="foot note" instead.

[01:11:01.0000] <yecril71>
So this is not a good counterexample at all.

[01:13:00.0000] <yecril71>
The solution presented in the specification is may be inconsistent but it is at least practical.

[01:14:00.0000] <yecril71>
A class is a semantic property, it is not exclusively for the sake of styling.

[01:15:00.0000] <yecril71>
The cases where class bears no semantics are exactly the cases where id can (and should) be used instead.

[01:16:00.0000] <yecril71>
Sort of "it really means nothing, but I would like this text to be yellow".

[01:16:01.0000] <yecril71>
Where "this text" is replaced by an identifier.

[01:17:00.0000] <yecril71>
On the other hand, class=xyzzy means that this text is an instance of class xyzzy.

[01:18:00.0000] <yecril71>
And this is semantic information.

[01:18:01.0000] <yecril71>
Or rather denotes an instance of class xyzzy.

[01:23:00.0000] <yecril71>
It seems irc-logs do not work.

[01:24:00.0000] <yecril71>
So I am just typing in vain.

[02:37:00.0000] <Philip`>
yecril71: http://krijnhoetmer.nl/irc-logs/whatwg/20081101 - the logs do seem to work :-)

[03:35:00.0000] <yecril71>
They work today but there is a hole of yesterday.

[03:35:01.0000] <yecril71>
So the whole think is sort of unreliable.

[03:35:02.0000] <yecril71>
The whole thing is sort of unreliable.

[03:36:00.0000] <yecril71>
It is funny how I tend to make typos here.  It does not happen in e-mail that often.

[04:09:00.0000] <Lachy>
yecril71, most IRC logging systems are somewhat unreliable because they depend upon the up time and connectivity of the server running the log bot

[04:14:00.0000] <Dashiva>
Several bots running on each server on the network, connecting from different locations.

[06:29:00.0000] <erlehmann>
If you don't celebrate Halloween, PROCEED AS NORMAL.

[08:56:00.0000] <yecril71>
The recommended way of making footnotes is inconsistent because

[08:56:01.0000] <yecril71>
there are three very different ways, depending on the context.

[08:56:02.0000] <Hixie>
that's not inconsistent

[08:58:00.0000] <yecril71>
Moreover, the usage boundaries are not sharp.

[08:58:01.0000] <yecril71>
So, when the text of your footnote changes, you have to change the footnote syntax.

[08:58:02.0000] <yecril71>
If that is not inconsistent, then what is?

[08:59:00.0000] <yecril71>
And please update the IRC link on Wiki.

[09:01:00.0000] <Hixie>
being inconsistent would be like having three footnotes use different styles in the same document

[09:01:01.0000] <Hixie>
the spec suggests different ways to do it, and allows authors to pick whichever one they want

[09:01:02.0000] <Hixie>
which link in the wiki?

[09:01:03.0000] <yecril71>
See your talk page

[09:02:00.0000] <yecril71>
Did you mean "styles" as in CSS?

[09:03:00.0000] <yecril71>
Using different styles in various places need not be inconsistent.

[09:03:01.0000] <yecril71>
Using different constructs for similar semantics is inconsistent.

[09:05:00.0000] <yecril71>
Exercise: Find all footnotes in the current document.

[09:05:01.0000] <yecril71>
The length of the solution is the measure of consistency.

[09:05:02.0000] <Hixie>
oh feel free to edit the page

[09:06:00.0000] <yecril71>
I feel but it is protected

[09:06:01.0000] <Hixie>
huh, why was it protected

[09:06:02.0000] <yecril71>
Use action=protect

[09:06:03.0000] <Hixie>
try now

[09:11:00.0000] <yecril71>
Done.

[09:13:00.0000] <Hixie>
thanks

[09:13:01.0000] <Hixie>
bbl

[10:23:00.0000] <yecril71>
User agents, such as web crawlers, do have interest in semantics.

[10:27:00.0000] <yecril71>
It would be nice to specify title and behaviour using CSS.

[10:27:01.0000] <yecril71>
For the time being, behaviour relies on JavaScript and

[10:28:00.0000] <yecril71>
it can be applied via JavaScript, although this is a bit inefficient.

[10:28:01.0000] <yecril71>
There is no workaround for title.

[10:29:00.0000] <yecril71>
(where no script is available, that is).

[10:29:01.0000] <yecril71>
However, managing hyperlinks in the same way would be insane.

[10:30:00.0000] <yecril71>
Images and hiding are subject to CSS.

[10:31:00.0000] <yecril71>
Ordinarily, there should be just one element called "edit".

[10:31:01.0000] <yecril71>
Or no such elements, of course.

[10:32:00.0000] <yecril71>
I can see nothing wrong in having to change the stylesheet.

[10:33:00.0000] <yecril71>
CSS is modular because of @import.

[10:33:01.0000] <yecril71>
The only problem with CSS is that you cannot explicitly inherit style

[10:33:02.0000] <yecril71>
from a structurally unrelated class.

[10:34:00.0000] <yecril71>
That problem clearly wants a solution but it is not a problem in HTML.

[10:34:01.0000] <yecril71>
Well, here you are, Bert.

[13:48:00.0000] <Hixie>
106

[14:48:00.0000] <takkaria>
107


2008-11-02
[17:38:00.0000] <utilisateur>
ok so..

[18:10:00.0000] <utilisateur>
na nannaaaaaaa

[05:56:00.0000] <gsnedders_mibbit>
/me waves

[05:56:01.0000] <gsnedders_mibbit>
my computer is broken :\

[05:57:00.0000] <Philip`>
Clearly you need more computers

[05:57:01.0000] <Philip`>
Redundancy is the solution to all problems

[05:57:02.0000] <gsnedders_mibbit>
clearly

[08:38:00.0000] <yecril71>
I think a footnote deserves a dedicated element.

[08:39:00.0000] <yecril71>
It belongs to the standard repertoire of typesetting and publishing.

[08:40:00.0000] <yecril71>
If we recommend using aside.footnote, why not div.article?

[08:41:00.0000] <yecril71>
A footnote is much more used than aside, and it has a different meaning.

[08:41:01.0000] <yecril71>
An aside contains an illustration for the main text, 

[08:42:00.0000] <yecril71>
a footnote is for in-depth explanation of details.

[09:28:00.0000] <mpt>
I think asides are more common on the Web than footnotes are

[09:28:01.0000] <mpt>
In books probably the reverse is true

[09:29:00.0000] <Philip`>
/me wonders if that's just a consequence of the commonly available layout models

[09:31:00.0000] <Philip`>
(Book pages tend not to have enough space for putting asides in sidebars, and web pages tend not to have nearby feet in which to put footnotes)

[10:14:00.0000] <takkaria>
the footnote issue is a weird one

[10:41:00.0000] <hsivonen>
/me learns that KML uses escaped presentational HTML inside XML

[12:34:00.0000] <yecril71>
Several publishers want to have their own version of HTML.

[12:34:01.0000] <yecril71>
This is all right, as long as they write for each other.

[12:35:00.0000] <yecril71>
They will have to implement and share the same browser as well.

[12:35:01.0000] <yecril71>
The Web is for people the publisher has never heard of.

[12:36:00.0000] <yecril71>
The publisher cannot expect them to use their browser.

[12:36:01.0000] <yecril71>
So this initiative is doomed to end up in a closed circle.

[12:36:02.0000] <Lachy>
yecril71, WTF?

[12:37:00.0000] <yecril71>
I mean, Pentasis and Pascual.

[12:38:00.0000] <yecril71>
Such custom formats are better served with XML though.

[12:38:01.0000] <Lachy>
oh, I haven't read that thread yet

[14:22:00.0000] <BenMillard>
Philip`, are you active? I'm going through pages in your q-tags.txt but some uses of <q> have gone: http://philip.html5.org/data/q-tags.txt

[14:22:01.0000] <BenMillard>
specifically, I can't find <q> on http://phoenix.lowtech.org/

[14:28:00.0000] <BenMillard>
also http://www.cotswoldlibdems.org.uk/

[15:33:00.0000] <BenMillard>
there's another LibDem site which uses the same markup and design as the Cotswolds one, so I've given them the same analysis

[15:34:00.0000] <Philip`>
BenMillard: I'm not especially active

[15:40:00.0000] <takkaria>
just the occasional walk, then?

[15:44:00.0000] <BenMillard>
Philip`, no worries. I'm noting "Gone" and reviewing as best I can from the sample markup in that text file.

[15:44:01.0000] <BenMillard>
so far, I've been able to review all of them thanks to the sample :)

[15:45:00.0000] <Philip`>
BenMillard: I don't have an entirely trivial way to see the content from the versions of the pages that I downloaded, so it's probably easier to try archive.org for them


2008-11-03
[16:04:00.0000] <BenMillard>
Philip`, oh yeah! thanks for reminding me

[00:24:00.0000] <cyclist>
Does the table processing model describe the algorithm to caclulate the column widths?

[00:27:00.0000] <cyclist>
from my understanding it only describes how to determine which column/row *slots* are assigned to each cell

[00:31:00.0000] <roc>
calculating column widths is a CSS issue

[00:34:00.0000] <cyclist>
ok

[04:28:00.0000] <hsivonen>
http://lists.w3.org/Archives/Public/www-style/2008Nov/0008.html I wonder what the rationale for three sets of metrics is...

[04:37:00.0000] <zcorpan>
hmm... wonder which cases to consider when designing how to be bandwidth-conservative with <audio> and <video>

[04:37:01.0000] <zcorpan>
new Audio(src) is likely to be used for games so should be downloaded in full

[04:38:00.0000] <zcorpan>
or perhaps we should have a way for authors to say "i want to preload the whole thing"

[04:39:00.0000] <zcorpan>
there's no new Video(src)

[04:40:00.0000] <zcorpan>
perhaps there's no need to preload videos

[04:40:01.0000] <zcorpan>
or perhaps the offline stuff should be used i one wants to preload stuff

[07:25:00.0000] <Philip`>
/me wonders why pressing backspace to delete the first character in a contenteditable span in FF3 causes it to not delete the character, and instead to insert a space before the span

[07:27:00.0000] <Philip`>
...but only when the span is preceded by a space character

[07:35:00.0000] <Philip`>
Ah, looks like https://bugzilla.mozilla.org/show_bug.cgi?id=455992

[07:36:00.0000] <Philip`>
/me hopes someone gets around to fixing it some time

[08:02:00.0000] <shepazu>
hsivonen: http://www.economist.com/vote2008/?mode=description

[08:41:00.0000] <Philip`>
shepazu: It's odd that they wait until the last paragraph of their analysis (saying "our readers in America also massively favoured the Democrat over the Republican, by 81% to 19%") to indicate that their polling method is rubbish and totally unrepresentative of reality

[08:49:00.0000] <gavin>
heh

[08:50:00.0000] <gavin>
wow, in context they even admit that the number doesn't match reality

[08:51:00.0000] <gsnedders>
/me waves

[08:53:00.0000] <Philip`>
/me particles

[08:54:00.0000] <Philip`>
gavin: But they only say the American number doesn't match reality, and don't retract their claims about the views of the rest of the world

[08:54:01.0000] <gsnedders>
Am I the only person who finds Philip` ridiculously funny on IRC?

[08:54:02.0000] <gavin>
yeah

[08:55:00.0000] <Philip`>
(Never mind that a (not even random) minimum sample of 10 people from each country is pretty useless)

[08:55:01.0000] <Philip`>
gavin: He is the only one? :-(

[08:55:02.0000] <gavin>
(to Philip`)

[08:55:03.0000] <Philip`>
Ah, phew

[08:58:00.0000] <gsnedders>
My standards suck interview still isn't up? Those guys suck.

[10:37:00.0000] <Philip`>
/me loves Gentoo

[10:38:00.0000] <smedero>
why's that?

[10:38:01.0000] <Philip`>
The com_err package was blocking an upgrade of e2fsprogs, presumably because its functionality is now bundled into e2fsprogs for some strange reason

[10:38:02.0000] <Philip`>
so I uninstalled com_err, so that it would let me upgrade e2fsprogs

[10:38:03.0000] <Philip`>
But wget relies on com_err, so it became unable to download e2fsprogs :-(

[10:39:00.0000] <Philip`>
s/it/the package installation system/

[10:41:00.0000] <gsnedders>
c'est la vie :P

[10:42:00.0000] <Philip`>
(curl relies on it too, but fortunately GET (from Perl's LWP) doesn't, so it let me download the package into the expected location, and it's all okay now)

[10:42:01.0000] <gsnedders>
/me got conditional offer from York (at AH BB in maths and physics)

[11:13:00.0000] <aboodman>
Hixie: ping on worker feedback

[11:13:01.0000] <Hixie>
i'm on vacation til 11/11, but it's on my list of things to reply to when i get back

[11:14:00.0000] <BenMillard>
I've got 20 URLs left to categorise from Philip's <q> collection: http://philip.html5.org/data/q-tags.txt

[11:14:01.0000] <BenMillard>
39 entries so far (a couple of URLs have multiple entries)

[11:15:00.0000] <BenMillard>
s/20 URLs left/22 URLs left/

[11:32:00.0000] <Philip`>
/me doesn't like how Gmail makes it harder to mark a spam message as spam than to simply delete it

[11:32:01.0000] <Philip`>
(If you mark it as spam then you have to delete it from the spam folder later)

[11:34:00.0000] <Philip`>
Only the tenuous hope that marking as spam will inform some heuristics and reduce the likelihood of future spam causes me to keep doing so

[12:19:00.0000] <gsnedders>
q really is the new acronym.

[14:54:00.0000] <sicking>
Hixie, aboodman: ping

[14:54:01.0000] <Hixie>
hey

[14:54:02.0000] <sicking>
Hixie, so we realized that there is no spec in webworkers for figuring out what useragent is running

[14:55:00.0000] <sicking>
Hixie, i.e. something like navigator.userAgent

[14:55:01.0000] <Hixie>
yeah

[14:55:02.0000] <sicking>
Hixie, it's trivial to implement, and seems like it might help future compat a lot

[14:55:03.0000] <sicking>
Hixie, so i think we'll want to try to get it in for this release

[14:55:04.0000] <Hixie>
we should just expose the relevant parts of navigator

[14:55:05.0000] <sicking>
Hixie, should we simply expose a .navigator object which has the properties that are defined in html5?

[14:55:06.0000] <Hixie>
which is defined now in html5

[14:55:07.0000] <Hixie>
yeah

[14:55:08.0000] <Hixie>
what you said

[14:56:00.0000] <sicking>
  readonly attribute DOMString appName;

[14:56:01.0000] <sicking>
  readonly attribute DOMString appVersion;

[14:56:02.0000] <sicking>
  readonly attribute DOMString platform;

[14:56:03.0000] <sicking>
  readonly attribute DOMString userAgent;

[14:56:04.0000] <Hixie>
right

[14:56:05.0000] <Hixie>
i'm on vacation right now, will be back 11/11, but please do send feedback so i can add it when i get back

[14:56:06.0000] <sicking>
ok, we'll probably implement before then

[14:56:07.0000] <sicking>
but i'll send to list

[14:57:00.0000] <othermaciej>
sicking: send to the list, I'll ask Alexey to add it to ours too

[14:59:00.0000] <sicking>
othermaciej, cool

[15:27:00.0000] <aboodman>
hixie, sicking ... catching up

[15:28:00.0000] <aboodman>
sicking: what hixie said

[15:28:01.0000] <sicking>
cool

[15:28:02.0000] <aboodman>
sicking: i would like to come to agreement wrt dual interfaces for workers.

[15:28:03.0000] <aboodman>
i am still uncomfortable with it the way that it is.

[15:29:00.0000] <aboodman>
and i know you guys are getting ready to ship, so i think it's important to get consensus.

[15:29:01.0000] <sicking>
aboodman, i don't feel strongly either way. I'm happy with the way it is but don't really mind your changes.

[15:29:02.0000] <aboodman>
oh, cool.

[15:29:03.0000] <sicking>
aboodman, we don't implement shared workers for now at all, so as far as that goes i think it doesn't matter much

[15:29:04.0000] <sicking>
and we don't implement MessagePorts

[15:30:00.0000] <aboodman>
well, it makes a difference only in that the dedicated worker interfaces don't exist in my proposal.

[15:30:01.0000] <aboodman>
so you'd have to do some subset of the sharedworker interface.

[15:31:00.0000] <sicking>
we don't implement the dedicated interfaces separately anyway, i don't think that's detectable from the worker

[15:31:01.0000] <sicking>
we simply have a nsIDOMWorker interface

[15:31:02.0000] <aboodman>
right, but it has nsIDOMWorker::sendMessage(), I assume.

[15:31:03.0000] <sicking>
right, but that was the case in your proposal still right?

[15:32:00.0000] <aboodman>
no.

[15:32:01.0000] <olliej>
aboodman: you should talk to ap and weinig

[15:32:02.0000] <aboodman>
gar s/sendMessage/postMessage/g :)

[15:32:03.0000] <sicking>
i though it was in the "compromise"

[15:32:04.0000] <sicking>
heh

[15:32:05.0000] <olliej>
aboodman: i believe ap is getting all the worker apis implemented in webkit atm

[15:34:00.0000] <aboodman>
I can start a new thread with the full set of changes I suggest. Will mozilla and webkit review it?

[15:34:01.0000] <aboodman>
or would something else be easier?

[15:34:02.0000] <sicking>
aboodman, you got a link to the old thread? or remember the subject?

[15:35:00.0000] <aboodman>
sicking: looking

[15:36:00.0000] <sicking>
aboodman, found it

[15:36:01.0000] <aboodman>
ko.

[15:36:02.0000] <sicking>
aboodman, in the message you sent 9/24 11:38 it looks like postmessage is still there

[15:36:03.0000] <sicking>
aboodman, though not sure if that changes later in the thread

[15:37:00.0000] <sicking>
oh

[15:37:01.0000] <sicking>
no

[15:37:02.0000] <sicking>
it's not

[15:37:03.0000] <aboodman>
sicking: that is me summarizing our conversation I believe.

[15:37:04.0000] <aboodman>
that is what we compromised to.

[15:37:05.0000] <aboodman>
I wanted to keep the whatwg list up to date on the changes.

[15:38:00.0000] <sicking>
right, what's in the message is what i remember we compromised on, and still has postMessage/onmessage directly on the worker context

[15:39:00.0000] <sicking>
aboodman, basically it adds 'connect' to dedicated workers

[15:39:01.0000] <aboodman>
sicking: yes, we are on the same page. however, i did not view discussion on worker.postMessage() as closed.

[15:40:00.0000] <aboodman>
i was just basically checkpointing what we had agreed to up at that point.

[15:40:01.0000] <sicking>
aboodman, ok

[15:40:02.0000] <aboodman>
i would still like to resolve the remaining disagreements.

[15:40:03.0000] <sicking>
aboodman, sorry, didn't realize that was the case. It's getting really close to shipping here

[15:41:00.0000] <sicking>
aboodman, which isn't neccesarily a problem unless we decide to remove postMessage directly on the worker

[15:41:01.0000] <aboodman>
sicking: I definitely feel that it would be better to not have that interface, and to only use connect().

[15:42:00.0000] <aboodman>
It seems to me that the majority case is using connect() anyway.

[15:42:01.0000] <sicking>
aboodman, i know we disagree on that :)

[15:42:02.0000] <aboodman>
heh.

[15:42:03.0000] <aboodman>
usually it falls to hixie to resolve disagreements on merit, but he is away, and i just don't want this to fall off the radar and then get implemented.

[15:43:00.0000] <sicking>
i think the problem here is figuring out which use case is the common one

[15:45:00.0000] <aboodman>
well, do you agree that connect() is a superset?

[15:45:01.0000] <sicking>
yeah

[15:46:00.0000] <sicking>
well

[15:46:01.0000] <sicking>
it's a different syntax

[15:46:02.0000] <aboodman>
i mean a superset of capabilities

[15:46:03.0000] <aboodman>
agree that it isn't a compatible superset, hence my concern.

[15:46:04.0000] <sicking>
seems capabailties are the same in both syntaxes, can't think of anything you can do in one but not the other

[15:46:05.0000] <aboodman>
technically, we are allowed to change the whatwg proposals after vendors implement them.

[15:46:06.0000] <aboodman>
in reality, it becomes harder after vendors implement them.

[15:47:00.0000] <sicking>
absolutely

[15:48:00.0000] <aboodman>
how about this: I will start a new thread attempting to summarize my concerns as succinctly as possible. I'll also pester the relevant apple people to chime in.

[15:48:01.0000] <sicking>
aboodman, sounds good, and hixie if possible


2008-11-04
[17:57:00.0000] <Hixie>
aboodman, sicking: personally i think of the proposals i've seen i like what's in the spec best, but i'll spec whatever you end up agreeing on if you do come to an agreeemnt that's good

[17:58:00.0000] <Hixie>
aboodman, sicking: however i'd recommend against compromising on things just to get agreement, you should definitely think that whatever you agree with is the best if you agree to it, not just agree for the sake of getting progress

[17:58:01.0000] <Hixie>
that way lies compromise-driven madness

[18:00:00.0000] <sicking>
yay madness!

[18:07:00.0000] <othermaciej>
Hixie: are you willing to compromise on your stance against compromise-driven madness?

[18:07:01.0000] <Hixie>
no :-)

[18:07:02.0000] <Hixie>
though i'll spec whatever gets implemented, so there are ways to override me :-)

[18:09:00.0000] <Dashiva>
othermaciej: He'll have to wait for the committee to finish evaluating

[23:07:00.0000] <aboodman>
hi weinig, just posted that mail i promised

[23:07:01.0000] <weinig>
aboodman: sweet!

[23:10:00.0000] <aboodman>
argh, please replace all occurences of 'sendMessage()' with 'postMessage()'. I keep mixing up the gears and html5 names.

[01:02:00.0000] <zcorpan>
hmm both safari and firefox seem crash prone if you throw 10 video elements at them in the same page

[01:08:00.0000] <olliej>
zcorpan: could you file a bug at bugs.webkit.org? :D

[01:20:00.0000] <hsivonen>
zcorpan: would be good to have your test case on b.m.o, too :-)

[01:26:00.0000] <doublec>
zcorpan, what version of firefox?

[01:31:00.0000] <zcorpan>
Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.1b2pre) Gecko/20081031 Minefield/3.1b2pre

[01:36:00.0000] <zcorpan>
olliej, hsivonen: see http://simon.html5.org/test/html/semantics/video/loading-demos/

[01:36:01.0000] <olliej>
zcorpan: is it first click safe?

[01:37:00.0000] <zcorpan>
yes it's a dir page

[01:37:01.0000] <zcorpan>
don't have time right now to file bugs though sorry

[01:45:00.0000] <roc>
it's not crashing for me, although it isn't really working either

[01:46:00.0000] <roc>
oh, I think I hit the HTTP connection limit or someting

[01:46:01.0000] <roc>
now they're all playing

[01:46:02.0000] <roc>
poorly

[01:46:03.0000] <doublec>
yes I see the same

[01:57:00.0000] <zcorpan>
roc: it doesn't crash for m,e every time

[01:59:00.0000] <zcorpan>
firefox actually first showed a gray area instead of the page when i tried to load 002.html, and then did the same for any other page, then when i tried to quit firefox it crashed. i pressed the restart firefox button and then it showed the uninstall minefield dialog (and restarted firefox)

[01:59:01.0000] <zcorpan>
very weird

[02:01:00.0000] <roc>
mmm

[02:52:00.0000] <Philip`>
http://radar.barackobama.com/ - wow, it's a fancy animated map thing that isn't Flash - I think that's the first time I've ever seen one that isn't

[02:52:01.0000] <Philip`>
The layout's a bit broken in Opera, but it's the thought that counts

[02:53:00.0000] <Philip`>
/me presumes it's using jQuery for animatedness

[03:08:00.0000] <zcorpan>
i presume this is spam http://forums.whatwg.org/viewtopic.php?t=4025

[03:09:00.0000] <zcorpan>
same ip

[03:10:00.0000] <hsivonen>
are robocall in the U.S. legal for political advertising? I thought they were made illegal for telemarketing in general.

[03:10:01.0000] <Philip`>
http://www.google.com/search?q=%22I+need+to+ask+you+guyz+a+question%22

[03:15:00.0000] <Philip`>
hsivonen: I believe they are legal, purely on the basis that if they weren't then a lot of people would have complained very loudly about it, and I haven't heard anyone complaining that they're anything other than rude and irritating

[03:19:00.0000] <Philip`>
(Also, Wikipedia says lots of states have different rules for political organisations than commercial ones, for automated phone calls)

[03:20:00.0000] <Philip`>
(That combination of methods of proof-by-not-hearing-anyone-else-claim-it's-untrue and proof-by-Wikipedia is clearly infallible)

[03:22:00.0000] <Philip`>
/me discovers http://uncyclopedia.wikia.com/wiki/Proof as a useful repository of such proof methods

[04:23:00.0000] <jcranmer>
Philip`: I love Proof-by-Wikipedia

[04:28:00.0000] <jcranmer>
I suppose that was a dumb idea

[04:29:00.0000] <jcranmer>
if I choose to reply to this email, I will probably embroil myself in a debate as to whether or not the HTML specification should be roughly as precise as a law

[04:30:00.0000] <jcranmer>
/me chooses not to respond

[04:30:01.0000] <hsivonen>
/me notes that law has very different precision in different countries

[04:31:00.0000] <hsivonen>
some countries assume you #include "commonsense.h"

[04:33:00.0000] <jcranmer>
jcranmer $ locate commonsense.h

[04:33:01.0000] <jcranmer>
/usr/include/bitbucket/commonsense.h

[04:33:02.0000] <jcranmer>
jcranmer $ ls -l /usr/include/bitbucket/commonsense.h

[04:35:00.0000] <jcranmer>
-r--r--r--  1 root root     18 2008-01-01 00:00 /usr/include/bitbucket/commonsense.h -> /dev/null

[04:35:01.0000] <jcranmer>
:-(

[04:38:00.0000] <Philip`>
/me wonders which email jcranmer is referring to

[04:40:00.0000] <jcranmer>
Philip`: off-list

[04:40:01.0000] <Philip`>
Ah

[06:55:00.0000] <zcorpan>
Hixie: you probably want to hide the "watch for updates" box for print media

[06:56:00.0000] <Philip`>
What if you print onto dynamic e-paper?

[06:59:00.0000] <zcorpan>
then it still overlaps the other text

[07:00:00.0000] <zcorpan>
and dynamic e-paper might well use the screen media instead

[07:00:01.0000] <zcorpan>
s/and/or/

[07:01:00.0000] <zcorpan>
i wonder why the status boxes aren't in the print version

[07:07:00.0000] <zcorpan>
"???? (KUROSAWA Takeshi)" says the pdf

[07:27:00.0000] <Philip`>
The link underlining in http://yaml.org/spec/1.2/#Introduction looks disturbingly weird to me

[07:28:00.0000] <Philip`>
It looks like they intentionally made it look that way, but I can't imagine why

[08:06:00.0000] <jcranmer>
egads

[08:07:00.0000] <jcranmer>
anyone here remember Dmitry Turin, or did he only do stuff on the CSS lists?

[08:08:00.0000] <jcranmer>
(well, I'm sure that many core people here watch both lists anyways)

[08:08:01.0000] <Philip`>
He was on public-html too

[08:08:02.0000] <jcranmer>
ah

[08:08:03.0000] <Philip`>
Sadly he seems to have been absent lately

[08:08:04.0000] <jcranmer>
well, consider him back as of today

[08:09:00.0000] <jcranmer>
he just made some posts to the mozilla newsgroups of all places

[08:09:01.0000] <Philip`>
Excellent!

[08:09:02.0000] <jcranmer>
Subject: HTML6 budget

[08:09:03.0000] <jcranmer>
it goes downhill from there, IMO

[08:09:04.0000] <Philip`>
That's a very promising start

[08:10:00.0000] <Philip`>
I liked it when he gave a link to a 200-slide Powerpoint presentation

[08:26:00.0000] <MikeSmith>
Good to hear that Dmitry Turin is back

[08:26:01.0000] <MikeSmith>
I was worried that he might have given up on his dreams

[08:28:00.0000] <jcranmer>
I'm not in the mood to deal with him in this case

[08:29:00.0000] <jcranmer>
I'm already embroiled in three or four long threads

[08:31:00.0000] <Philip`>
As far as I can tell, nobody has extracted any value from discussions with him on public-html, so it seems safe to ignore him

[08:31:01.0000] <jcranmer>
I definitely will do that

[08:34:00.0000] <MikeSmith>
I think in many cases Dmitry is not necessarily expecting a response anyway

[08:34:01.0000] <Philip`>
(People have suggested that he should try identifying problems and use cases rather than just presenting a complete solution without explaining why it's worthwhile, but I haven't seen him follow that advice yet)

[08:35:00.0000] <jcranmer>
I think he is

[08:35:01.0000] <jcranmer>
expecting

[08:35:02.0000] <Dashiva>
Philip`: He saves us all the work of going through those phases by giving us the complete solution

[08:36:00.0000] <Philip`>
(He does seem to have put a lot of effort into his work, so it'd be nice if he could redirect that effort towards something that would actually be practical)

[08:36:01.0000] <hsivonen>
mozilla.dev.planning it seems

[08:36:02.0000] <jcranmer>
Philip`: "He needs concrete budget. So i'm asking you to estimate and say, how much will it cost."

[08:37:00.0000] <jcranmer>
he wants a response, it seems

[08:37:01.0000] <MikeSmith>
I don't think Dmitry's interested in mundane things like examining and documenting use cases or investigating whether there may be some spec or technology that already solves the problems he's created complete solutions for

[08:37:02.0000] <jcranmer>
MikeSmith: of course not

[08:38:00.0000] <Philip`>
jcranmer: Did you mean s/Philip`/MikeSmith/?

[08:38:01.0000] <jcranmer>
Philip`: oh, right

[08:38:02.0000] <jcranmer>
sorry

[08:38:03.0000] <jcranmer>
MikeSmith: it only works with his new solutions

[08:38:04.0000] <Philip`>
Apology accepted

[08:38:05.0000] <Philip`>
:-p

[08:38:06.0000] <MikeSmith>
jcranmer: right

[08:39:00.0000] <MikeSmith>
He wants to build the whole complete rocket ship to the moon, from scratch

[08:39:01.0000] <Philip`>
MikeSmith: I think he is aware of existing technology - e.g. he looked at the existing XPath technology, and then changed the syntax a bit so that he liked it more

[08:39:02.0000] <MikeSmith>
Philip`: OK

[08:40:00.0000] <Philip`>
or at least gave a few examples of what the syntax ought to look like

[08:40:01.0000] <Philip`>
which is totally enough to justify making a new XML path selector standard

[08:43:00.0000] <Philip`>
MikeSmith: Hmm, how long ago did people realise that making a moon rocket would require a lot of work? I think it was some Jules Verne thing where one guy built one himself, whereas nowadays everyone realises you need giant corporations or government organisations to do it, but I have no idea when people switched from one view to the other...

[08:44:00.0000] <MikeSmith>
Philip`: me neither..

[08:44:01.0000] <Philip`>
MikeSmith: Alas :-(

[08:55:00.0000] <MikeSmith>
related to that "how long ago.." question - http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2008-November/016985.html seems sort of relevant

[08:55:01.0000] <MikeSmith>
the part where Mike Schinkel writes, "The irony is that if TBL had gotten this kind of resistance"...

[09:01:00.0000] <MikeSmith>
and Hixie points out that TimBL actually implemented what he spec'ed

[09:05:00.0000] <Dashiva>
"This is very much a case of empowering serendipity" <- What?

[09:07:00.0000] <MikeSmith>
the thing with Dmitry Turin -- and I guess a good number of others as well -- is that they spin these grand plans without any intention of actually implementing them. The expect that somebody else will implement it for them.

[09:07:01.0000] <MikeSmith>
the word empowering should be banned from all discussions. anybody who uses it should receive a thrubbing

[09:07:02.0000] <Dashiva>
And that everyone will use them once implemented.

[09:08:00.0000] <Philip`>
/me made an entry for the ICFP contest where you had to write an AI bot that would navigate a map and find food and push competing robots out of the way, and called it Serendipity because that was the only way it would ever actually find any food

[09:08:01.0000] <MikeSmith>
heh

[09:09:00.0000] <Dashiva>
Oh, that quote. "what does your robot do, sam?" "it collects data about the surrounding environment, then discards it and drives into walls"

[09:11:00.0000] <Philip`>
It would be funnier if it weren't so true

[09:12:00.0000] <Philip`>
/me once made a robot which triggered its own reset switch every time it hit an obstacle

[09:13:00.0000] <MikeSmith>
robots shouldn't do practical stuff -- the should to crazy, unpredictable stuff, and make strange sounds

[09:14:00.0000] <MikeSmith>
http://en.wikipedia.org/wiki/Dorkbot

[09:14:01.0000] <Philip`>
Also, it's unfortunate that the power of Lego motors is greater than the strength of the Lego chassis in which I usually put them

[09:14:02.0000] <MikeSmith>
"people doing strange things with electricity"

[09:15:00.0000] <MikeSmith>
Philip`: out-of-control torque is a great thing to watch

[09:15:01.0000] <erlehmann>
Philip`: pixxx!

[09:16:00.0000] <Philip`>
Attaching four Lego motors (powered from a 12V adapater) to a half-meter S shape made of Lego monorail track is great fun, particularly when mounted on top of a combat robot

[09:17:00.0000] <Philip`>
*adapter

[09:17:01.0000] <Philip`>
You can get quite a lot of kinetic energy out of them

[09:19:00.0000] <erlehmann>
wat

[09:20:00.0000] <erlehmann>
back to biz - are there any deprecated items in the HTML5 spec ?

[09:21:00.0000] <Philip`>
erlehmann: No - you're either allowed to write things, or not allowed to

[09:21:01.0000] <erlehmann>
also, i believe that section 4.8.2.1.3 should contain strong wording to use CSS

[09:21:02.0000] <zcorpan>
erlehmann: why?

[09:21:03.0000] <Philip`>
(though of course you can still write things that you're not allowed to, and they'll still work)

[09:22:00.0000] <Philip`>
(for some understanding of 'work')

[09:22:01.0000] <erlehmann>
CSS3 content replacement would probably make the section obsolete anyway

[09:23:00.0000] <zcorpan>
erlehmann: why would you use css for logos and icons?

[09:25:00.0000] <erlehmann>
zcorpan: i already do, if they conver no additional semantic info. for navlists as presented in the spec i just change the "list icon" (how is it called).

[09:25:01.0000] <zcorpan>
erlehmann: you didn't answer my question though :)

[09:25:02.0000] <erlehmann>
common icons usually are deco that can be replaced without changing semantics, therefore i use CSS.

[09:26:00.0000] <zcorpan>
ah

[09:26:01.0000] <erlehmann>
like doing a web site in tango style, then switching over to aqua look.

[09:26:02.0000] <erlehmann>
most icons have the same purposes as colors - merely being stylistic elements

[09:27:00.0000] <zcorpan>
i don't quite agree -- icons are closer to the content than the color scheme

[09:27:01.0000] <zcorpan>
pretty much like italics vs font face

[09:28:00.0000] <erlehmann>
well yes, but they usually convey semantics that are already there.

[09:29:00.0000] <Philip`>
The semantics usually aren't there unless you specifically add class names as hooks for the images

[09:29:01.0000] <erlehmann>
consider a change of a corporations' logo. on the corporation website, it is purely stylistic. on a web site discussing design and typography the logo is semantic content.

[09:30:00.0000] <jcranmer>
... must ... resist ... troll ... posting

[09:30:01.0000] <jcranmer>
...

[09:30:02.0000] <zcorpan>
it's not just stylistic -- it's important for users to recognize where they are

[09:31:00.0000] <Philip`>
If you're changing the logo, you'd just overwrite /images/logo.gif and it doesn't matter how you're referencing it in a page :-)

[09:31:01.0000] <Philip`>
Uh oh

[09:32:00.0000] <Philip`>
Dmitry came back to public-html

[09:45:00.0000] <jcranmer>
oh, I just realized why I didn't see that post... I subscribe to all but public-html :-)

[10:17:00.0000] <erlehmann>
Philip`: that proves that the image by itself has no semantic meaning, doesn't it ?

[10:17:01.0000] <erlehmann>
which was my point, btw

[11:15:00.0000] <Philip`>
/me wonders why people are being so rude to Pentasis

[11:15:01.0000] <jcranmer>
Philip`: his assertion that specs shouldn't be written by/for browser vendors?

[11:19:00.0000] <Philip`>
jcranmer: That's a reason to strongly disagree, particularly when he's making those claims within the context of the WHATWG which is fundamentally designed in opposition to his claims

[11:19:01.0000] <Philip`>
but it doesn't seem like a reason to explicitly laugh at him or to say "Now GTFO my web"

[11:22:00.0000] <jcranmer>
Philip`: I haven't seen any of the recent messages, so I can't riposte

[11:29:00.0000] <erlehmann>
Philip`: the point is that his criticism is pointless. he claims that WHATWG has "not the right" to evolve HTML. fine, fine, go make another standard ! what pentasis does is akin to flaming on the KDE mailing list that they should use GTK.

[11:42:00.0000] <jcranmer>
I wonder what Hixie's position is

[12:16:00.0000] <Hixie>
my position on the guy who said "GTFO my web" is that i told him that was unacceptable; if he does it again he'll get banned for a week

[12:20:00.0000] <erlehmann>
Hixie: oh, that would be me.

[12:20:01.0000] <erlehmann>
i'll watch my output, then.

[12:45:00.0000] <gsnedders>
oh hell

[12:46:00.0000] <gsnedders>
/. commenhs are now worse than digg at tines

[12:46:01.0000] <jcranmer>
weren't they always?

[12:46:02.0000] <doublec>
this the theora thread?

[12:48:00.0000] <gsnedders>
doublec: yep. totally not reading what I'm saying

[12:48:01.0000] <gsnedders>
jcranmer: Rarely.

[12:50:00.0000] <gsnedders>
Apparently I'm suggesting Ogg has patent problems. I'm not. I'm saying Theora's patent status beyond On2's patents is unknown.

[12:50:01.0000] <Philip`>
But when talking about patents, being unknown is a problem

[12:51:00.0000] <gsnedders>
But Ogg != Theora.

[12:53:00.0000] <Philip`>
Ah

[12:53:01.0000] <Philip`>
But they're close enough :-p

[12:55:00.0000] <gsnedders>
The guy who is arguing with me knows the difference, but thinks I am saying Ogg has problems and can't believe he's arguing with someone who thinks that.

[12:55:01.0000] <Dashiva>
Why don't we just encode the video as sound

[12:55:02.0000] <Dashiva>
That way we're safe!

[12:56:00.0000] <Philip`>
Dashiva: That won't work - it wouldn't be accessible to deaf people

[12:56:01.0000] <gsnedders>
:D

[12:57:00.0000] <Dashiva>
They could convert back to video. It's all just waves anyway

[12:57:01.0000] <Philip`>
Ah, cunning

[12:58:00.0000] <jcranmer>
gsnedders: I see what you mean

[12:58:01.0000] <Philip`>
Of course, somebody will have patented the audio/video conversion algorithm

[12:59:00.0000] <gsnedders>
jcranmer: /. comments in general or the thread specifically?

[13:02:00.0000] <jcranmer>
gsnedders: specific thread

[13:02:01.0000] <gsnedders>
Can I be bothered to write a long comment?

[13:03:00.0000] <gsnedders>
http://slashdot.org/comments.pl?threshold=-1&commentsort=0&mode=nested&cid=25625761

[13:03:01.0000] <gsnedders>
hmm

[13:03:02.0000] <jcranmer>
/me attempts to gauge how much will make it through

[13:04:00.0000] <gsnedders>
that URI doesn't work

[13:05:00.0000] <jcranmer>
Hixie: good thing I did some discussion off-list, then :-)

[13:08:00.0000] <gsnedders>
I had a discussion like it at TPAC. Some people were taking it up with timbl, who grabbed me (or rather called me, first by another name beginning with g, then correctly) as a passing HTML WG member.

[13:09:00.0000] <gsnedders>
Got some history about HTML's relation to SGML

[13:10:00.0000] <Dashiva>
"You look like a suitable decoy"

[13:10:01.0000] <gsnedders>
:D

[13:28:00.0000] <aboodman>
sicking, othermaciej_: i wanted to make sure that you saw i resumed the thread about workers as promised.

[13:28:01.0000] <othermaciej_>
aboodman: I did

[13:28:02.0000] <sicking>
aboodman, yeah, will comment

[13:29:00.0000] <sicking>
aboodman, it's a little fuzzy on what new API you are proposing though

[13:29:01.0000] <sicking>
aboodman, such as, is there an implicit call to 'connect' when a shared worker is created?

[13:29:02.0000] <aboodman>
sicking: yes. should i follow up with the IDL?

[13:30:00.0000] <sicking>
aboodman, i'll comment first

[13:31:00.0000] <aboodman>
sicking: i misread your question. there is no implicit connect() call. callers must always call connect() explicitly.

[13:31:01.0000] <sicking>
aboodman, ah, that's what i suspected, though you don't mention that change

[13:32:00.0000] <aboodman>
sicking: my apologies, i wasn't sure what would be the least lossy way to communicate my ideas.

[13:50:00.0000] <sicking>
othermaciej: I don't see your reply yet on the whatwg list btw

[13:53:00.0000] <othermaciej_>
I did not reply

[13:53:01.0000] <othermaciej_>
I did see the thread

[13:55:00.0000] <othermaciej_>
I agree with Alexey's reply

[13:55:01.0000] <othermaciej_>
and also the general sentiment that it would be good to reduce the number of interfaces to the messaging mechanism

[14:12:00.0000] <sicking>
othermaciej_, aboodman: my concern with forcing connect() to be called on a dedicated worker is that it makes the simple case so complicated for the user. I've detailed in a response

[14:13:00.0000] <sicking>
othermaciej_, aboodman: in any case this is something that we need to decide on very soon. We're freezing today, but i might be able to get changes done for a few more days

[14:20:00.0000] <othermaciej>
sicking: it doesn't seem that much more complicated to me, particularly considering that Workers are an advanced feature in the first place

[14:21:00.0000] <othermaciej>
sicking: I tend to think overall API simplicity is more important here than optimizing the simple case

[14:21:01.0000] <sicking>
othermaciej, even with the nested event handlers?

[14:22:00.0000] <sicking>
othermaciej, I do think there is value in making the simple case simpler, if it's the common use case

[14:23:00.0000] <Hixie>
/me thought what we had today was pretty simple as it was

[14:23:01.0000] <sicking>
othermaciej, creating a single API that fits all use cases doesn't always yield the simplest API for any of them

[14:23:02.0000] <othermaciej>
sicking: so you are really concerned about one extra line of code in 4-5 line code sequences?

[14:23:03.0000] <sicking>
/me agrees with Hixie

[14:23:04.0000] <othermaciej>
I have to admit I don't care much about this either way, other than wanting a resolution

[14:24:00.0000] <othermaciej>
but I agree with aboodman that having three different ways to do worker-related messaging overall seems like a mess

[14:24:01.0000] <sicking>
othermaciej, mostly the nested functions, with two different events in scope are complicated

[14:24:02.0000] <Hixie>
i don't really see that we have three ways

[14:24:03.0000] <Hixie>
they're all the same way

[14:24:04.0000] <sicking>
well, startConversaion is just a convinence method, not really a new mechanism. I'm fine with dropping that

[14:25:00.0000] <Hixie>
and startConversation was only added in response to aaron's request :-)

[14:25:01.0000] <sicking>
ultimately there is only one communication mechanism: postMessage

[14:26:00.0000] <othermaciej>
I have a hard time getting worked up about it either way

[14:26:01.0000] <othermaciej>
like I said

[14:26:02.0000] <sicking>
with aarons proposal you'll actually always have to use two mechanisms, connect() and postMessage()

[14:26:03.0000] <othermaciej>
I think this is an advanced feature in any case

[14:26:04.0000] <sicking>
i mostly agree, i'm ok with either, though it might be hard to get postMessage dropped from FF3.1 at this point

[14:27:00.0000] <othermaciej>
I'd be more concerned about the complexity of determining worker lifetime if anything

[14:27:01.0000] <othermaciej>
although that's more a concern for implementations than an API issue

[14:28:00.0000] <Hixie>
i'm ok with whatever API people want, but in the face of apathy i'm strongly in favour of the status quo

[14:28:01.0000] <Hixie>
(if the arguments aren't convincing, as, imho, in this case)

[14:50:00.0000] <aboodman>
Jonas, hixie, I don't see how you think there is only one mechanism.

[14:51:00.0000] <aboodman>
Maybe we are having a terminology problem.

[14:51:01.0000] <sicking>
aboodman, the only way you can send data is through 'postMessage'

[14:51:02.0000] <aboodman>
yes, but the method is located on different objects.

[14:51:03.0000] <aboodman>
and must be initialized differently.

[14:51:04.0000] <Hixie>
the only mechanism is the message channel mechanism

[14:51:05.0000] <Hixie>
and it's initialised automatically whenever possible

[14:53:00.0000] <aboodman>
so again, in the case of dedicated workers, you have worker.postMessage(). In the case of shared workers, you have worker.port.postMessage().

[14:53:01.0000] <aboodman>
the inside is also different.

[14:54:00.0000] <sicking>
i guess if you consider them different or not is a matter of definition. But they aren't *that* different IMHO

[14:55:00.0000] <aboodman>
well it is different. a developer must remember that in the case of dedicated workers you can call postMessage() on the worker object, but not on shared workers.

[14:55:01.0000] <aboodman>
I don't see why this should be the case.

[14:55:02.0000] <Hixie>
in the case of shared workers, it can't be optimised further since there are multiple ports

[14:56:00.0000] <Hixie>
it's not worker.port.postMessage()

[14:56:01.0000] <Hixie>
it's just myPort.postMessage()

[14:56:02.0000] <Hixie>
oh you mean on the outside

[14:56:03.0000] <Hixie>
not the inside

[14:56:04.0000] <sicking>
aboodman, making all cases as complicated as the most complicated isn't neccesarily a win IMHO

[14:56:05.0000] <sicking>
necessarily

[14:56:06.0000] <aboodman>
Hixie: yes, i was talking about the outside.

[14:56:07.0000] <Hixie>
we could remove the .port on the shared worker case, but that would be an oversimplification of the API, imho

[14:57:00.0000] <Hixie>
i don't understand why we'd want to expose the port only on one side

[14:57:01.0000] <aboodman>
Hixie: what difference would that make?

[14:57:02.0000] <Hixie>
that would be even mroe confusing

[14:57:03.0000] <aboodman>
Hixie: why would that be more confusing?

[14:58:00.0000] <sicking>
Hixie, the advantage of not having an implicit connect() is that you remove the correlation one-page-one-connection

[14:58:01.0000] <Hixie>
the answer to the question "Is this a MessagePort or something special?" would be "well..."

[14:58:02.0000] <sicking>
Hixie, so if you have some service that does server-db access for example and you want several such connections from a single page, you act the same way for each such connection

[14:59:00.0000] <sicking>
Hixie, and on the inside you don't care about what connection is from a new page and what is a new connection from an already existing page

[14:59:01.0000] <aboodman>
Since people are not that thrilled with startConversation/connect, I can make an alternate proposal that also collapses the shared worker/dedicated worker duality.

[14:59:02.0000] <aboodman>
but does not have startConversation

[15:00:00.0000] <aboodman>
OUTSIDE: var worker = new SharedWorker("foo.js", "foo"); worker.postMessage("ping");

[15:00:01.0000] <sicking>
aboodman, i do like the separation between shared and dedicated. Because it allows us to keep the dedicated simpler than the shared one inheritly can be

[15:01:00.0000] <aboodman>
INSIDE: onmessage = function(e) { e.messagePort.sendMessage("pong") }

[15:01:01.0000] <Hixie>
look i don't mind going back to the generic mechanism we used to have, where shared and dedicated workers were indistinguishable, and there was exactly one mechanism

[15:01:02.0000] <Hixie>
but nobody liked that

[15:01:03.0000] <Hixie>
i don't really understand the point of trying to get that withotu actually getting it

[15:02:00.0000] <Hixie>
i don't have a problem with startConversation()

[15:02:01.0000] <Hixie>
i think it's fine

[15:03:00.0000] <aboodman>
Hixie, I think that my last proposal is almost exactly the same as your initial one, except that there is a first class Worker object.

[15:03:01.0000] <aboodman>
But the Worker object, does not have any postMessage() related apis on it.

[15:03:02.0000] <Hixie>
how does this address sicking's original complaints?

[15:04:00.0000] <Hixie>
(for a while the spec did have an api with a first class worker object with a pipe on it, btw)

[15:04:01.0000] <aboodman>
it doesn't, I was just clarifying that basically I am in favor of going back to your original design with a few minor alterations.

[15:05:00.0000] <Hixie>
i'm not, sicking said he wouldn't implement it :-)

[15:05:01.0000] <aboodman>
gotcha.

[15:07:00.0000] <aboodman>
sicking: I think that if we get rid of connect/startConveration then we can combine the two interfaces w/o adding any code at all to the DedicatedWorker case. I think this would be a win.

[15:07:01.0000] <Hixie>
why?

[15:07:02.0000] <Hixie>
(just curious)

[15:07:03.0000] <Hixie>
(not saying there's no reason it'd be a win)

[15:08:00.0000] <aboodman>
the same reason code reuse is usually good

[15:08:01.0000] <sicking>
aboodman, well, you'll have a weird implicit port created on the inside that isn't reachable until someone posts

[15:08:02.0000] <aboodman>
less opportunity for bugs.

[15:08:03.0000] <sicking>
aboodman, i know someone (maciej?) wasn't too exited about not being able to post out until someone had posted in

[15:08:04.0000] <Hixie>
are we talking inside or outside here? i'm confused. what are you saying you want to change exactly?

[15:09:00.0000] <Hixie>
there are use cases for posting in both directions first, we shouldn't preclude that in either case

[15:09:01.0000] <aboodman>
I don't think we need to preclude posting from outwardly first.

[15:09:02.0000] <aboodman>
s/from//

[15:09:03.0000] <sicking>
aboodman, how would you post out then?

[15:09:04.0000] <Hixie>
what are you saying you want to change exactly? relative to the current spec

[15:10:00.0000] <sicking>
aboodman, i.e. what object would you call postMessage on?

[15:10:01.0000] <aboodman>
in both the dedicated and shared case you'd call postmessage on the worker object.

[15:10:02.0000] <sicking>
aboodman, what would that send a message to on the shared worker

[15:10:03.0000] <sicking>
?

[15:10:04.0000] <aboodman>
getting to that :)

[15:10:05.0000] <aboodman>
on the inside there are onconnect and onmessage events

[15:12:00.0000] <aboodman>
in both events you get a "port" (maybe not exactly the same thing currently called MessagePort) in the event object which allows you to talk back to the outer interface that sent you  a message or connected.

[15:13:00.0000] <aboodman>
note: there is no global postMessage() or port object on the inside of workers for either the dedicated or shared case.

[15:13:01.0000] <sicking>
aboodman, 'in the event'? if you want to post outward first there is no event, no?

[15:13:02.0000] <sicking>
oh

[15:13:03.0000] <aboodman>
sicking: you get onconnect

[15:13:04.0000] <aboodman>
which is basically onload

[15:13:05.0000] <aboodman>
in the case of dedicated workers

[15:15:00.0000] <sicking>
hm

[15:15:01.0000] <sicking>
it still looks to me like you're adding complexity to the dedicated worker solely for the case of making it more similar to the shared one

[15:16:00.0000] <aboodman>
i have reduced the amount of added complexity to only the case where the worker wants to send messages outward first.

[15:17:00.0000] <sicking>
sure, you're adding just a little complexiy, but still seems like it's only for the sake of making it more similar to the dedicated worker

[15:18:00.0000] <sicking>
hm, btw, where do messages appear?

[15:18:01.0000] <aboodman>
i suppose. the global complexity is lower though. it seems like a great trade-off to me.

[15:18:02.0000] <sicking>
if i do   w = new Worker(...); w.postMessage('foo');

[15:18:03.0000] <sicking>
where do i catch that on the inside?

[15:18:04.0000] <aboodman>
global onmessage event

[15:19:00.0000] <sicking>
so there's a global onmessage but no global postMessage?

[15:19:01.0000] <aboodman>
yes.

[15:19:02.0000] <aboodman>
that is the proposal.

[15:19:03.0000] <aboodman>
which makes sense functionally -- the relationship between workers and clients is 1 to many.

[15:20:00.0000] <sicking>
that seems unexpected given that we don't have that anywhere else

[15:21:00.0000] <aboodman>
not sure what to tell you. it doesn't seem weird to me.

[15:22:00.0000] <sicking>
i still prefer what we currently have (possibly modulo removing startConversaion and the implicit call to connect() on shared workers)

[15:22:01.0000] <sicking>
aboodman, but do post to the list and see what others feel

[15:22:02.0000] <aboodman>
sicking: how do you think that shared workers should talk to the outside?

[15:24:00.0000] <sicking>
aboodman, if we remove implicit connect()?

[15:25:00.0000] <aboodman>
sicking: yeah, i don't understand what your proposal is for shared workers in general.

[15:26:00.0000] <aboodman>
the current spec for them doesn't make a lot of sense to me

[15:26:01.0000] <sicking>
start with as now

[15:26:02.0000] <sicking>
add a connect() method

[15:26:03.0000] <sicking>
remove .port

[15:26:04.0000] <aboodman>
i see.

[15:26:05.0000] <sicking>
when connect() is called a port is returned

[15:27:00.0000] <sicking>
and a 'connect' event is fired inside the worker

[15:27:01.0000] <sicking>
the connect event object holds the port for the other end of the channel

[15:27:02.0000] <aboodman>
i think part of the disconnect (hah!) is that you think connect() is a feature of shared workers only. i think it is useful for all workers. i actually think it is _less_ useful for shared workers, since you can get basically the same effect by just creating multiple instances of the shared worker.

[15:27:03.0000] <aboodman>
so if it isn't on dedicated workers, i wouldn't put it on shared workers

[15:28:00.0000] <sicking>
hmm

[15:28:01.0000] <sicking>
interesting idea

[15:28:02.0000] <sicking>
so just talking about shared workers for a sec

[15:28:03.0000] <Hixie>
wait, what's connect)?

[15:28:04.0000] <Hixie>
connect()?

[15:29:00.0000] <sicking>
Hixie, in my described solution above it raises a connect event inside the worker with a port, and then returns the other port

[15:29:01.0000] <sicking>
aboodman, so just talking about shared workers for a sec, how would you do communication with a shared worker?

[15:30:00.0000] <aboodman>
i think that what i proposed above would be my ideal solution

[15:30:01.0000] <sicking>
so there is a worker.postMessage

[15:30:02.0000] <aboodman>
var worker = new SharedWorker("foo.js", "foo"); worker.postMessage("ping"); ||| onmessage = function(e) { e.messagePort.postMessage("pong"); }

[15:33:00.0000] <sicking>
aboodman, (still just talking about shared workers). So that looks nice, except the disconnect between onmessage/no-postMessage. So how about this: take that proposal, except that when a new shared worker is instantiated a 'connect' message is raised. That message contains a port that has postMessage/onmessage

[15:34:00.0000] <Hixie>
sicking: why would we want that?

[15:34:01.0000] <sicking>
Hixie, want what?

[15:34:02.0000] <aboodman>
sicking: i prefer it w/o special cases for dedicated workers. i don't think the lack of a global postMessage() is a big deal.

[15:34:03.0000] <Hixie>
connect()

[15:34:04.0000] <sicking>
Hixie, there seems to have been some disconnect about that

[15:34:05.0000] <Hixie>
yeah sorry i'm not really paying much attention here, supposed to be on vacation :-)

[15:35:00.0000] <Hixie>
watching electrion coverage and replying to two e-mails at the same time too

[15:37:00.0000] <sicking>
aboodman, still feels like doing that is just adding complexity to the dedicated worker for the sole purpose of having it more similar to the shared one

[15:37:01.0000] <sicking>
aboodman, dedicated workers are by nature simpler, why not let them be

[15:38:00.0000] <sicking>
aboodman, with this new proposal the difference is minimal

[15:39:00.0000] <sicking>
aboodman, no difference on the outside, and on the inside the only difference is that shared workers have to listen to 'connect' to get new incoming communication channels

[15:39:01.0000] <sicking>
aboodman, whereas the dedicated worker simply lets the global scope be the communication channel

[15:40:00.0000] <aboodman>
sicking: yes, it is a big improvement. i don't think the remaining differences are that big a deal.

[15:40:01.0000] <aboodman>
i still think it would be better if the inner interfaces were the same, but this difference won't bother me that much if it is what browser supported.

[15:41:00.0000] <sicking>
aboodman, i do really like to keep onmessage/postMessage in sync

[15:42:00.0000] <aboodman>
<shrug> ok.

[15:42:01.0000] <sicking>
wanna mail to the list or should I?

[15:42:02.0000] <aboodman>
i will

[15:42:03.0000] <sicking>
cool, do bring up both proposals

[15:42:04.0000] <aboodman>
which ones?

[15:43:00.0000] <aboodman>
oh you mean the two that are very close together.

[15:43:01.0000] <sicking>
right

[15:43:02.0000] <aboodman>
ok, sure.

[15:43:03.0000] <sicking>
thanks

[15:43:04.0000] <aboodman>
on the bright side, i believe this would mean no changes to things you've already implemented.

[15:43:05.0000] <sicking>
yup

[15:44:00.0000] <sicking>
that'd be the case with either of them


2008-11-05
[23:32:00.0000] <BenMillard>
I've finished analysing the <q> collection Philip` gathered: http://projectcerbera.com/web/study/2008/quotes#pt

[23:35:00.0000] <BenMillard>
I e-mailed the list with my findings: http://lists.w3.org/Archives/Public/public-html/2008Nov/0011.html

[23:52:00.0000] <BenMillard>
I also reviewed the implementability of internationalised automatic quote mark generation on <q> in relation to what HTML4 says and what HTML5 could say: http://lists.w3.org/Archives/Public/public-html/2008Nov/0012.html

[23:52:01.0000] <BenMillard>
basically, the feature seems far more trouble than it's worth

[23:53:00.0000] <BenMillard>
hmm, despite proofreading and revising those 2 messages several times over the past week there are still some bad typos :(

[01:54:00.0000] <yecril71>
Why should interruptions be more problematic for accessibility than any other part of speech?

[01:59:00.0000] <yecril71>
If you want to mark up content for consistently and flexibly,

[01:59:01.0000] <yecril71>
there are other open standards that achieve them.

[02:00:00.0000] <yecril71>
However, you probably need a strong scientific background

[02:00:01.0000] <yecril71>
and several months of preparation to use them correctly.

[02:00:02.0000] <yecril71>
Otherwise your document will turn out into a big blob of nonsense markup.

[02:02:00.0000] <yecril71>
HTML should be simple enough for a high school child to use.

[02:04:00.0000] <yecril71>
An abbreviation is no more a reference than any word is.

[02:04:01.0000] <hsivonen>
yecril71: are you referring to a particular email message?

[02:04:02.0000] <yecril71>
Re: [whatwg] ---

[02:05:00.0000] <yecril71>
Every word, as it is written, is a reference to a notion,

[02:05:01.0000] <yecril71>
except where it is used obliquely, as a reference to itself.

[02:05:02.0000] <yecril71>
Those cases, however, are rare.

[02:07:00.0000] <yecril71>
So a reference is too broad a term to name an element.

[02:09:00.0000] <yecril71>
Sup and sub are not style elements.

[02:10:00.0000] <yecril71>
Text browsers cannot just render them transparently with no harm to the meaning.

[02:10:01.0000] <yecril71>
PRE is not a style element because it alters the parsing mode.

[02:11:00.0000] <yecril71>
VAR is not the same as function; 

[02:12:00.0000] <yecril71>
you can have <VAR >f</VAR > and <I >f</I > with different meanings.

[02:12:01.0000] <yecril71>
I mean, both are functions, 

[02:12:02.0000] <yecril71>
but the second refers to a well-known function

[02:13:00.0000] <yecril71>
and the first is sort of local, incidental.

[02:13:01.0000] <hsivonen>
yecril71: does a reader appreciate the distinction?

[02:14:00.0000] <yecril71>
I think so; it is better to know what kind of a symbol is presented.

[02:14:01.0000] <yecril71>
Symbols are hard to read and any clues are welcome.

[02:15:00.0000] <yecril71>
(That is why code editors mark various elements with different colours).

[02:15:01.0000] <hsivonen>
I'd argue that one's literary expression sucks if it is so ambiguous that it wouldn't work on paper and the reader needs to inspect markup

[02:16:00.0000] <yecril71>
Who says it would not work on paper?

[02:16:01.0000] <hsivonen>
(assuming that usually sin() and f() aren't presented in different colors)

[02:16:02.0000] <yecril71>
But they are presented in different typefaces.

[02:17:00.0000] <hsivonen>
sin is often presented without italics, so your <i> example above doesn't apply

[02:17:01.0000] <hsivonen>
bad example from me

[02:17:02.0000] <yecril71>
(Actually, the ISO standard recomments well-known symbols to be typeset upright).

[02:17:03.0000] <yecril71>
recommends.

[02:17:04.0000] <hsivonen>
the ISO standard?

[02:18:00.0000] <yecril71>
The italic typeface could be styled out in a context where only symbols appear.

[02:18:01.0000] <yecril71>
There is an ISO standard for typesetting scientific publications.

[02:18:02.0000] <yecril71>
And they say well-known objects should be upright.

[02:19:00.0000] <yecril71>
It is rarely used because the TeX community did not fully embrace it.

[02:20:00.0000] <hsivonen>
:-)

[02:24:00.0000] <yecril71>
The case of symbols inserted into running text is a bit different.

[02:25:00.0000] <yecril71>
I think it would be acceptable to leave them in italics 

[02:25:01.0000] <yecril71>
if typeface distinction is not available.

[02:27:00.0000] <yecril71>
I agree documents should be readable without styling;

[02:27:01.0000] <yecril71>
they are just easier to read with proper styling than without.

[02:41:00.0000] <yecril71>
One example of something unreadable without styling

[02:41:01.0000] <yecril71>
is 14<SUP ><U >25</U ></SUP >

[02:41:02.0000] <yecril71>
which in Polish means 2:15 PM.

[02:43:00.0000] <yecril71>
This probably should be coded as

[02:44:00.0000] <yecril71>
<time ><hrs >14</hrs ><sep >:</sep ><mins >25</mins ></time >

[02:45:00.0000] <yecril71>
where sep is display:none.

[02:49:00.0000] <yecril71>
End body tag otherwise means nothing when there is no body element open.

[02:56:00.0000] <yecril71>
Any markup language can be used to describe how content is tagged.

[02:57:00.0000] <yecril71>
The "HT" in "HTML" means it is text with hyperlinks for easy publishing on the Web.

[02:57:01.0000] <yecril71>
There are other markup schemes to do scientific research on text corpora.

[02:58:00.0000] <yecril71>
Not everything must be coded in HTML.

[03:36:00.0000] <yecril71>
An attribute is a bad choice for a title because attributes are not rendered.

[03:36:01.0000] <yecril71>
At least should not be.

[04:23:00.0000] <Lachy>
Oh No! Microsoft have discontinued Windows 3.x. http://tech.slashdot.org/article.pl?sid=08/11/05/0257202&from=rss

[04:26:00.0000] <jcranmer>
Lachy: did you just join the 21st century?

[04:26:01.0000] <hsivonen>
/me wonders what it feels like to maintain a severely legacy piece of software knowing what the state of the art is

[04:40:00.0000] <operairc>
hello

[04:42:00.0000] <Lachy>
operairc, hi

[06:14:00.0000] <Lachy>
any suggestions for what font I should use in my presentation slides? It has to be done in Powerpoint and so the font has to be available on Windows

[06:21:00.0000] <hsivonen>
Lachy: Arial in that cas

[06:21:01.0000] <hsivonen>
e

[06:21:02.0000] <Lachy>
I wanted to avoid Arial

[06:22:00.0000] <hsivonen>
(alternatively, you could do your thing in Keynote, export to a bunch of TIFF, drag&drop the TIFF set onto a blank Keynote presentation to import them back and then export the resulting bitmap slides to PowerPoint)

[06:22:01.0000] <Philip`>
What version of Windows?

[06:22:02.0000] <Lachy>
I've chosen Book Antiqua for now

[06:22:03.0000] <Philip`>
(Vista has nicer fonts)

[06:22:04.0000] <Lachy>
I only have XP available

[06:23:00.0000] <Philip`>
Ah

[06:23:01.0000] <Lachy>
and I think the machines at the conference have XP too

[06:23:02.0000] <Philip`>
Comic Sans is clearly the sensible choice

[06:23:03.0000] <Lachy>
true, but I figured that's what everyone else would be using and wanted to be different

[06:24:00.0000] <Philip`>
You could use italic Comic Sans instead

[06:25:00.0000] <Lachy>
exporting keynote is possible. But I also have to create it with OperaShow to keep chaals happy. I'm supposed to present using OperaShow if possible, but I still need to have the powerpoint as a backup

[06:25:01.0000] <hsivonen>
wow. Windows has crazy bugs. now I need to sniff if my Python script runs on Windows...

[06:26:00.0000] <hsivonen>
one needs to stuff an extra quote into the argument of os.system() so that Windows can swallow the quote and then see the right command

[06:27:00.0000] <Lachy>
This is the panel I'm on, in which I need to give a 15 min presentation on HTML5. http://www.pubcon.com/sessions.cgi?action=view&record=150

[06:35:00.0000] <Philip`>
hsivonen: Do you really have to use os.system() and rely on the system shell, rather than using the subprocess module or something?

[06:37:00.0000] <hsivonen>
Philip`: I have no idea. os.system has served me well on OS X and Ubuntu

[06:39:00.0000] <Philip`>
http://docs.python.org/library/os.html#os.system suggests subprocess

[06:39:01.0000] <Philip`>
which should be more reliable since it doesn't use the shell, so there aren't all the problems of quoting and escaping

[06:41:00.0000] <hsivonen>
Philip`: that looks like more work than stuffing an extra quote in there

[06:44:00.0000] <hsivonen>
what should I use on windows in place of os.symlink?

[06:45:00.0000] <Philip`>
Pre-Vista doesn't have anything equivalent to symlinks, so you'd need to reconsider your problem from a higher level

[06:45:01.0000] <hsivonen>
zcorpan: do you read the WHATWG implemontors list?

[06:46:00.0000] <zcorpan>
hsivonen: yes

[06:47:00.0000] <zcorpan>
though you had already replied when i saw the email

[06:47:01.0000] <hsivonen>
zcorpan: OK. then I don't try to reply to the reply to my reply.

[06:50:00.0000] <hsivonen>
Philip`: what about shortcuts? if I create shortcuts to make a directory appear in multiple places, will Python shutil follow shortcuts?

[06:51:00.0000] <hsivonen>
and will Java file IO follow shortcuts in paths?

[06:51:01.0000] <Philip`>
hsivonen: No (unless shutil is really crazy, which I don't think it is) - shortcuts are purely an Explorer thing, not a kernel or filesystem thing

[06:52:00.0000] <hsivonen>
hmm. that's not cool

[06:52:01.0000] <Philip`>
so the Win32 API doesn't understand shortcuts, and just treats them like plain files (which is all they are)

[06:53:00.0000] <hsivonen>
I guess I need to write a Python function that copies instead of symlinking on Windows

[06:53:01.0000] <hsivonen>
sigh

[06:53:02.0000] <Philip`>
(Pre-Vista NTFS does have junction points, which are kind of like symlinks to directories, but they're pretty obscure and nobody uses them)

[06:53:03.0000] <jcranmer>
 /nick jcranmer|away

[07:02:00.0000] <Hixie>
junction points are directory hard links

[07:02:01.0000] <Hixie>
(and mount points)

[07:03:00.0000] <Philip`>
They're not hard links in the Unixy sense, i.e. just a different name for the same node on disk - there's a distinction between the original directory and the junction points pointing to it

[07:04:00.0000] <Hixie>
oh?

[07:04:01.0000] <Hixie>
yeah i guess so

[07:05:00.0000] <Hixie>
ntfs also has multiple streams per file, which as far as i can tell is used exclusively by malware authors

[07:05:01.0000] <Philip`>
e.g. Wikipedia says "The dir command in Windows 2000 or later recognizes junction points, displaying <JUNCTION> instead of <DIR> in directory listings"

[07:05:02.0000] <Hixie>
yeah

[07:06:00.0000] <Philip`>
Maybe that's because anyone other than malware authors wants to avoid getting support calls from users with FAT filesystems :-)

[07:07:00.0000] <othermaciej>
hard links in the Unixy sense *are* different names for the same node on disk

[07:07:01.0000] <othermaciej>
whereas symlinks are different names for a path in the filesystem

[07:08:00.0000] <othermaciej>
it may be that junction points are not exactly like either

[07:08:01.0000] <Philip`>
othermaciej: That's what I meant about hard links

[07:08:02.0000] <othermaciej>
I see

[07:08:03.0000] <Philip`>
so I agree with you :-)

[07:09:00.0000] <othermaciej>
are junction points references to the filesystem object directly, or to a path?

[07:10:00.0000] <othermaciej>
to me that's the biggest difference between symlinks and hardlinks, more so than being indistinguishable from the original

[07:10:01.0000] <Philip`>
I guess the issue there is: if you move the pointed-at directory, does the junction point still point at it?

[07:10:02.0000] <Hixie>
junction points are basically mount points for other paths, iirc

[07:11:00.0000] <Hixie>
but what do i know

[07:11:01.0000] <Philip`>
and I don't know the answer to that

[07:11:02.0000] <Hixie>
ask me an html question!

[07:11:03.0000] <Hixie>
:-)

[07:14:00.0000] <Philip`>
/me can't actually think of any HTML questions :-(

[07:14:01.0000] <Hixie>
hah

[07:14:02.0000] <Hixie>
i have 1000s, it's ok

[07:16:00.0000] <Philip`>
Hmm, the fun fair outside my bedroom window has been playing loud music for an hour, and it's not even going to open until three hours from now

[07:18:00.0000] <Philip`>
(Hooray for bonfire night, where we celebrate the capture and torture and mutilation and death of a suspected terrorist)

[08:07:00.0000] <gsnedders>
Philip`: You still living where you were in July?

[08:26:00.0000] <Philip`>
gsnedders: No

[08:27:00.0000] <gsnedders>
Philip`: Where do you live now?

[08:27:01.0000] <gsnedders>
Philip`: Somewhere overlooking a park I guess :P

[08:27:02.0000] <Philip`>
gsnedders: Around http://maps.google.co.uk/?ie=UTF8&ll=52.209553,0.131621&spn=0.003373,0.010085&t=h&z=17

[08:29:00.0000] <Philip`>
I have a nicer view from my room than I did in July :-)

[08:44:00.0000] <gsnedders>
Yeah, you wouldn't see much from where you were then :)

[08:44:01.0000] <gsnedders>
Only 59 hours by train from Le Mans to Istabul! :P

[09:25:00.0000] <Hixie>
89.

[09:35:00.0000] <yecril71>
What is wrong with

[09:35:01.0000] <yecril71>
<STRONG class="requirements" >Enter an odd number! <INPUT ></STRONG >?

[09:37:00.0000] <yecril71>
And consider that Hixie thinks it is right

[09:38:00.0000] <yecril71>
that labels do not work in Internet Explorer.

[09:38:01.0000] <yecril71>
Since they officially should not work, 

[09:38:02.0000] <yecril71>
it is not a big deal that they actually do not.

[09:48:00.0000] <Hixie>
?

[09:48:01.0000] <gsnedders>
Hixie: ?

[09:50:00.0000] <Hixie>
not sure what you meant

[09:58:00.0000] <gsnedders>
Hixie: I mean the 89

[10:00:00.0000] <gsnedders>
Hixie: My statement of the 59 hours made sense :P

[10:02:00.0000] <Hixie>
er i meant not sure what yecril71 meant, sorry

[10:07:00.0000] <yecril71>
I meant labels, Internet Explorer, inputs and error messages, of course.

[10:08:00.0000] <yecril71>
What if we put the input inside the validation message?

[10:08:01.0000] <gsnedders>
Hixie: But what was your 89?

[10:09:00.0000] <yecril71>
My 89 was fine, thanks.

[10:10:00.0000] <yecril71>
I graduated from high school.

[13:14:00.0000] <gsnedders>
w00t

[13:14:01.0000] <gsnedders>
I really have made it with the latest last week post!


2008-11-06
[00:12:00.0000] <hsivonen>
I wish Apple shipped a more robust file system.

[00:13:00.0000] <hsivonen>
here I am with a corrupt HFS+ and in the middle of a SuperDuper! and Time Machine dance

[00:58:00.0000] <zcorpan>
hmm i'm a bit disappointed about markp's misinformation

[01:02:00.0000] <hsivonen>
zcorpan: in this week?

[01:03:00.0000] <zcorpan>
hsivonen: in http://blog.whatwg.org/the-road-to-html-5-episode-1-the-section-element

[01:03:01.0000] <hsivonen>
I'm disappointed at Time Machine running the same brittle file system that it is backing up

[01:04:00.0000] <hsivonen>
I'm going to lose at least this day to hfs+

[01:05:00.0000] <hsivonen>
fortunately, I've got and offsite boot disk clone from immediately before tpac

[01:06:00.0000] <hsivonen>
I should probably go get it in case time machine turns out to be useless

[01:19:00.0000] <hsivonen>
I wonder if there's any way  to get feedback of waht's going on when Finder beach balls to perform force fsck on the tm disk

[01:21:00.0000] <zcorpan>
http://www.w3.org/MarkUp/2008/ED-xhtmlmime-20081024/ is getting a bit better

[01:25:00.0000] <zcorpan>
"DO ensure that any CSS properties on the html element are also specified on the body element." hmm that wasn't there when i reviewed the document last time

[01:26:00.0000] <hsivonen>
zcorpan: did you get a response to your comments?

[01:26:01.0000] <zcorpan>
hsivonen: yes

[01:27:00.0000] <zcorpan>
http://lists.w3.org/Archives/Public/public-xhtml2/2008Oct/thread.html#msg25

[01:27:01.0000] <zcorpan>
though not all of my comments

[01:29:00.0000] <hsivonen>
more positive than the initial telecon reaction

[01:30:00.0000] <zcorpan>
well i was a bit harsh so that's understandable

[01:35:00.0000] <zcorpan>
/me wonders which browser tommy is working on

[01:38:00.0000] <zcorpan>
http://www.kvaleberg.com/ maybe

[01:41:00.0000] <pergj>
zcorpan: if you ask around in the (Opera) office, someone will know. :)

[01:45:00.0000] <zcorpan>
maybe

[01:46:00.0000] <zcorpan>
hah i didn't even look at his email address

[01:49:00.0000] <pergj>
There is only one name of a (former) browser vendor on our webpage, so you can probably figure it out by yourself.

[01:50:00.0000] <zcorpan>
openwave

[01:53:00.0000] <zcorpan>
"Browser

[01:53:01.0000] <zcorpan>
Bookmarks, enter url, history"

[01:53:02.0000] <zcorpan>
compelling features :)

[01:54:00.0000] <zcorpan>
do you use the html5 parser for application/xhtml+xml?

[01:55:00.0000] <pergj>
the product described on the webpage does not have much to do with the browser

[01:55:01.0000] <zcorpan>
is there more detailed information about the browser somewhere?

[01:56:00.0000] <pergj>
http://www.purplelabs.com/news-press-release-view.php?code=90

[01:57:00.0000] <pergj>
at the moment we are using the html5 parser for xhtml too

[01:58:00.0000] <zcorpan>
do you plan to change that?

[01:59:00.0000] <pergj>
we should at some point, but right now there are other priorities. 

[02:00:00.0000] <zcorpan>
hey tthorsen

[02:01:00.0000] <tthorsen>
Hi. I saw the irc log, but it took me some time to get XChat installed

[02:02:00.0000] <tthorsen>
did that mail I sent to the list make sense?

[02:03:00.0000] <zcorpan>
not sure if the parser change works but i noticed that none of the top 4 browsers match html5 right now

[02:03:01.0000] <zcorpan>
oh wait maybe firefox does

[02:03:02.0000] <zcorpan>
yep

[02:04:00.0000] <zcorpan>
did it break some site to follow html5?

[02:04:01.0000] <tthorsen>
Yes. http://bankrate.com

[02:05:00.0000] <pergj>
for this particular case Opera is the same as the spec while firefox is not

[02:05:01.0000] <pergj>
I meant that Opera does the same as Tommy's proposed change

[02:05:02.0000] <zcorpan>
oh sorry i was looking at the wrong dom viewer. firefox also doesn't match the spec

[02:06:00.0000] <pergj>
What Firefox does is maybe slightly better

[02:06:01.0000] <tthorsen>
yes. Actually they both handle my example markup identically, but if you add some cdata after "thirddiv" (but still inside "firstform") you'll see a difference

[02:08:00.0000] <tthorsen>
Seems firefox is able to see that the first </form> belongs to the ignored "secondform" element, whereas Opera does it pretty much like my proposed change.

[02:09:00.0000] <zcorpan>
the proposed change seems to match ie's rendering

[02:09:01.0000] <zcorpan>
(try to set a border on the form)

[02:10:00.0000] <zcorpan>
(the first form i.e.)

[02:15:00.0000] <zcorpan>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3Cdiv%20id%3D%22firstdiv%22%3E%0D%0A%20%20%20A%0D%0A%20%20%20%3Cdiv%20id%3D%22seconddiv%22%3E%0D%0A%20%20%20%20%20%20%20%3Cform%20id%3D%22firstform%22%20style%3Dborder%3Asolid%3E%0D%0A%20%20%20%20%20%20%20%20%20%20%20%3Cdiv%20id%3D%22thirddiv%22%20style%3Dborder%3Adotted%20red%3E%0D%0A%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%3C%2Fform%3Exx%0D%0A%20%20%20%20%20%20%20%20%20%20%20%3C%2Fdiv

[02:15:01.0000] <zcorpan>
%3Eyy%0D%0A%20%20%20%20%20%20%20%3C%2Fform%3E%0D%0A%20%20%20%3C%2Fdiv%3E%0D%0A%20%20%20B%0D%0A%3C%2Fdiv%3E

[02:15:02.0000] <zcorpan>
(uploaded)

[02:16:00.0000] <zcorpan>
i think this has changed back and forth and whatever you do will break some site

[02:18:00.0000] <tthorsen>
well, so long as a site works in the most popular browsers, then there must be some algorithm which will work

[02:19:00.0000] <pergj>
At least the "original" bankrate case is working in all browsers. Tommy's extended case is a different matter though.

[02:19:01.0000] <zcorpan>
btw did you implement the svg/mathml parts in the html5 parser?

[02:20:00.0000] <pergj>
no

[02:20:01.0000] <zcorpan>
ok

[02:20:02.0000] <pergj>
we never supported mathml or inline svg

[02:25:00.0000] <zcorpan>
btw let me know if web dom core is of any help to you

[02:28:00.0000] <pergj>
ok. At the moment our DOM implementation has not caused us as much pain as our old handling of broken pages. :)

[02:29:00.0000] <zcorpan>
makes sense :)

[03:03:00.0000] <virtuelv>
hmph

[03:04:00.0000] <virtuelv>
the single-document html5 spec is way too big, and the multipage version doesn't allow me to glance over an element index, or IDL defined by the spec

[03:16:00.0000] <Philip`>
virtuelv: You need a better browser, so it can handle the single page version with no problems

[03:16:01.0000] <Philip`>
virtuelv: Does the (multipage) table of contents not work as a rough index of elements?

[03:47:00.0000] <virtuelv>
Philip`: not when you're looking for something like showNotification or showModalDialog

[03:52:00.0000] <Philip`>
virtuelv: Ah, right

[03:53:00.0000] <Philip`>
If you can't use a decent browser, you could perhaps just download the HTML and use grep

[03:53:01.0000] <Philip`>
but that's not a very elegant solution :-(

[03:54:00.0000] <Philip`>
If the multipage version could include some extra index pages, what kind of thing would be most useful?

[03:54:01.0000] <Philip`>
(like, maybe a list of all the linked terms in the whole document? or a concatenation of all the IDL blocks? or something like that?)

[03:56:00.0000] <virtuelv>
I *have* a decent browser

[04:06:00.0000] <Philip`>
Then you should be able to handle the single page version with no problems :-)

[04:38:00.0000] <hsivonen>
hmm. apparently even Time Capsule wouldn't have avoided the problem I have

[04:40:00.0000] <hsivonen>
Time Capsule exposes a sparse image instead of a file system access protocol, so presumably the kernel of the failing system that is being backed up manages the backup file system and can break it

[04:42:00.0000] <hsivonen>
I want a Time Capsule where the backup file system isn't managed by the computer whose failure is the threat that TC is supposed to insure against

[05:24:00.0000] <Lachy>
hsivonen, what happened to your file system?

[05:44:00.0000] <Philip`>
hsivonen: You want two computers and a network and rsync

[06:07:00.0000] <hsivonen>
Lachy: a windowserver malfunction lead to an unclean reboot with ctrl-cmd-power which apparently didn't flush file systems properly and corrupted both the boot and TM file systems

[06:08:00.0000] <hsivonen>
Philip`: rather, I want my file systems to be robust and isolated from the os x kernel

[06:08:01.0000] <hsivonen>
a raid nas with zfs internally and an nfs export would be great

[06:09:00.0000] <Philip`>
But you also want it to be cheap and easy to set up? :-)

[06:09:01.0000] <hsivonen>
of course

[06:09:02.0000] <hsivonen>
The Time Capsule form factor would be nice

[06:48:00.0000] <zcorpan>
Hixie: id and hidden are in the wrong order in the list of global attributes

[06:51:00.0000] <zcorpan>
/me updated http://simon.html5.org/html5-elements

[07:04:00.0000] <Dashiva>
Is there a specific term for all those crazy words Dmitry keeps coming up with?

[07:04:01.0000] <Dashiva>
Sament and bament and all that

[07:05:00.0000] <Dashiva>
Taking the start of the first word and the end of the last word

[07:07:00.0000] <yecril71>
Presenting a confirmation dialogue box before the file selection dialogue box is just silly.

[07:07:01.0000] <yecril71>
What would the text be?

[07:08:00.0000] <yecril71>
"The following dialogue will allow you to choose a file from your local storage

[07:08:01.0000] <yecril71>
that the browser will be able to upload.

[07:08:02.0000] <Lachy>
I couldn't figure out what "Bament" was supposed to mean.

[07:09:00.0000] <yecril71>
Do you really want to do this?"

[07:09:01.0000] <yecril71>
LOL

[07:10:00.0000] <Dashiva>
Lachy: Apparently there's a "Baribute" too

[07:10:01.0000] <zcorpan>
Dashiva: http://en.wikipedia.org/wiki/Portmanteau_word ?

[07:11:00.0000] <Lachy>
hah

[07:12:00.0000] <Dashiva>
I can understand wanting to optimize for smaller documents, but optimizing for a smaller spec by making terms into nonsense...

[07:12:01.0000] <Lachy>
I wish someone would respond to Dmitry and explain to him why he's not getting a response to his absurd budget request

[07:12:02.0000] <Philip`>
I wish nobody would respond to Dmitry

[07:13:00.0000] <Lachy>
I meant off-list

[07:13:01.0000] <Lachy>
I can't do it cause he's on my do-not-respond list

[07:13:02.0000] <zcorpan>
he's on my do-not-read list

[07:14:00.0000] <Lachy>
some of his emails have entertainment value, like the first 3 lines of the budget request

[07:14:01.0000] <Lachy>
I didn't read beyond that

[07:14:02.0000] <yecril71>
I think Portugal is an interesting country to live in.

[07:15:00.0000] <yecril71>
Computers tend to come to people in Portugal by themselves.

[07:15:01.0000] <Lachy>
yecril71, what?

[07:15:02.0000] <yecril71>
Moreover, they come equipped with Microsoft XP in English.

[07:15:03.0000] <Lachy>
are your comments related to something in particular?

[07:15:04.0000] <yecril71>
And the poor Portuguese people cannot do anything with them.

[07:16:00.0000] <yecril71>
My comments are related to the file submission browse button label.

[07:16:01.0000] <yecril71>
I am also interested how the Catalan people would react to a label in Catalan

[07:16:02.0000] <yecril71>
on a Spanish page.

[07:17:00.0000] <Philip`>
According to his 209-slide presentation, "Sament" = "SAg eleMENT"

[07:17:01.0000] <Philip`>
and "Sag" = "Sent tAG"

[07:17:02.0000] <Philip`>
I'm guessing that "Bament" comes from "Browser tag element", but the etymology doesn't seem to be documented here

[07:18:00.0000] <Philip`>
and I can't guess what Kament is

[07:19:00.0000] <yecril71>
Having it the other way round makes them very uneasy.

[07:22:00.0000] <yecril71>
I must say I am not very surprised.

[07:24:00.0000] <yecril71>
I like the idea of having semantics separate from structure very much.

[07:25:00.0000] <yecril71>
That would e.g. allow declaring the title attribute based on a set of rules.

[07:26:00.0000] <yecril71>
Repeating it for every item I want to describe is very inefficient and/or troublesome.

[07:27:00.0000] <yecril71>
From the point of view of HTML, it would be LINK[rel=semantics]

[07:27:01.0000] <yecril71>
and perhaps SCRIPT[type="text/xml"], or perhaps "text/something+xml".

[07:28:00.0000] <yecril71>
Where the LINK target can be a bookmark set on the SCRIPT.

[07:29:00.0000] <yecril71>
Of course, it is not clear that XML would be the medium for semantical information.

[07:29:01.0000] <yecril71>
But I think it can be.

[07:30:00.0000] <yecril71>
The big question is, what semantic information should be put there.

[07:30:01.0000] <yecril71>
And that depends on the author�s anticipation of what the user agents will be able to do with this information.

[07:34:00.0000] <zcorpan>
does ie do something with <script type="text/xml-script">?

[07:35:00.0000] <yecril71>
I do not think so; "text/xml" is recognized.

[07:35:01.0000] <zcorpan>
what does it do with text/xml?

[07:35:02.0000] <yecril71>
What is "text/xml-script" for?

[07:35:03.0000] <zcorpan>
don't know but it's used all over the place

[07:35:04.0000] <yecril71>
It makes an XML data island, available through

[07:35:05.0000] <zcorpan>
ah

[07:36:00.0000] <yecril71>
document.getElementByID("xmlscript").xml

[07:37:00.0000] <yecril71>
Where is "text/xml-script" used?

[07:38:00.0000] <zcorpan>
http://www.google.com/codesearch?q=text%2Fxml-script&hl=en&btnG=Search+Code

[07:38:01.0000] <yecril71>
(Dangerous bend: an external script of type "text/xml" fails to load

[07:38:02.0000] <yecril71>
if it has an XML PI with encoding)

[07:40:00.0000] <yecril71>
Your search shows that it is not for Internet Explorer,

[07:40:01.0000] <yecril71>
it is for Internet Information Server.

[07:41:00.0000] <zcorpan>
ok

[07:45:00.0000] <Philip`>
<script type=text/xml> and <script language=xml> both do the funky XML thing in IE

[07:45:01.0000] <Philip`>
(unlike (text/)xml-script)

[07:45:02.0000] <Philip`>
at least in IE6

[07:53:00.0000] <yecril71>
except that script[language=xml] means that handlers to events dispatched to the script

[07:53:01.0000] <yecril71>
should be processed by the XML interpreter.

[07:54:00.0000] <yecril71>
I am unable to find any usage for this.

[07:57:00.0000] <Philip`>
I'm not quite sure what you mean

[07:59:00.0000] <Philip`>
I'd assume <script type=text/xml>/<script language=xml> are equivalent in the same way that <script type=text/vbscript>/<script language=vbscript> are equivalent

[08:07:00.0000] <yecril71>
Exactly in the same way.

[08:08:00.0000] <yecril71>
SCRIPT[language=vbscript] means that

[08:08:01.0000] <yecril71>
handlers to events dispatched to the script

[08:08:02.0000] <yecril71>
should be processed by Visual Basic Scripting edition.

[08:09:00.0000] <yecril71>
Example: <script type="text/javascript" language="vbscript" onload="MSGBOX &quot;Loaded&quot;,, &quot;Info&quot;" >

[08:10:00.0000] <yecril71>
Rather funny, and probably unsupported.

[08:16:00.0000] <Philip`>
Ah, that's odd - it looks like <script language> has two totally separate meanings

[08:17:00.0000] <Philip`>
(It sets the language of the loaded script, like what <script type> does; but it also sets the language for on* event handlers, like it does on every other element)

[08:19:00.0000] <Hixie>
language exists on other elements?

[08:20:00.0000] <Philip`>
Yes

[08:20:01.0000] <Philip`>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Cbody%20onload%3D%22alert%20'hello%20world'%22%20language%3D%22vbscript%22%3E

[08:20:02.0000] <Philip`>
says 'hello world' in IE6

[08:20:03.0000] <Philip`>
(and doesn't if you remove the language)

[08:21:00.0000] <Philip`>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Ca%20href%3D%23%20onclick%3D%22alert%20'hello%20world'%22%20language%3Dvbscript%3Etest%3Ca%3E too

[08:21:01.0000] <yecril71>
I think Internet Explorer infers script type from script language.

[08:22:00.0000] <yecril71>
It is an extension since a script without a type does nothing

[08:22:01.0000] <yecril71>
(unless the type is provided by the Web server that serves the script

[08:22:02.0000] <yecril71>
as an external resource)

[08:23:00.0000] <Philip`>
I don't think that's true - it defaults to JScript if nothing is specified

[08:23:01.0000] <Philip`>
(e.g. if you use <script src=foo> and foo is sent as text/plain)

[08:23:02.0000] <yecril71>
Right, taking that back.

[08:24:00.0000] <yecril71>
In this case, it is a nonconforming extension.

[08:26:00.0000] <yecril71>
Philip`, how did you get Visual basic to understand ticks as string delimiters?

[08:28:00.0000] <Philip`>
yecril71: Uh, that's a good question

[08:29:00.0000] <Philip`>
I think the answer is that IE is crazy

[08:29:01.0000] <yecril71>
Sorta.

[08:29:02.0000] <Philip`>
If I write <body onload="alert '1'" language=vbscript><script language=vbscript>alert '2'</script> then it alerts "" and then "1"

[08:30:00.0000] <Philip`>
i.e. in the <script> it treats ' as a comment character, so it's an empty alert call

[08:30:01.0000] <Philip`>
where in the onload attribute, ' is just a quote character

[08:30:02.0000] <Philip`>
s/where/whereas/

[08:31:00.0000] <Philip`>
so the basic syntax of the VBScript language depends on the context in which it's used

[08:31:01.0000] <Philip`>
which is clearly insane

[08:59:00.0000] <Hixie>
vbscript is insane, yes

[12:00:00.0000] <aaronlev>
hsivonen: hi

[12:01:00.0000] <hsivonen>
aaronlev: hi

[12:01:01.0000] <aaronlev>
hsivonen: have you seen silvia's requirements doc for video a11y?

[12:01:02.0000] <hsivonen>
aaronlev: from before TPAC or after?

[12:01:03.0000] <hsivonen>
I have seen one from before

[12:01:04.0000] <hsivonen>
I haven't after

[12:02:00.0000] <aaronlev>
she just sent it out a couple days ago, https://wiki.mozilla.org/Accessibility/Video_a11y_requirements

[12:02:01.0000] <aaronlev>
was wondering what you thought of it

[12:02:02.0000] <hsivonen>
I'll review it.

[12:02:03.0000] <hsivonen>
/me has lost the whole day recovering from HFS+ damage

[12:03:00.0000] <aaronlev>
ouch


2008-11-07
[23:25:00.0000] <jgraham>
So, did I miss anything exciting?

[02:08:00.0000] <hsivonen>
Philip`: I think the "other computer" I should have is a Netgear ReadyNAS

[02:08:01.0000] <hsivonen>
it's expensive, but not as expensive as another computer with RAID

[02:09:00.0000] <hsivonen>
and it's relatively not so expensive considering how much time I lose annually to recovering from HFS+ problems and how many disks I need to buy in order to enable those recoveries

[02:10:00.0000] <Philip`>
Why would the other computer need RAID?

[02:10:01.0000] <Philip`>
since it's just for backup storage, and doesn't need especially high reliability and uptime and everything

[02:10:02.0000] <hsivonen>
Philip`: if the disks aren't in sync all the time, I need to merge file systems manually upon failure like I'm doing now

[02:11:00.0000] <Philip`>
If they're not in sync, you can just lose a few hours of work and it won't be the end of the world :-)

[02:12:00.0000] <hsivonen>
well, I could also consolidate my current external non-backup disks to an NFS RAID

[02:12:01.0000] <hsivonen>
I'm not sure if getting a RAID NAS is crazy, but yesterday I sure regretted that I hadn't bought one a year before

[02:14:00.0000] <hsivonen>
after all, juggling 6 external disks at around 100 euros a pop plus time doesn't make the cheap way so cheap after all

[02:39:00.0000] <Philip`>
/me finds it hard to juggle more than three disks at once, and then he still tends to drop them onto the floor quite often

[02:41:00.0000] <hsivonen>
where does Eclipse on Mac store its indent prefs?

[02:46:00.0000] <Philip`>
Maybe in the workspace project directory's .settings/org.eclipse.jdt.core.prefs file?

[02:46:01.0000] <hsivonen>
/me starts eclipse to see if everything is intact

[02:54:00.0000] <zcorpan>
/me added <input type> values to http://simon.html5.org/html5-elements

[02:56:00.0000] <Philip`>
zcorpan: It might be nice if those lists could be sorted alphabetically

[02:56:01.0000] <Philip`>
(possibly as a user-choosable option, since it's also nice to have the non-alphabetic semi-logical order)

[02:57:00.0000] <zcorpan>
Philip`: right now it's in the spec's order

[02:57:01.0000] <zcorpan>
Philip`: if you write a javascript that sorts the lists i'll be happy to include it :)

[03:02:00.0000] <zcorpan>
/me decides to use the quotation marks in markup instead of in css

[03:02:01.0000] <Philip`>
Hixie: Could you add an HTMLOListElement.sortKey = function (li) { return key for automatic sorting of list items } feature so that I can just wait a few years and then implement html5-elements list sorting for zcorpan with minimal effort?

[03:06:00.0000] <zcorpan>
Philip`: i'm not sure it's less effort -- you'd have to write lots of testcases for such a feature

[03:06:01.0000] <Philip`>
zcorpan: That makes it somebody else's problem

[03:06:02.0000] <Philip`>
since it's too boring a feature for me to write test cases myself

[03:07:00.0000] <zcorpan>
fair enough :P

[03:11:00.0000] <hsivonen>
how's this "DRE" different from DRM?

[03:12:00.0000] <Lachy>
hsivonen, have you considered getting a drobo and droboshare for your NAS?

[03:13:00.0000] <Lachy>
hsivonen, what's DRE?

[03:15:00.0000] <Philip`>
hsivonen: http://lessig.org/blog/2003/04/a_respectful_quibble_with_the.html - "DRE is [...] DRM minus the management. A DRE system simply enables an efficient way for people to say what freedoms they are enabling."

[03:31:00.0000] <Lachy>
"Our CC licenses expressly state that you can't use our technology with a DRM system that does not adequately protect "fair use."" - I didn't know that.

[03:32:00.0000] <Lachy>
but I wonder what that means for Amazon, who's selling a DRM-infected ebook of Lessig's own Free Culture

[03:32:01.0000] <Philip`>
/me wonders how many people who use CC licences ever actually read what they say

[03:33:00.0000] <Philip`>
MIT/BSD licences have the quite significant advantage that you can read them

[03:33:01.0000] <Lachy>
most people would just read the human readable versions

[03:34:00.0000] <Lachy>
I tried reading the legal stuff once, but it's difficult to understand

[03:48:00.0000] <hsivonen>
Lachy: according to www-style EOT is DRE--not DRM

[03:49:00.0000] <hsivonen>
Lachy: I don't know what a drobo is

[03:50:00.0000] <hsivonen>
Lachy: this thing? http://www.drobo.com/Products/BeyondRAID.html

[03:50:01.0000] <hsivonen>
Lachy: it looks like the OS X kernel would run a HFS+ file system in that case. correct?

[03:51:00.0000] <hsivonen>
I don't want a lot of space. I want to get rid of HFS+ and to move file system cache management to a different kernel that doesn't die along with OS X.

[03:52:00.0000] <hsivonen>
Lachy: the "human readable" versions don't mention important things like the TPM ban.

[03:53:00.0000] <hsivonen>
/me wonders if EOT constitutes banned TPM from the CC point of view...

[03:58:00.0000] <zcorpan>
Philip`: sort feature implemented

[03:59:00.0000] <Lachy>
hsivonen, yeah, that's the drobo

[04:01:00.0000] <hsivonen>
Lachy: Its seems that Drobo + DroboShare + disks is more expensive than ReadyNAS NV+. (and even a ReadyNAS Duo might suffice for me)

[04:02:00.0000] <hsivonen>
Lachy: thanks for the pointer

[04:04:00.0000] <Lachy>
The claims that EOT is DRE is bogus, because it has an enforcement mechanism built into it. So it's definitely DRM

[04:04:01.0000] <Lachy>
even if the encryption isn't particularly effective

[04:14:00.0000] <Lachy>
Wow, the arguments in this are rather pathetic. http://lists.w3.org/Archives/Public/www-style/2008Nov/0046.html - His argument for claiming that plain TTF/OT files don't meet the needs of developers is based on wanting to give into the demands of the font foundries

[04:15:00.0000] <Philip`>
I assume there are technical arguments for using the EOT format rather than OpenType (e.g. to allow compression, to make useful metadata easily readable without parsing the whole file, to express digital rights, etc), but has anyone suggested any non-DRM purpose for the XORing?

[04:15:01.0000] <Lachy>
the fact is the, technologically, plain TTF/OT files do meet the needs of developers.

[04:15:02.0000] <Lachy>
s/is the/is that/

[04:19:00.0000] <Philip`>
Lachy: It's not much good being technologically satisfactory if there are other practical (social, legal, etc) problems

[04:20:00.0000] <hsivonen>
how do I detect the existence of os.symlink in Python the right way?

[04:20:01.0000] <Philip`>
like how it's no good for HTML5 to tell UAs that they must support Theora - maybe the technology is fine and it does exactly what developers want, but it's never actually going to work

[04:22:00.0000] <Philip`>
hsivonen: Detecting whether the method exists in the os module or not?

[04:22:01.0000] <hsivonen>
btw, did the open source release of Android include an H.264 decoder? was there any legal stuff about its patent status and Apache License 2.0?

[04:22:02.0000] <Philip`>
If so: 'symlink' in dir(os)

[04:22:03.0000] <hsivonen>
Philip`: right

[04:22:04.0000] <hsivonen>
Philip`: thanks

[04:22:05.0000] <Philip`>
or try: os.symlink(...); expect AttributeError: ...

[04:29:00.0000] <hsivonen>
aaargh. is the shutil module Windows-incompatible?

[04:30:00.0000] <Philip`>
It shouldn't be

[04:30:01.0000] <Philip`>
Is there some particular part that's a problem?

[04:32:00.0000] <hsivonen>
http://pastebin.mozilla.org/570525

[04:33:00.0000] <hsivonen>
looks like the /*.* bit comes from shutil

[04:33:01.0000] <hsivonen>
or is ..\\ the offending part?

[04:34:00.0000] <Lachy>
hsivonen, is the h.264 decoder in Android covered by the Apache licence too?

[04:35:00.0000] <hsivonen>
Lachy: I don't know

[04:35:01.0000] <Philip`>
hsivonen: What is the src passed to shutil.copytree?

[04:36:00.0000] <hsivonen>
os.path.join("..", "syntax"), os.path.join(buildRoot, "local-entities", "syntax")

[04:36:01.0000] <hsivonen>
so src is os.path.join("..", "syntax")

[04:38:00.0000] <hsivonen>
returns "..\\syntax" on Windows and "../syntax" on OS X

[04:40:00.0000] <Philip`>
Looks like os.listdir is adding the /*.*

[04:40:01.0000] <Philip`>
and it returns that error message if ..\syntax doesn't exist

[04:41:00.0000] <Philip`>
(It does seem to handle both \ and / correctly)

[04:42:00.0000] <hsivonen>
aaah, right. In the non-symlink case, I need to resolve the first path against the second

[04:42:01.0000] <Philip`>
If ..\syntax doesn't exist, that could be the problem :-)

[04:43:00.0000] <hsivonen>
yeah :-)

[04:43:01.0000] <Philip`>
(Do you have to have two separate code paths, rather than doing the non-symlink thing on every platform?)

[04:44:00.0000] <hsivonen>
I'm probably going to remove the whole thing in due course. I just want to unblock other people working on Windows ASAP.

[04:44:01.0000] <hsivonen>
also, I don't want to copy stuff on non-Windows

[05:14:00.0000] <jgraham>
FWIW I think doing is foo in dir(bar) is explicitly bad practice

[05:14:01.0000] <jgraham>
s/is/if/

[05:18:00.0000] <zcorpan>
Hixie: "URL" in the <input> summary table should point to #url-state instead of #url

[05:23:00.0000] <hsivonen>
jgraham: what's the right way?

[05:27:00.0000] <jgraham>
hsivonen: I think using try: except: or hasattr(os, "symlink") is the right way 

[05:29:00.0000] <hsivonen>
jgraham: thnks

[05:29:01.0000] <hsivonen>
now I've dealt with the symlink issue

[05:30:00.0000] <hsivonen>
next it appears that javac behaves differently on Windows

[05:30:01.0000] <hsivonen>
sigh

[05:33:00.0000] <hsivonen>
Hmm. if users can deal with an English or Castellano browser UI, why is the file upload button so bad?

[05:34:00.0000] <zcorpan>
hsivonen: because it's in the rendering viewport and hence the author's responsibility to get right

[05:35:00.0000] <zcorpan>
users mostly just interact with what's in the rendering viewport and the back button

[05:35:01.0000] <zcorpan>
and the close button

[05:35:02.0000] <hsivonen>
fair enough

[05:49:00.0000] <Philip`>
/me forgot about hasattr

[06:23:00.0000] <hsivonen>
aargh. the windows quoting bug strikes again

[06:23:01.0000] <Lachy>
what's the windows quoting bug?

[06:24:00.0000] <hsivonen>
Windows does crazy things with quoted string in the argument to Python os.system()

[06:50:00.0000] <hsivonen>
what ever happened to the historical posix compliance of Windows NT...

[06:51:00.0000] <hsivonen>
it's sad how many crazy differences and bugs there are compared to unix platforms

[06:51:01.0000] <jcranmer>
you need to have the Unix subsystem installed

[06:56:00.0000] <hsivonen>
I'd prefer it to behave in a normal way by default

[06:58:00.0000] <Philip`>
Windows' POSIX system is kind of parallel to its Win32 subsystem, and if you use one then you can't use the other, and most people want to use Win32

[06:58:01.0000] <jcranmer>
Microsoft wouldn't :-)

[06:58:02.0000] <Philip`>
(or at least that's what I think I've heard)

[07:03:00.0000] <Philip`>
/me recently heard people complaining how Windows' directory change notification system sometimes gives you the long filename when a file is created and the short (8.3) name when it is deleted, and once it's deleted it's too late for you to do the short->long name conversion to work out which file it was so it's impossible to tell

[07:04:00.0000] <Philip`>
so it's constantly irritating trying to write code and finding all these weird cases that are there for backward compatbility :-(

[07:04:01.0000] <jcranmer>
short names must die

[07:05:00.0000] <jcranmer>
most of Windows' eccentricities must die, in fact

[07:06:00.0000] <Philip`>
(Then again, the FAM directory change notification system on Linux is (seemingly fundamentally and unavoidably) prone to deadlocks, so that's no fun either)

[07:15:00.0000] <Hixie>
zcorpan: as far as i can tell, it does

[07:18:00.0000] <hsivonen>
why is it that people assum commercial foundries wouldn't sell fonts for Web use if there was an unserved market there ready to use someone else's fonts as an alternative?

[07:20:00.0000] <Philip`>
Probably because they talked to commercial foundries, who said "we won't sell our fonts for Web use"

[07:22:00.0000] <hsivonen>
I guess that leaves a business opportunity for new foundries

[07:22:01.0000] <Lachy>
I wonder why font foundries consider selling fonts for web use any different from existing companies that sell stock photos for web use

[07:23:00.0000] <Hixie>
why is it that people care, is what i really want to know

[07:23:01.0000] <Lachy>
because we'll be stuck with DRM encumbered crap like EOT

[07:24:00.0000] <hsivonen>
Lachy: the explanation I've heard is that photos are usually used in a different way from fonts and synthetizer samples

[07:25:00.0000] <hsivonen>
Hixie: care about what?

[07:26:00.0000] <Philip`>
Lachy: If you copy a stock photo I've used on my site, you're not getting much more value from it than I am; but if you copy a font I've used, you can use it for lots of new text or for a giant logo, so it's much more useful to you, which makes it a different situation

[07:27:00.0000] <Philip`>
(E.g. I might pay some amount for the photo/font that's proportionate to the value I'm getting from it, but you might copy the font and get a lot more value and the poor font company won't have received any compensation for that at all)

[07:30:00.0000] <Philip`>
Also, stock photos will usually be resized and recoloured and shaded and shaped and munged to fit the site's design, which prevents anyone from copying the original version that was sold

[07:31:00.0000] <hsivonen>
I wonder if there are video producers who approve of .flv in Flash because a random Windows user can't play .flv but disapproves of .mp4 in Flash because it plays with QuickTime

[07:33:00.0000] <Philip`>
There seem to be people who disapprove of both because you can save them to disk and play them in reasonably common tools like VLC (I guess?), and instead use Flash's proprietary not-yet-reverse-engineered streaming protocol instead

[07:36:00.0000] <Hixie>
hsivonen: care about supporting the font foundries as if they were charity cases

[07:36:01.0000] <Philip`>
/me wonders why people haven't complained much about the BBC iPlayer displaying unfair favouritism towards the iPhone, by providing programmes as standard downloadable MP4 to it but not to any other MP4-capable device

[07:41:00.0000] <takkaria>
Philip`: but the BBC know that there are tools available to get the MP4 files, and they're not doing anything to stop them

[07:41:01.0000] <Lachy>
Philip`, people complain about the iPlayer restrictions all the time

[07:42:00.0000] <jcranmer>
"Foundries will adapt, just like music labels have adapted,"

[07:42:01.0000] <jcranmer>
I don't think the music labels have quite adapted yet

[07:42:02.0000] <jcranmer>
they're adapt/ing/, but adapt/ed/...

[07:43:00.0000] <Lachy>
jcranmer, they almost have. Hopefully by next year, we'll see more stores worldwide with DRM free music

[07:44:00.0000] <Hixie>
jcranmer: the only remaining bastian of DRM is iTunes, and that has more to do with hurting apple than the usual reasons for DRM

[07:44:01.0000] <jcranmer>
Lachy: contact me again when the RIAA stops randomly selecting people to sue for copyright infringement

[07:44:02.0000] <Hixie>
that has nothing to do with DRM

[07:44:03.0000] <jcranmer>
it deals with piracy; separated by at most one degree of freedom

[07:45:00.0000] <jcranmer>
maybe I should just shut up

[07:45:01.0000] <Lachy>
the RIAA's sue-their-customers strategy is a seaparate problem

[07:45:02.0000] <Hixie>
i have problems with copyright too, but those are unrelated to the problems with DRM

[07:46:00.0000] <Hixie>
DRM could exist even if there was no copyright

[07:46:01.0000] <Hixie>
and would be evil

[07:46:02.0000] <Hixie>
copyright isn't evil, it's just a societal decision that i happen to disagree with

[07:47:00.0000] <jcranmer>
I would say that copyright is A Good Thing™ if done right, but the present incarnation is not an example of `doing right'

[07:48:00.0000] <jcranmer>
although patents are much more broken than copyrights in this day and age

[07:49:00.0000] <Lachy>
the problems with DRM are made worse by horrible copyright laws, like the DMCA and its equivalents in other countries, which impose severe penalties for circumvention

[07:50:00.0000] <hsivonen>
It sucks that being able to buy an iPlayer subscription is limited to people living on a particular set of islands

[07:50:01.0000] <takkaria>
you can get a subscription?  I did not know that

[07:50:02.0000] <jcranmer>
IIRC, doesn't the EU have less stringent laws on the matter than the US?

[07:51:00.0000] <hsivonen>
takkaria: well pay for the BBC fee/tax/whatever-it's-called

[07:52:00.0000] <Lachy>
hmm, this discussion reminds me that I really need to finish and post my blog entry about all these restrictions

[07:52:01.0000] <hsivonen>
I had wished that the French culture proponents in France had realized that they could spread their stuff more effectively than the Americans spread theirs if the French didn't do DRM

[07:53:00.0000] <hsivonen>
but, sadly, the EU is as crazy as ever when it comes to DRM and copyright

[07:55:00.0000] <Lachy>
that's because the governements are increasingly financed and influenced by rich organisations wanting to push their own agenda

[08:01:00.0000] <jcranmer>
... sorry I brought up this OT discussion

[08:01:01.0000] <jcranmer>
/me goes and sits in the corner

[08:24:00.0000] <Philip`>
takkaria: They did a fair amount to stop those tools at first - it became more than just a token effort of checking the UA string

[08:25:00.0000] <Philip`>
and they only seem to have stopped once they reached the point where it was no longer possible to distinguish an iPhone UA from a custom-made Ruby/etc UA

[08:26:00.0000] <Philip`>
hsivonen: You don't have to pay the licence fee to (legally) use the iPlayer - the licence fee is only required for live broadcasts

[08:29:00.0000] <Philip`>
(though I'm not quite sure how it legally works in the cases where the BBC has live streaming video (e.g. of important news) on their web site)

[12:55:00.0000] <phroggy>
so there was a big discussion on the mailing list about whether classes have semantic meaning, and it was pointed out that the original purpose of the class attribute was exactly that.  But I just don't get it.  The only meaning of classes is 1) in the developer's own mind, and 2) presentational.  What good are semantics if the semantic meaning isn't really accessible?

[12:57:00.0000] <phroggy>
what I mean by that is, in order to ascribe semantic meaning to a particular class, it takes an intelligent human looking at the source code and trying to understand the intentions of the author.  People talk about how semantic markup allows e.g. search engines to better understand the content, but classes don't help with that at all, because their interpretation can't be automated.

[12:57:01.0000] <phroggy>
the only automated thing you can do with classes is apply presentational styles as defined by CSS.

[12:58:00.0000] <phroggy>
comments?

[13:28:00.0000] <Lachy>
phroggy, classes are also useful for scripting and for use in situations where there is some agreement between the author and consumer about the meaning of the class names

[13:29:00.0000] <Lachy>
microformats effectively provide such an agreement for a set a class names

[13:30:00.0000] <mpt>
The original purpose of the <div> element was semantic, too

[13:32:00.0000] <Lachy>
it is a little semantic, it just doesn't have much meaning beyond conveying a little structural information

[13:42:00.0000] <phroggy>
I confess I've failed to read up on microformats.  What's the basic idea there?

[13:46:00.0000] <Lachy>
phroggy, http://microformats.org/get-started/

[13:46:01.0000] <Philip`>
phroggy: The idea is that I can write <span class="vcard"><span class="fn">Mr Blobby</span></span> and then people will be able to follow the microformat parsing rules to determine that that's a name

[13:46:02.0000] <phroggy>
hmmmm, interesting.

[15:08:00.0000] <justben>
Minor issue with html5lib CSS sanitizer. Anyone here who can accept a patch?

[15:12:00.0000] <justben>
Well, assuming someone will see this sooner or later in logs. One of the regexes backtracks horribly if it's passed a long css string.

[15:12:01.0000] <justben>
Sample at http://ben.lateralfricative.net/patches/html5lib/css-sanitizer-re-tweak/evil-table.html.gz

[15:12:02.0000] <smedero>
hrm, well probably jgraham or Philip` I suppose... but I imagine you could just create a new issue on the google project and attach the patch. http://code.google.com/p/html5lib/issues/entry

[15:13:00.0000] <justben>
Alas, that appears to require a google account, and my brain is about full up on web accounts right now.

[15:13:01.0000] <justben>
Anyway, I've got a tiny patch at http://ben.lateralfricative.net/patches/html5lib/css-sanitizer-re-tweak/css-sanitizer-re-tweak.patch.gz that eliminates some backtracking.

[15:14:00.0000] <justben>
On my computer it reduces the processing for evil-table.html from ~45mins to ~2secs.

[15:15:00.0000] <justben>
I don't hang around here typically, but I'm xmpp: and mailto:ben⊙ln if anyone wants to talk about it.

[15:27:00.0000] <justben>
Cheers!


2008-11-08
[00:48:00.0000] <takkaria>
Philip`: live streaming video means you need a licence.  trufact

[03:26:00.0000] <Philip`>
takkaria: Well, obviously not in general, but I'm thinking of the live streaming BBC channels

[03:27:00.0000] <Philip`>
since that doesn't seem much different to a PC with a TV tuner card plugged into an aerial, for which you do need a licence

[03:29:00.0000] <Lachy>
Philip`, since when do you need a licence to to have a tuner card in a PC?

[03:29:01.0000] <Lachy>
or, in which country?

[03:32:00.0000] <Philip`>
Lachy: Since probably 1946

[03:32:01.0000] <Philip`>
in the UK

[03:33:00.0000] <Philip`>
http://www.tvlicensing.co.uk/information/index.jsp#link1 - "You need a TV Licence to use any television receiving equipment such as a TV set, digital box, DVD or video recorder, PC, laptop or mobile phone to watch or record television programmes as they're being shown on TV."

[03:44:00.0000] <Lachy>
oh, wow

[03:44:01.0000] <Lachy>
There are no such licences in Australia

[03:46:00.0000] <Lachy>
woah! So you pay £139.50 each year just for a TV?!

[03:47:00.0000] <Lachy>
is that per-TV or per household?

[03:48:00.0000] <Lachy>
how do they enforce it, since it wouldn't be possible to know whether or not you have a TV unless they inspected your house

[03:49:00.0000] <Philip`>
Per household

[03:50:00.0000] <Philip`>
They're not being too unreasonable - you get a discount if you have a black-and-white TV or are blind

[03:50:01.0000] <Lachy>
I noticed the discount for B&W, but surely everyone has colour these days

[03:51:00.0000] <Philip`>
In my experience, they enforce it by sending increasingly threatening letters telling you that it's a criminal offence to watch TV without a licence, and that's about it

[03:51:01.0000] <Philip`>
Apparently they do sometimes send people to check, but you can just tell them to go away and not let them in

[03:52:00.0000] <Philip`>
but I think they can then get a warrant or something, to allow them to search for illicit TVs

[03:53:00.0000] <koredn>
It is the same in germany - only it is about 400€ / year

[03:54:00.0000] <Philip`>
Hmm, http://en.wikipedia.org/wiki/TV_licence#Germany says €204.36/year

[03:55:00.0000] <koredn>
OK that depends on some factors, like if you additionally have radio and so on...

[03:55:01.0000] <Philip`>
Ah, okay

[03:57:00.0000] <koredn>
Looks like I imagined the exact amount of money wrong...

[04:03:00.0000] <Philip`>
Do people complain a lot about the licence fee in Germany?

[04:12:00.0000] <Lachy>
So our taxes pay about $0.08 per day for the ABC, which works out about $30 per year, which is reasonable. But even then, I think that's only for people earning over about $50,000 per year. http://en.wikipedia.org/wiki/TV_licence#Australia

[04:19:00.0000] <Philip`>
Is ABC's output good?

[04:22:00.0000] <Lachy>
yeah, it's pretty good for some people.

[04:22:01.0000] <Lachy>
They have a lot of programs aimed at kids or at the older generation (over 40s). I don't really like many of their programs myself

[06:29:00.0000] <ehird>
http://www.bluishcoder.co.nz/2008/11/video-audio-and-cross-domain-usage.html Umm, what's next, doing this for <img> too?

[06:35:00.0000] <Philip`>
ehird: No, because that would break backward compatibility to an unacceptable extent

[06:35:01.0000] <ehird>
It doesn't exactly, you know, make sense.

[06:36:00.0000] <Philip`>
so instead we have more complex security rules for <img> (e.g. if it's cross-domain then you can't draw it to a <canvas> then read pixels back)

[06:36:01.0000] <ehird>
That's...really ridiculous.

[06:37:00.0000] <Philip`>
See this channel's topic :-)

[06:38:00.0000] <Philip`>
http://lists.xiph.org/pipermail/theora/2008-November/001958.html seems like a good description of the various relevant issues

[06:41:00.0000] <ehird>
Philip`: This'll screw up that amazon thing whose name I forget.

[06:41:01.0000] <Philip`>
S3?

[06:41:02.0000] <Philip`>
or CDNs is general, where you get someone else to host all your big media files?

[06:41:03.0000] <ehird>
Yeah.

[06:41:04.0000] <Philip`>
*in general

[06:41:05.0000] <ehird>
It'll be years upon years before they add the access control stuff, I bet.

[06:42:00.0000] <ehird>
so it'll slow adoption of <video> &c

[06:42:01.0000] <Philip`>
Access Control isn't hard to add - it's just a few HTTP headers

[06:42:02.0000] <Philip`>
and if their customers demand it, I don't see why they wouldn't be happy to add it quickly

[06:42:03.0000] <ehird>
Philip`: will their customers demand it? More likely they won't just use <video>, IMO...

[06:46:00.0000] <Philip`>
Amazon S3 apparently already lets you set custom HTTP headers for your files, so I think you could just add a "Access-Control-Allow-Origin: *" and it'd be fine

[06:48:00.0000] <ehird>
Oh, okay.

[06:48:01.0000] <Philip`>
Oh, actually, I might be wrong

[06:48:02.0000] <ehird>
Oh, okay.

[06:49:00.0000] <Philip`>
so I don't know really

[06:49:01.0000] <Philip`>
but if it's not possible already, it doesn't seem like the kind of change that would take them years to implement if you asked them nicely

[06:50:00.0000] <Philip`>
and once one person has asked, everyone else will be able to use that feature too

[06:50:01.0000] <Philip`>
and then they'll expect all the competing services to offer that feature too

[06:50:02.0000] <ehird>
mm

[12:11:00.0000] <gsnedders>
/me stretches

[13:07:00.0000] <takkaria>
Philip`: live streaming bbc stuff of any kind requires a tv licence

[13:10:00.0000] <Philip`>
takkaria: It's strange that they never advertise or warn about that with the streaming stuff on their web site - I guess they assume it's too minor a problem to be concerned with, since most people will have real TVs anyway

[13:12:00.0000] <Philip`>
/me presumes nobody would really be fined £1000 for visiting the BBC news web site and clicking 'play' on the video when they don't have a TV licence

[13:15:00.0000] <takkaria>
I saw a warning up on there once about it

[13:17:00.0000] <Dashiva>
What about us dirty foreigners?

[13:19:00.0000] <takkaria>
you probably get warned about something too

[13:20:00.0000] <takkaria>
dunno, never been a foreigner

[13:20:01.0000] <Philip`>
Oh, I've never noticed a warning

[13:20:02.0000] <Philip`>
Foreigners already get a different version of the news site, with adverts

[13:20:03.0000] <Philip`>
so I guess they remove the live streaming video too

[13:20:04.0000] <jmb>
they do

[13:20:05.0000] <Philip`>
(based on IP)

[13:20:06.0000] <jmb>
it looks truly odd :)

[13:21:00.0000] <Lachy>
what about us dirty foreigners with ad blockers?

[13:21:01.0000] <Philip`>
(What they need is for ISPs to provide an IP -> postal address lookup service, which can be correlated with the TV licensing database, to determine whether the user should be allowed to see the video or not)

[13:21:02.0000] <jmb>
cool. then they can lose that!

[13:22:00.0000] <Philip`>
Lachy: You're just leeches on British society

[13:23:00.0000] <Lachy>
actually, other than just now, I can't remember the last time I even looked at the BBC website

[13:24:00.0000] <Lachy>
although i suppose I have downloaded a number of documentaries


2008-11-09
[00:40:00.0000] <hsivonen>
/me sees Ahem.ttf on android.git.kernel.org. Does Android ship with Ahem?

[04:09:00.0000] <jgraham>
hsivonen: If the doctype name is folded, is there any constraint on which way it should be folded? Fold-o-lowercase makes the most sense to me but isn't what Gecko does at the moment.

[04:09:01.0000] <jgraham>
/me is catching up with some of the email from the last month

[04:13:00.0000] <hsivonen>
jgraham: lower case seems more elegant and more sane when mapping to XML 1.0 but browser-wise, I'd just want to get an interned string

[04:13:01.0000] <jgraham>
hsivonen: OK. 

[04:40:00.0000] <hsivonen>
heh. Geolocation is now seen as part of "html5" on twitter

[10:04:00.0000] <yecril71>
Microsoft Internet Explorer treats all unknown elements as empty.

[10:05:00.0000] <yecril71>
Only custom elements with namespace prefixes can contain other elements in the DOM.

[10:06:00.0000] <Philip`>
That's sort of true, except if you do document.createElement('foo') then it'll start parsing <foo> as non-empty

[10:07:00.0000] <Philip`>
i.e. <foo>a</foo><script>document.createElement('foo')</script><foo>b</foo> will produce an empty FOO element, then 'a', then an empty /FOO element, then a script element, then a foo element that contains 'b'

[10:07:01.0000] <yecril71>
That requires script support and does not work for restricted sites.

[10:07:02.0000] <Philip`>
Why would it be different in restricted sites?

[10:08:00.0000] <yecril71>
Because restricted sites do not allow scripts to run.

[10:08:01.0000] <Philip`>
Ah

[10:08:02.0000] <yecril71>
I know because I had to restrict several for my own use.

[10:09:00.0000] <yecril71>
Anyway, if the author wants arbitrary elements, he can use XML+CSS

[10:09:01.0000] <yecril71>
or XML+XSLT

[10:10:00.0000] <Philip`>
or a different browser

[10:10:01.0000] <yecril71>
There is no need to push them into HTML.

[10:10:02.0000] <yecril71>
The author surely can use a different browser.

[10:10:03.0000] <yecril71>
The problem is the viewer cannot.

[10:12:00.0000] <Philip`>
/me has to go away for a while

[10:13:00.0000] <Dashiva>
The combination of restricted site and restricted browser sounds like intranet

[10:16:00.0000] <yecril71>
Restricting sites is useful at home as well.

[10:17:00.0000] <yecril71>
Nobody likes the browser to hang in the middle of something.

[10:23:00.0000] <Dashiva>
There's no restriction on browsers at home, though.

[10:25:00.0000] <yecril71>
Internet Explorer has restricted zones.

[10:25:01.0000] <yecril71>
A site that belongs to a restricted zone is restricted.

[10:26:00.0000] <yecril71>
The decision on whether a site should be restricted is on the home user.

[10:41:00.0000] <Philip`>
If a user configures their browser so that some web sites stop working, that's the user's problem and it shouldn't be the authors' responsibility to cope with those atypical restrictions

[10:42:00.0000] <Philip`>
Is it just me, or is "Re: [whatwg] SPOOFED: Re: SPOOFED: Re: ---" a pretty rubbish subject line for an email thread?

[10:50:00.0000] <Lachy>
Philip`, the original message with the subject "[whatwg] ---" was even more rubbish

[10:53:00.0000] <Philip`>
Lachy: I disagree - I think the longer subject exhibits more of what kids nowadays would call "epic fail" than the original

[10:54:00.0000] <yecril71>
I think placing sites into the restricted zone is a typical setting.

[10:54:01.0000] <yecril71>
It is better to have them restricted than to have them break the browser.

[10:54:02.0000] <hsivonen>
hmm. Opera's "always standards mode" really means "map quirks to standards but keep almost standards as almost standards"

[10:55:00.0000] <Lachy>
the longer one is a result of poor spam filtering software and a lack of effort to manually clean it up, but the shorter one is a result of consiously using a bogus subject line

[10:57:00.0000] <yecril71>
The author�s responsibility is to publish valid content.

[10:58:00.0000] <yecril71>
Content that uses custom elements should remain invalid.

[11:00:00.0000] <yecril71>
"SPOOFED" means the presence of Envelope-Sender in this case.

[11:00:01.0000] <Philip`>
Lachy: So the shorter one is merely the result of consciously using a bogus subject line, whereas the longer one is a result of that plus poor spam filtering software plus another four authors who didn't bother fixing the subject and is therefore more rubbish

[11:01:00.0000] <yecril71>
It means WHATWG is posing as the author.

[11:02:00.0000] <Philip`>
That's because it's a pretty standard mailing list, so it's silly for a mail client to think that's a problem

[11:03:00.0000] <yecril71>
The world is a place of all sorts of weird things.

[11:03:01.0000] <Philip`>
That's what makes it interesting enough to bother living in :-)


2008-11-10
[04:12:00.0000] <zcorpan>
Hixie: look at the next table, specifically the table header that contains "Text, E-mail, URL" (re #url vs #url-state)

[05:10:00.0000] <zcorpan>
hmm, <!doctype html><title></title><noscript><style/></style></noscript> is valid html5 (with script)

[05:11:00.0000] <zcorpan>
so is <!doctype html><title></title><noscript><style></noscript>

[05:18:00.0000] <hesslow>
I want to use links like spotify:search:test but I want some kind of fallback for users that don't have the spotify-protocol registered. Is there anyway to do that? I found the type-attributes for a-tags byt that just gives the browser a hint and there is no way to specify how to fallback. What I want to do is to load a normal http-url which have more information about the spotify-url. Is there anyway to do what I want?

[05:20:00.0000] <tthorsen>
zcorpan: Either would at least cause a few parse errors, wouldn't they?

[05:22:00.0000] <hsivonen>
b

[05:23:00.0000] <zcorpan>
tthorsen: not when scripting is enabled. (the fragment parsing algorithm would then generate parse errors but the spec doesn't say that's an error for the purposes of the content model)

[05:24:00.0000] <tthorsen>
right. I ran them through the html5lib parser, and it agrees that they're alright.

[05:26:00.0000] <zcorpan>
filed http://www.w3.org/Bugs/Public/show_bug.cgi?id=6213

[05:42:00.0000] <hesslow>
The reason I ask here is that I don't think there is any way so my follow up question would be shouldn't there be a way to fallback.

[05:51:00.0000] <hsivonen>
zcorpan: btw, Validator.nu now builds on Windows (but crashes on startup for a yet unknown reason)

[05:55:00.0000] <zcorpan>
hsivonen: cool

[05:56:00.0000] <hsivonen>
(requires the JAVA_HOME env variable to be set)

[06:38:00.0000] <hsivonen>
can anyone recommend a free simple text editor for windows that uses UTF-8 and LF line breaks?

[06:43:00.0000] <zcorpan>
i use notepad2

[06:44:00.0000] <zcorpan>
though it doesn't handle large files too well

[06:44:01.0000] <csarven>
hsivonen PSPad

[06:46:00.0000] <hsivonen>
zcorpan: thanks

[06:47:00.0000] <hsivonen>
csarven: how is PSPad different from notepad2?

[06:47:01.0000] <hsivonen>
PSPad has a softpedia badge, which is not a good sign

[06:48:00.0000] <zcorpan>
pspad seems to have lots of features that notepad2 doesn't, like ftp client

[06:53:00.0000] <Philip`>
I've heard lots of people say positive things about Notepad++

[06:56:00.0000] <hsivonen>
aaargh. the symlink thing is still the problem

[06:58:00.0000] <zcorpan>
hmm, insertRow() without arguments

[06:59:00.0000] <zcorpan>
acts like insertRow(0) in webkit and opera, acts like insertRow(rows.length) in ie, throws in firefox

[07:23:00.0000] <hsivonen>
an incredible amount of trouble flowing out of using symlinks in the first place and Windows not having them

[07:24:00.0000] <hsivonen>
aaargh. if I use shutil.copytree to copy a tree that contains .svn directories, why isn't shutil.rmtree allowed to remove it?

[07:25:00.0000] <hsivonen>
all-wcprops is access denied

[07:25:01.0000] <hsivonen>
why oh why

[08:19:00.0000] <Philip`>
hsivonen: Is all-wcprops set as readonly?

[08:20:00.0000] <Philip`>
I think it's either that or hidden (or both)

[08:20:01.0000] <Philip`>
but I'm not entirely sure why SVN would choose to do that

[08:23:00.0000] <hsivonen>
Philip`: yes, it has been marked read-only

[08:23:01.0000] <hsivonen>
how am I supposed to delete something I checked out from svn?

[09:09:00.0000] <Philip`>
hsivonen: By clicking on the folder icon then pressing the delete key then selecting "yes I really want to delete all these files"

[09:19:00.0000] <hsivonen>
Philip`: is that gesture meant to be different from selecting Delete from the context menu?

[09:21:00.0000] <Philip`>
hsivonen: No - it's just meant to be different to using shutil.rmtree

[09:22:00.0000] <hsivonen>
Philip`: ok. interestingly, for me it did behave differently from the context menu

[09:22:01.0000] <hsivonen>
(context menu failed, delete key worked)

[09:22:02.0000] <hsivonen>
same confirmations and progress UI

[09:22:03.0000] <hsivonen>
Windows <3

[09:23:00.0000] <hsivonen>
Philip`: thanks

[09:45:00.0000] <Philip`>
hsivonen: Oh, odd

[12:18:00.0000] <Lachy>
/me waves from New York :-)

[12:19:00.0000] <Lachy>
I have to site in the airport for anouther hour before flying to Vegas

[12:19:01.0000] <Lachy>
s/site/sit/

[12:22:00.0000] <Hixie>
i'm at JFK terminal 3

[12:22:01.0000] <Lachy>
I'm at Newark airport

[12:22:02.0000] <Hixie>
heh

[12:23:00.0000] <Lachy>
where are you flying to?

[12:23:01.0000] <Hixie>
home

[12:24:00.0000] <Lachy>
oh. Where have you been?

[12:25:00.0000] <Hixie>
new york state, mostly

[12:25:01.0000] <Hixie>
"vacation"

[12:29:00.0000] <Lachy>
this is absurd http://www.techdirt.com/articles/20081107/0118162765.shtml

[12:38:00.0000] <Hixie>
that's the least of haliburton's transgressions

[12:40:00.0000] <Philip`>
I'm glad they presented that process as a flowchart - I would have found it difficult to follow the dependencies and ordering between the steps if they had simply written it as a list

[12:41:00.0000] <Dashiva>
Philip`: Maybe someone patented patent applications using lists

[12:42:00.0000] <Philip`>
I hope nobody patents the idea of criticising absurd patents

[12:42:01.0000] <Lachy>
hmm. It seems the "This idea was invented by Shampoo" thing is turning into a new slashdot meme.


2008-11-11
[17:58:00.0000] <jcranmer>
hmm, still asks for password

[00:17:00.0000] <hsivonen>
Validator.nu now builds on Windows but running it requires deleting the HTML 4 / XHTML 1.0 -related lines from validator/presets.txt

[00:53:00.0000] <hsivonen>
the EOT thread took an interesting turn...

[01:39:00.0000] <zcorpan>
silli's both doctypes trigger quirks in O9.6

[03:45:00.0000] <zcorpan>
tthorsen: <select>TITLE</select> is what you get in ie

[03:58:00.0000] <tthorsen>
really? I hadn't tested IE

[03:58:01.0000] <tthorsen>
So that's consistent with the html 5 specification, but does it make sense?

[04:01:00.0000] <takkaria>
I imagine very few sites rely on its behaviour, so it doesn't matter so much what happens to it, I guess

[04:01:01.0000] <hsivonen>
tthorsen: compat is more important than making sense

[04:02:00.0000] <zcorpan>
tthorsen: see /topic :)

[04:03:00.0000] <tthorsen>
of course, but in this case everyone does it differently. Isn't that an opportunity to pick the behaviour that makes the most sense?

[04:03:01.0000] <takkaria>
tthorsen: fwiw, your latest problem with the parsing algorithm is fallout from splitting CDATA/RCDATA processing into its own insertion mode

[04:04:00.0000] <hsivonen>
tthorsen: of the behaviors available, I prefer just doing the IE thing here

[04:04:01.0000] <hsivonen>
tthorsen: with my validator developer hat on, I want to avoid non-streamable recovery where possible given compat

[04:05:00.0000] <zcorpan>
hsivonen: ignoring the characters is certainly streamable

[04:05:01.0000] <hsivonen>
zcorpan: right. but I don't want to move stuff to head any more than absolutely necessary

[04:06:00.0000] <zcorpan>
hsivonen: ah

[04:06:01.0000] <zcorpan>
i thought you were talking about select

[04:06:02.0000] <hsivonen>
/me thought moving to head was one of the options discussed

[04:06:03.0000] <hsivonen>
I guess I'm not properly in sync with what's being discussed

[04:07:00.0000] <tthorsen>
I only meant to suggest ignoring any cdata in select unless the current node is an option element.

[04:08:00.0000] <zcorpan>
which matches opera and mostly matches firefox

[04:09:00.0000] <takkaria>
zcorpan: are Opera implementing HTML5 parsing, do you know

[04:09:01.0000] <jgraham>
I would be happy with that fwiw

[04:09:02.0000] <takkaria>
+?

[04:10:00.0000] <zcorpan>
takkaria: not yet, alghough we are aligning our current parser

[04:10:01.0000] <zcorpan>
although

[04:10:02.0000] <zcorpan>
hmm.. <select><style> and <select><script>

[04:11:00.0000] <zcorpan>
or at least the latter

[04:11:01.0000] <jgraham>
hsivonen: How far have you got in making the validator.nu parser talk to gecko?

[04:11:02.0000] <zcorpan>
the script is inserted in all top-4 browsers

[04:13:00.0000] <hsivonen>
jgraham: the parser core translates to C++ and compiles against NSPR/XPCOM types. It doesn't talk to Gecko yet. Right now, I'm working on the entity table generator.

[04:13:01.0000] <zcorpan>
<style> works in webkit and firefox

[04:13:02.0000] <hsivonen>
jgraham: So no progress since the start of TPAC

[04:15:00.0000] <jgraham>
hsivonen: Sounds cool. I didn't know what you'd done before TPAC either

[04:15:01.0000] <hsivonen>
I made a detour into Jing development

[04:16:00.0000] <hsivonen>
James Clark made a comeback since a 5-year hiatus, so I figured I should get my changes merged upsteam now that the project is under active development

[04:16:01.0000] <malware>
hsivonen: cool to hear that you now have commit access to the upstream Jing sources

[04:17:00.0000] <hsivonen>
malware: yeah. Right now there's a validator-nu branch. the trunk isn't suitable for V.nu yet

[04:18:00.0000] <malware>
I see. well, regardless, I think this may be the first time in history that James has actually enabled collaborative development on any code he's currently working on

[04:30:00.0000] <malware>
re: messages from Tommy Thorsen indicating he's doing active development on a browser engine -- I assume for Kvaleberg -- makes me wonder how high they are aiming and what the business case is for building the engine in-house instead of just using WebKit

[04:30:01.0000] <hsivonen>
/me notes that thanks to all the Java eclipse goodness, it's nicer to write in Java and translate to C++ than to write in C++.

[04:30:02.0000] <zcorpan>
wow, tfoot, th etc break out of <select> in opera

[04:31:00.0000] <zcorpan>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Cbody%3E%0D%0A%3Cscript%3E%0D%0Avar%20elms%20%3D%20%22a%20abbr%20acronym%20address%20area%20b%20base%20basefont%20bdo%20bgsound%20big%20blink%20blockquote%20body%20br%20button%20caption%20center%20cite%20code%20col%20colgroup%20comment%20dd%20del%20dfn%20dir%20div%20dl%20dt%20em%20embed%20fieldset%20font%20form%20frame%20head%20hr%20html%20h1%20h2%20h3%20h4%20h5%2

[04:31:01.0000] <zcorpan>
0h6%20i%20input%20iframe%20ilayer%20img%20inlineinput%20ins%20kbd%20keygen%20label%20layer%20legend%20li%20link%20listing%20map%20marquee%20menu%20meta%20multicol%20nextid%20nobr%20noembed%20noframes%20nolayer%20noscript%20object%20ol%20optgroup%20option%20p%20param%20plaintext%20pre%20q%20rt%20ruby%20s%20samp%20script%20select%20small%20sound%20spacer%20span%20spell%20strike%20strong%20style%20sub%20sup%20table%20tbody%20td%20textarea%20tfoot%20

[04:31:02.0000] <zcorpan>
th%20thead%20title%20tr%20tt%20u%20ul%20var%20wbr%20xmp%22.split(%22%20%22)%3B%0D%0Avar%20div%20%3D%20document.createElement('div')%3B%0D%0Avar%20current%3B%0D%0Afor%20(var%20i%20%3D%200%3B%20i%20%3C%20elms.length%3B%20%2B%2Bi)%20%7B%0D%0A%20%20current%20%3D%20div.cloneNode(false)%3B%0D%0A%20%20current.innerHTML%20%3D%20elms%5Bi%5D%20%2B%20'%3Cselect%3E1%3C'%20%2B%20elms%5Bi%5D%20%2B%20'%3E2%3C%2F'%20%2B%20elms%5Bi%5D%20%2B%20'%3E3%3C%2Fselect%3E

[04:31:03.0000] <zcorpan>
'%3B%0D%0A%20%20document.body.appendChild(current)%3B%0D%0A%7D%0D%0A%3C%2Fscript%3E

[04:31:04.0000] <zcorpan>
oops

[04:32:00.0000] <zcorpan>
ie inserts unknown elements

[04:34:00.0000] <zcorpan>
#text: 12</plaintext>3</select>

[04:34:01.0000] <zcorpan>
interesting

[04:35:00.0000] <zcorpan>
<table> breaks out of <select> in ie

[04:36:00.0000] <zcorpan>
and all other table elements

[04:39:00.0000] <zcorpan>
<select><plaintext> is certainly funny in webkit and ie

[04:41:00.0000] <Philip`>
Lots of people put <script>s and <br>s inside <select>

[04:41:01.0000] <tthorsen>
the <script> I can understand, but what's the <br> supposed to do?

[04:42:00.0000] <zcorpan>
<br> doesn't do anything

[04:42:01.0000] <Philip`>
Don't ask me to make sense of the web :-)

[04:42:02.0000] <Philip`>
Also lots of comments in there

[04:43:00.0000] <zcorpan>
not <style>?

[04:43:01.0000] <zcorpan>
or <td>

[04:43:02.0000] <Philip`>
and an <a> (on http://www.dahmen-quilt.com/)

[04:44:00.0000] <Philip`>
zcorpan: None that I see, when grepping some thousands of pages and looking through the output manually

[04:44:01.0000] <zcorpan>
<table><td><select><td>

[04:45:00.0000] <zcorpan>
oh the spec has in table in select

[04:45:01.0000] <Philip`>
http://www.inmobiliariaportilla.com/ has a <select><font>

[04:45:02.0000] <Philip`>
and http://www.roommatelocator.com/ has a <select></font>

[04:46:00.0000] <Philip`>
http://www.rentjillshouse.com/ has a <td><select></td>

[04:47:00.0000] <zcorpan>
it seems the spec needs to fix <select><script>

[04:48:00.0000] <zcorpan>
/me files a bug

[04:48:01.0000] <krijnh>
Yay, more mail :)

[04:52:00.0000] <zcorpan>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=6216

[04:53:00.0000] <Philip`>
http://www.keysmusic.co.uk - <select><ul>

[04:55:00.0000] <zcorpan>
<isindex> "works" in select in ie

[05:04:00.0000] <tthorsen>
zcorpan: fwiw, my preliminary fix for scripts in selects will be to do exactly what we do in "in body", which is: 'Process the token using the rules for the "in head" insertion mode.'

[05:05:00.0000] <zcorpan>
sounds reasonable

[05:05:01.0000] <tthorsen>
works for my simple test, too

[05:12:00.0000] <jgraham>
tthorsen: Are you using the html5lib test suite?

[05:16:00.0000] <tthorsen>
I do test markup with it from time to time, but I'm not using it very much.

[05:17:00.0000] <tthorsen>
Oh. I read you wrong. I've run our engine through some of the tests in the test suite, but not all

[05:18:00.0000] <tthorsen>
I don't have anything automated, so it would take me a long time to try them all.

[05:18:01.0000] <jgraham>
OK, just wondering becuase if you have new tests that yu can release under the MIT license it would be great to add them to the test suite

[05:20:00.0000] <zcorpan>
hmm anne and i wrote a browser runner for the html5lib tests before

[05:20:01.0000] <tthorsen>
ah, no I'm afraid most of our testing consists of proprietarily licensed tests and testing real-world web sites

[05:20:02.0000] <zcorpan>
dunno where it is

[05:23:00.0000] <jgraham>
http://html5.org/parsing-tests/testrunner.htm

[05:24:00.0000] <takkaria>
tthorsen: you should run against the html5lib test suite anyway, it's very useful

[05:24:01.0000] <takkaria>
picked up dozens and dozens of issues when I was hacking on hubbub (a C html5 parsing library)

[05:25:00.0000] <tthorsen>
yeah. I did find a bunch of bugs last time I ran a few of them

[05:26:00.0000] <takkaria>
actually, I have a bunch of character-encoding tests that I should contribute back to html5lib

[06:33:00.0000] <yecril71>
The problem of external pages sniffing the content of

[06:33:01.0000] <yecril71>
intranet files, or just some characteristics of it,

[06:33:02.0000] <yecril71>
is similar to the problem of Web sites loading 

[06:33:03.0000] <yecril71>
local resources via the file URL scheme,

[06:34:00.0000] <yecril71>
only that now the URL scheme is not a distinguishing feature.

[06:34:01.0000] <yecril71>
Should such limitations be ever implemented, the exact behaviour 

[06:35:00.0000] <yecril71>
should depend on the Internet zone of the embedding document,

[06:35:01.0000] <yecril71>
just as with XMLHTTPRequest.

[06:36:00.0000] <yecril71>
For example, Internet sites should not be allowed

[06:36:01.0000] <yecril71>
to get information about intranet content 

[06:36:02.0000] <yecril71>
or load it.

[06:53:00.0000] <takkaria>
Internet zoning an IE concept, though, not a concept of the Web per se

[07:12:00.0000] <yecril71>
Interet zoning is a concept that is adequate to the situation of the question.

[07:13:00.0000] <yecril71>
Which is unintentionally leaking internal data, which is not a concept of the Web either.

[07:25:00.0000] <takkaria>
well, yes, but it's something best brought to the attention of browser developers and not mentioned in passing on an IRC channel :)

[07:29:00.0000] <Philip`>
The same-origin security restrictions are needed to avoid providing ways around any firewall restrictions, not just for the internet/intranet divide

[07:37:00.0000] <yecril71>
Can you describe a situation when the data leak does not sit in the Intranet?

[07:39:00.0000] <yecril71>
And the purpose of this channel is to formulate recommendations for all browser developers

[07:39:01.0000] <yecril71>
where they feel the need to act in the same way.

[07:57:00.0000] <Philip`>
yecril71: There are lots of internet sites where access is restricted to certain IPs, e.g. I can access various academic papers if I'm using my university's internet connection, and it would be a data leak if someone else on the internet could access those resources via my browser

[08:25:00.0000] <Lachy>
/me is surpised by what they call a continental breakfast over here in the US. There wasn't even any toast!

[08:29:00.0000] <MikeSmith>
Lachy: "continental" is this case is a euphemism for "miniscule"

[08:31:00.0000] <Philip`>
Where "minuscule" is a euphemism for "not all-you-can-eat pancakes"?

[08:34:00.0000] <MikeSmith>
eating lots of pancakes is a great way for visitors to fit in with America life, by helping them to reach the typical "American size" (euphemism for "fat")

[08:35:00.0000] <jmb>
/me so failed at that last week

[08:36:00.0000] <MikeSmith>
follow up your pancake breakfast with a stop by AM/PM to get a 32-ounce Coca-Cola, and in the evening, enjoy a great movie at the cinema while eating popcorn for a container the size of a garbage pail

[08:38:00.0000] <jcranmer>
continental = coffee + donuts, in my experience

[08:38:01.0000] <jcranmer>
often times OJ as well; if you're lucky, you'll find a waffle machine

[11:02:00.0000] <hsivonen>
Hixie: the single-page version of the spec links to itself instead of the multipage version

[11:06:00.0000] <takkaria>
Leif has high hopes for web developers

[11:07:00.0000] <Lachy>
there doesn't really seem to be very much interesting sessions for me to attend today at pubcon. It's all about marketing :-(

[11:07:01.0000] <Lachy>
I might skip some sessions today and go explore Vegas

[11:08:00.0000] <blooberry>
lachy: a tuesday in November should be a pretty good time to do that. How are the throngs? And the weather?

[11:09:00.0000] <Lachy>
the weather is good today

[11:09:01.0000] <Lachy>
what are "throngs"?

[11:09:02.0000] <blooberry>
I was there last year at the end of november and it was great weather except for a torrential downpour that the Las Vegas drainage system seemed ill-equipped to handle

[11:09:03.0000] <blooberry>
throngs = the crowds

[11:09:04.0000] <Lachy>
in what language?

[11:10:00.0000] <blooberry>
english 8-} http://dictionary.reference.com/browse/throng

[11:12:00.0000] <Lachy>
ok

[11:13:00.0000] <blooberry>
what parts of vegas interest you? and are you on the strip? (if so, which end)

[11:13:01.0000] <Lachy>
I'm at the Las Vegas Convention Center

[11:13:02.0000] <Lachy>
it's not too far from the strip

[11:15:00.0000] <Lachy>
I have to go to some of the casinos later. Where exactly is the strip?

[11:16:00.0000] <Lachy>
http://maps.google.com/maps?f=q&hl=en&geocode=&q=3150+Paradise+Road,+Las+Vegas,+NV+89101&sll=36.128516,-115.162833&sspn=0.01409,0.021479&ie=UTF8&z=17&g=3150+Paradise+Road,+Las+Vegas,+NV+89101&iwloc=addr

[11:16:01.0000] <Lachy>
that's the map of the convention centre

[11:17:00.0000] <blooberry>
Once you get to the strip, there's a shuttle you can take up/down the strip. It is much longer than the maps make them look

[11:18:00.0000] <blooberry>
that's because all the hotels are soooo damn huge. Takes forever to get from one end of one to another.

[11:19:00.0000] <Lachy>
I'll probably go there later in the week, if not tonight, and gamble all my life savings away

[11:19:01.0000] <Lachy>
this session is over. I'm going to pack up and go. CYA

[11:19:02.0000] <blooberry>
woohoo! 8-}

[11:19:03.0000] <blooberry>
k.

[11:41:00.0000] <Philip`>
/me guesses that "pubcon" is not a con about pubs

[11:43:00.0000] <Philip`>
blooberry: Is the strip longer than it looks in GTA:SA?

[11:44:00.0000] <gsnedders>
Touché, Philip`.

[11:44:01.0000] <gsnedders>
But there again, that's the best I have to go by too.

[11:44:02.0000] <Philip`>
?

[11:44:03.0000] <hsivonen>
/me thought San Andreas was Los Angeles, not Las Vegas

[11:45:00.0000] <Philip`>
hsivonen: It's Los Angeles, San Francisco and Las Vegas

[11:45:01.0000] <gsnedders>
hsivonen: LA, SF, and LV

[11:45:02.0000] <gsnedders>
/me can't type quickly enough yet on this Dvorax keyboard

[11:45:03.0000] <gsnedders>
*Dvorak

[11:46:00.0000] <Philip`>
Dvorax sounds like the leader of a quite nasty alien species

[11:46:01.0000] <hsivonen>
ah

[11:46:02.0000] <gsnedders>
Philip`: Just using GTA as a reference :)

[11:47:00.0000] <hsivonen>
hrm. the entity table has lost line breaks since I last tried to scrape it

[11:47:01.0000] <gsnedders>
That has nothing to do with me.

[11:50:00.0000] <hsivonen>
scraping HTML with regexps is so brittle... :-(

[11:51:00.0000] <Philip`>
It's a shame you don't have a better way of parsing HTML

[12:09:00.0000] <hsivonen>
http://ejohn.org/blog/css-animations-and-javascript/#comment-322033

[12:10:00.0000] <hsivonen>
an outside-HTML5 case for using the event loop definition

[13:13:00.0000] <roc>
othermaciej: would you mind posting to the list to explain your position on cross-origin <video> loads?

[13:54:00.0000] <othermaciej>
roc: what more needs explanation?

[13:54:01.0000] <othermaciej>
roc: I said I think they should be supported

[13:54:02.0000] <othermaciej>
roc: I think the reasons why are obvious (you stated some of them)

[13:55:00.0000] <roc>
I guess I'm curious about which pros and cons carry weight with you and which don't

[13:56:00.0000] <roc>
obviously you're not concerned about information leakage from intranets, and I'd like to know why that is

[13:57:00.0000] <othermaciej>
I think leaking the size of files is not that huge a deal, at least if it is limited only to audio or video files

[13:59:00.0000] <roc>
ok

[14:00:00.0000] <roc>
What about duration? In the future, what about cue points? Caption data?

[14:01:00.0000] <othermaciej>
I think media plugins mean sites are already exposed to leaking that kind of information

[14:01:01.0000] <othermaciej>
well maybe not cue points or caption data as text embedded in the media file

[14:02:00.0000] <roc>
I'm torn over how much to weigh the media plugins argument. It's strong on its face, but on the other hand, being just as secure as plugins is kind of a low bar

[14:02:01.0000] <othermaciej>
I know, but this also isn't an area of plugin security people complain about

[14:03:00.0000] <othermaciej>
much as they don't complain that you can get width and height of a cross-site embedded image

[14:03:01.0000] <roc>
it's possible that's just because no-one's got around to figuring out how bad it is

[14:03:02.0000] <roc>
like clickjacking

[14:05:00.0000] <roc>
FWIW, Jonas and I were in favour of same-origin restrictions plus Access Control, but others aren't, and your opinion carries a lot of weight with me so now I'm totally unsure!

[14:05:01.0000] <Hixie>
i don't understand why people are even giving lip service to this ridiculous idea of inventing a new DRM format for fonts

[14:05:02.0000] <roc>
But I'd like Hixie to chime in since he filed that W3C bug

[14:05:03.0000] <Hixie>
/me looks up

[14:05:04.0000] <Hixie>
wassup?

[14:06:00.0000] <roc>
the question of same-origin restrictions for media elements

[14:07:00.0000] <Hixie>
my concern was caption data on intranet videos being made accessible in a future version of the spec, primarily

[14:07:01.0000] <roc>
ok

[14:07:02.0000] <Hixie>
(i don't want to have to introduce a "use access control because i want caption data" flag later)

[14:13:00.0000] <roc>
why would that flag be needed? Does Access-Control-Allow-Origin not come back unless we asked for it?

[14:18:00.0000] <hsivonen>
roc: have you analyzed how bad the safe part of the API would be if the unsafe parts of the API were turned off when loading cross-site without AC?

[14:18:01.0000] <roc>
it's not clear what is safe, really.

[14:19:00.0000] <roc>
is leaking the video size safe? The format? The duration?

[14:19:01.0000] <roc>
it's possible to imagine scenarios where each of these reveals something the intranet owner doesn't want revealed

[14:20:00.0000] <hsivonen>
I wonder if it's actually a good idea that MS has an intranet zone setting

[14:20:01.0000] <hsivonen>
(I've always thought it was a bad idea)

[14:20:02.0000] <roc>
it's helpful in some ways

[14:20:03.0000] <roc>
I don't know we'd want to rely on it

[14:31:00.0000] <hsivonen>
Hixie: do you count a font-specific compression scheme without root strings as DRM?

[14:33:00.0000] <hsivonen>
Hixie: <devilsadvocate>Is JPEG DRM when a stock photo licenser allows distribution as JPEG but not as CMYK TIFF?</devilsadvocate>

[14:37:00.0000] <Philip`>
HTML5 parses "<code><pre>...</code></pre>..." badly - it puts the second piece of text inside <code>, unlike browsers, which breaks http://blogs.sun.com/bblfish/entry/rest_apis_must_be_hypertext (around the JSON example)

[14:38:00.0000] <Hixie>
roc: it would be needed if we allowed cross-site today and then had to disable access to captions when we introduced captions

[14:39:00.0000] <Hixie>
hsivonen: no, compression schemes (that can be supported by the platform) are fine

[14:39:01.0000] <Hixie>
hsivonen: no, that's just a license

[14:40:00.0000] <Hixie>
Philip`: send feedbackj

[14:40:01.0000] <Hixie>
back

[14:40:02.0000] <Hixie>
no j

[14:40:03.0000] <roc>
Hixie: when you say "a flag", you mean an internal flag defined in the spec that's set for cross-origin videos, or some kind of flag that has to be set by users?

[14:41:00.0000] <Hixie>
i meant a flag for developers to set to say "ok this video should be loaded with AC so that i can get the captions", but i see that wouldn't be the only way to do it; regardless, i'd not really want a mode in which half the API is disabled because of lack of AC

[14:42:00.0000] <Hixie>
so "both"

[14:44:00.0000] <roc>
ok

[14:44:01.0000] <roc>
can you fight it out with maciej and let me know the result? thanks

[14:44:02.0000] <Hixie>
hehe

[14:44:03.0000] <Hixie>
maciej doesn't want restrictions?

[14:45:00.0000] <roc>
he wants either "cross-origin loads allowed with no restrictions on API" or "cross-origin loads allowed but with a restricted API"

[14:45:01.0000] <Hixie>
ah

[14:45:02.0000] <Hixie>
does he have reasons? (/me pokes othermaciej)

[14:45:03.0000] <roc>
I probed him about that just before in the channel

[14:45:04.0000] <othermaciej>
I think requiring Access-Control for any cross-origin media embedding is too much a developer hardship compared to plugin-based solutions that handle this fine

[14:46:00.0000] <Hixie>
fair point

[14:46:01.0000] <Hixie>
so we'd need anne to introduce a third mode in AC

[14:46:02.0000] <Hixie>
that returns the content despite it failing an ac check

[14:47:00.0000] <Hixie>
but with a "disable dangerous apis" option

[14:47:01.0000] <Hixie>
though frankly i'm a little concerned about exposing intranet videos of any kind

[14:48:00.0000] <Hixie>
i mean, if someone tricked me into going to their site, and had prior knowledge of where we put videos of our important internal events, they could get information about how long our internal meetings were going for

[14:48:01.0000] <Hixie>
which is a good indication of whether things are healthy or not at the company

[14:48:02.0000] <Hixie>
well, dunno about good

[14:48:03.0000] <Hixie>
but it's an indication anyway

[14:49:00.0000] <Hixie>
not to mention they could determine things like what bitrate we were using, information that i wouldn't personally be comfortable leaking explicitly, so why would i want to leak it accidentally?

[14:50:00.0000] <othermaciej>
I also believe the perceived security risks are partially imaginary as no one even complains about the same behavior for plugins; I could see the future API risks but I think exposing some parts of the API only for same-origin or if AC check passes is a better compromise than limiting any video to same-origin/AC

[14:50:01.0000] <gavin>
/me can't imagine why anyone would be worried about leaking the bitrate of internal videos

[14:50:02.0000] <othermaciej>
I would perhaps feel differently if AC were already more widely deployed in the infrastructure so it were not seen as much of an authoring hardship relative to plugin-based solutions

[14:51:00.0000] <Hixie>
i certainly understand and sympathise with your position

[14:51:01.0000] <roc>
so do I

[14:52:00.0000] <roc>
I wonder if XHR will be enough to drive deployment of AC, or we'll be stuck in a chicken-and-egg situation

[14:55:00.0000] <hsivonen>
How common is it today with plug-in-based solutions that people have videos neither on same origin nor on a video-specializing site like youtube?

[14:55:01.0000] <hsivonen>
I wonder how onerous it would be to put AC on Akamai-hosted Stevenotes

[14:56:00.0000] <roc>
Getting this resolved is a super-high priority issue for us, by the way. We're shipping soon and if we're going to have restrictions, we need to implement them now, or we'll never be able to

[14:57:00.0000] <Hixie>
i don't think <video> uptake is going to be so fast that we couldn

[14:57:01.0000] <Hixie>
't add restrictions in the next cycle

[14:57:02.0000] <Hixie>
/me looks at the API to see what is currently exposed

[14:58:00.0000] <Hixie>
dimensions, size, bit rate, bandwidth to hosting site

[14:58:01.0000] <Hixie>
and existence

[14:58:02.0000] <roc>
maybe you're right, we've already identified some sites that would break if we added same-origin restrictions right now, and that's only going to get bigger

[14:58:03.0000] <roc>
^ but

[14:58:04.0000] <Hixie>
sure

[14:59:00.0000] <roc>
duration

[14:59:01.0000] <roc>
format

[14:59:02.0000] <Hixie>
format?

[14:59:03.0000] <roc>
yeah

[14:59:04.0000] <Hixie>
i don't think we expose format explicitly in the spec

[14:59:05.0000] <roc>
I think you can fake it using <source> elements and redirects

[14:59:06.0000] <Hixie>
true

[14:59:07.0000] <roc>
er

[14:59:08.0000] <roc>
no redirects, just use currentSrc

[15:00:00.0000] <doublec>
hsivonen, having media upload/download from a subdomain is quite common

[15:00:01.0000] <doublec>
which breaks without AC

[15:01:00.0000] <hsivonen>
doublec: ok

[15:06:00.0000] <Hixie>
i guess what we should do is not have cross-site restrictions for now, but when we add them to make it so that if there is AC metadata and it is negative, to block everything

[15:06:01.0000] <Hixie>
allowing opt-out of all video access, and opt-in to full metadata access

[15:06:02.0000] <Hixie>
defaulting to basically the current API set

[15:07:00.0000] <Hixie>
but we should be prepared to block that altogether if someone comes up with an unexpected attack vector

[15:08:00.0000] <roc>
ok

[15:08:01.0000] <roc>
so we'll leak all those things you mentioned, for media types?

[15:08:02.0000] <Hixie>
yeah

[15:08:03.0000] <doublec>
is it worth not exposing anything about the media until metadata loaded?

[15:08:04.0000] <doublec>
so progress events, etc don't get fired until after that?

[15:09:00.0000] <Hixie>
but if you disagree with that, or aren't comfortable with it, then implement AC, and then you'll make the de-facto decision be to have access limitations :-)

[15:09:01.0000] <Hixie>
and i'll spec that :-)

[15:09:02.0000] <Hixie>
that's the first mover advantage :-) of course the disadvantage is that you might have to change more later.

[15:09:03.0000] <roc>
like I said, there isn't even consensus within Mozilla

[15:10:00.0000] <Hixie>
doublec: what is fired so far? just loadstart and progress?

[15:10:01.0000] <doublec>
yes

[15:10:02.0000] <Hixie>
loadstart seems safe, since that'll fire regardless

[15:10:03.0000] <doublec>
so progress gives filesize, etc for example

[15:11:00.0000] <Hixie>
i could see omitting progress until you have the headers down, but you can't give the size until you have the headers anyway

[15:11:01.0000] <Hixie>
so you'd only be able to tell if the site was down or not

[15:11:02.0000] <Hixie>
which you can already tell using new Image()

[15:13:00.0000] <roc>
if we use a platform backend, we may not be able to tell whether it's a usable media file immediately after loading the headers

[15:18:00.0000] <Hixie>
i don't think exposing whether it's valid or not is a problem

[15:18:01.0000] <Hixie>
the concern is over not exposing that the file is not 404 until we've checked for AC headers

[15:37:00.0000] <roc>
we don't want to fire progress events before we know it's a media file

[15:37:01.0000] <roc>
and yes, we need to make sure that non-media files and 404s appear identically to the API user


2008-11-12
[16:06:00.0000] <Hixie>
man i wish people would put blank lines between their paragraphs

[16:16:00.0000] <Lachy>
I generally don't bother wasting my time trying to read emails that have been formatted poorly

[16:17:00.0000] <Hixie>
unfortunately i don't have that privilege really

[16:20:00.0000] <Hixie>
Philip`: fixed set/group for you

[16:21:00.0000] <Philip`>
Hixie: Thanks - you are too kind

[16:22:00.0000] <Lachy>
I have too many emails to read. I've been trying to get through my whatwg folder, but given all my travelling and poor quality internet access while not at home, it's difficult to even keep the unread messages below 900 :-(

[16:23:00.0000] <Hixie>
no worries, my unread is at 2677

[16:23:01.0000] <Hixie>
http://www.whatwg.org/issues/data.html :-)

[16:23:02.0000] <Hixie>
(for some definition of "unread" that is really "unreplied")

[16:53:00.0000] <Lachy>
I'm looking for things to do in Vegas this week. This looks like fun! :-) http://www.vegasindoorskydiving.com/

[16:53:01.0000] <Lachy>
it's right next door to my hotel too

[16:54:00.0000] <jruderman>
Lachy: indoor skydiving? shouldn't that be ceilingdiving?

[16:56:00.0000] <gavin>
their "celebrity spotting" link points to file:///C|/Users/Cherlyn%20Fields/Desktop/pictures2.php

[16:57:00.0000] <Lachy>
LOL

[16:57:01.0000] <jruderman>
i just tried to click that link :(

[16:58:00.0000] <gavin>
http://www.vegasindoorskydiving.com/pictures2.php works

[16:58:01.0000] <gavin>
though the thumbnails are full images, judging by how fast they loaded

[16:58:02.0000] <Lachy>
from one page, it links to http://www.vegasindoorskydiving.com/pictures2.php

[17:00:00.0000] <Lachy>
I think I might go do that tonight and then head off to one of the casinos for dinner.

[17:01:00.0000] <Lachy>
But I still need to work out what to do for the rest of the week.

[17:01:01.0000] <Lachy>
Other than gambling away my life savings, I think I should go see a show or something

[17:02:00.0000] <Lachy>
any recommendations?

[17:02:01.0000] <Hixie>
avenue q

[17:03:00.0000] <Hixie>
though the vegas one is a truncated version

[17:04:00.0000] <Lachy>
is it still running?

[17:04:01.0000] <Hixie>
should be

[17:04:02.0000] <Hixie>
oh maybe not

[17:04:03.0000] <Lachy>
I found an New York Times article about it from 2006

[17:07:00.0000] <Hixie>
spamalot apparently replaced it

[17:07:01.0000] <Hixie>
i imagine that's probably fun too

[17:07:02.0000] <Hixie>
but i don't know it

[17:11:00.0000] <Lachy>
doesn't look like it is http://www.avenueq.com/tour/

[17:12:00.0000] <Hixie>
weee, they're coming to san jose in a few months!!!

[17:12:01.0000] <Hixie>
/me marks his calendar for when the tickets go on sale

[17:12:02.0000] <Hixie>
lol, my girlfriend already marked it

[17:22:00.0000] <Lachy>
damn, the wifi here is way too slow. I can't even read my email :-(

[17:22:01.0000] <Lachy>
oh well, I'm off to go dive into a giant fan. :-)

[17:23:00.0000] <Hixie>
later :-)

[19:28:00.0000] <Lachy>
I managed to survive my sky dive, that was awesome

[19:28:01.0000] <Lachy>
it's definitely harder than it looks though

[19:46:00.0000] <Lachy>
I'm guessing Chris Wilson thinks the term vendor-lock-in is offensive in this contenxt because its companies like Microsoft that always get accused of it

[19:48:00.0000] <roc>
what context?

[19:54:00.0000] <Lachy>
roc, see public-html

[19:54:01.0000] <roc>
oh

[19:54:02.0000] <roc>
no thanks!

[19:55:00.0000] <Lachy>
roc, why? Don't you read that list anymore?

[19:55:01.0000] <roc>
I never have

[19:55:02.0000] <jwalden>
/me was going to type "any more?"

[19:56:00.0000] <Lachy>
jwalden, yeah, I meant "any more", but I didn't press space properly

[19:56:01.0000] <jwalden>
no, I was echoing roc, not criticizing a typo -- although I did notice it :-P

[19:56:02.0000] <Lachy>
oh

[19:57:00.0000] <jwalden>
always looked like a huge cesspool when I've occasionally skimmed web archives

[19:57:01.0000] <jwalden>
too many people who think they can cook spoil the broth

[19:59:00.0000] <Lachy>
I wonder where I should go for dinner tonight. I suppose there will be restaurants around in the casinos, if not cheaper ones elsewhere

[19:59:01.0000] <jwalden>
actually, casino restaurants may be cheaper, to entice you to come in and make up their loss inside

[19:59:02.0000] <jwalden>
that's how it was when I was traveling through the west with family a number of years ago

[20:01:00.0000] <jwalden>
hm, judging by message count public-html may have improved since I last looked at it

[20:34:00.0000] <Lachy>
/me discovers that The Montecito casino featured in Las Vegas (the TV series) was fictional!

[20:35:00.0000] <Lachy>
until now, I just assumed it was a real one

[20:36:00.0000] <Lachy>
but at least the Bellagio from Ocean's 11 is real :-)

[20:47:00.0000] <roc>
for some definition of real

[01:00:00.0000] <Hixie>
oooh, leap second at the new year

[01:02:00.0000] <hsivonen>
/me still thinks that needing announcement-based coordination for something as fundamental as time sucks big time

[01:03:00.0000] <hsivonen>
HTML5 should probably say that the system clocks are supposed to observe UTC but the calculations happen as if the numbers were UT1 times

[01:04:00.0000] <Hixie>
i think what it says now is clearer

[01:05:00.0000] <Hixie>
and how else would you do it? assuming that you want to keep the clocks in sync with the sun, that is

[01:05:01.0000] <tthorsen>
Couldn't you just adjust the velocity and rotation of the earth?

[01:06:00.0000] <tthorsen>
giant rocket-engines, maybe

[01:06:01.0000] <hsivonen>
Hixie: I think keeping solar year and atomic year in sync on that level of accuracy falls in the theoretical purity bucket

[01:06:02.0000] <hsivonen>
Posix time is sane

[01:07:00.0000] <hsivonen>
Hixie: I think at this point it's OK to leave sundial legacy UAs as a bit inaccurate

[01:08:00.0000] <hsivonen>
for the next few thousand years timezones already cause more sundial incompatibility than strict atomic time

[01:08:01.0000] <Hixie>
if you don't do this, then a few decades or centuries from now, you'll be a few hours off, and after a few thousand years, your noon will be at midnight.

[01:09:00.0000] <hsivonen>
oh. last time I calculated the max drift it was a lot smaller

[01:09:01.0000] <hsivonen>
perhaps I miscalculated

[01:11:00.0000] <hsivonen>
there are 3600 seconds in an hour

[01:11:01.0000] <hsivonen>
you can have at most 2 leap seconds per year

[01:11:02.0000] <hsivonen>
but the average is closer to one

[01:11:03.0000] <hsivonen>
so to get an hour of drift, you'd need to wait 3600 years

[01:12:00.0000] <hsivonen>
what did I calculate wrong?

[01:13:00.0000] <hsivonen>
and we accept an hour of drift every year. twice

[01:18:00.0000] <Hixie>
the average is closer to 1 ever 2 or 3 years iirc

[01:19:00.0000] <Hixie>
every

[01:19:01.0000] <hsivonen>
Hixie: doesn't that make my point stronger?

[01:26:00.0000] <Hixie>
drift takes a long time, yes

[01:27:00.0000] <Hixie>
apparently the average is 1.4 milliseconds per day per century

[01:28:00.0000] <Hixie>
average deceleraton of time, that is

[01:28:01.0000] <Hixie>
well, deceleration of earth

[01:28:02.0000] <Hixie>
(meaning longer days)

[01:29:00.0000] <hsivonen>
Hixie: it seems to me that an hour of drift in the cultural notation over 4000 years or so is less of a problem than trying to accommodate leap seconds in software

[01:30:00.0000] <hsivonen>
after all, by some accounts, the Earth is now only 6000 years old, so 4000 years is a relative long time

[01:34:00.0000] <Hixie>
well, if you can convince the relevant governments to decouple their civil time from astronomical measurements, then we'll go talk to the ITU

[01:34:01.0000] <Hixie>
the alternative, replacing leap seconds with leap hours, imho causes even more confusion

[01:34:02.0000] <Hixie>
and is far more likely to involve painful software bugs

[01:35:00.0000] <Hixie>
code that runs in production once every 4000 years is unlikely to be bug-free

[01:36:00.0000] <hsivonen>
Code that depends on external announcements is sure to perform different calculations before and after receiving an announcement

[01:36:01.0000] <Hixie>
(alternatively, you could just use TAI or GPS time instead of UTC)

[01:36:02.0000] <Hixie>
most code doesn't need to care about leap seconds

[01:37:00.0000] <Hixie>
because the clocks are likely to drift by more than that in normal operation anyway

[01:37:01.0000] <hsivonen>
the concept of civil time sucks

[01:37:02.0000] <Hixie>
so the normal clock-sync mechanisms will fix leap seconds for you

[01:38:00.0000] <hsivonen>
Hixie: I think it's acceptable to pretend that the current time is UTC as long as calculations don't implement UTC

[01:38:01.0000] <hsivonen>
implementing real UTC in software would be crazy

[01:38:02.0000] <hsivonen>
i.e. it's OK to do Posix time calculations and set the current time UTC

[01:39:00.0000] <hsivonen>
Hixie: have you seen the serious proposals for smoothing leap second transitions? those are some serious complexity

[01:40:00.0000] <Hixie>
UTC-SLS?

[01:41:00.0000] <hsivonen>
yeah

[01:41:01.0000] <Hixie>
most software wouldn't need to know about it

[01:41:02.0000] <Hixie>
as far as i can tell

[01:44:00.0000] <roc>
the whole notion of simultaneity is flawed. Each processor must maintain its own local time and compensate for relative velocities when translating times into other frames of reference

[01:45:00.0000] <hsivonen>
Hixie: doesn't it bother you that different pieces of software would do different conversions so if you use midnight to represent dates you get random-looking *day*-level drift?

[01:45:01.0000] <Hixie>
not really, but maybe i don't understand what you mean

[01:46:00.0000] <jgraham>
roc: Don't forget variations in g

[01:46:01.0000] <hsivonen>
Hixie: the usual practice for representing dates to the precision of a day is to store seconds from a reference point and zero the least-significant seconds

[01:47:00.0000] <Hixie>
between hardware limitations (and, to a far lesser extent, relativistic, gravitational, and black-box radiation effects) and configuration errors, i doubt most consumer computers ever get anywhere near accurate enough for leap seconds to matter

[01:47:01.0000] <hsivonen>
Hixie: right, so leap seconds make no sense for "civil time"

[01:47:02.0000] <jgraham>
/me wonders h

[01:47:03.0000] <jgraham>
what

[01:47:04.0000] <hsivonen>
and military time doesn't use them

[01:47:05.0000] <hsivonen>
hence, theoretical purity needless complexity bucket

[01:47:06.0000] <jgraham>
Hixie means by "black-box radiation effects"

[01:48:00.0000] <zcorpan>
tthorsen: the spec actually matches ie, except that ie hides what's in object

[01:48:01.0000] <zcorpan>
tthorsen: try <p><object></p></object>x

[01:48:02.0000] <tthorsen>
how about the extra <p></p>?

[01:48:03.0000] <hsivonen>
aside, teaching people to say UTC instead of GMT is like teaching them to say URI instead of URL

[01:48:04.0000] <zcorpan>
tthorsen: stray </p> means <p></p>

[01:49:00.0000] <zcorpan>
tthorsen: why is it a problem?

[01:49:01.0000] <Hixie>
hsivonen: civil time in many countries is legally defined in terms of astronomical state. if we're going to use SI seconds and 36400 seconds-per-day, then that breaks. so we either have to change the laws, or have leap seconds (or hours, or whatever)

[01:49:02.0000] <Hixie>
jgraham: caesium atomic clocks vary in speed based on ambient temperature

[01:50:00.0000] <Hixie>
jgraham: i meant blackbody radiation, my bad

[01:50:01.0000] <tthorsen>
It's not really a problem for me, but I thought it was a bug in the spec. Firefox and opera usually makes empty <p> elements when they see stray </p>'s but not in this case

[01:50:02.0000] <Philip`>
36400 seconds per day would be pretty weird

[01:51:00.0000] <zcorpan>
tthorsen: that's because firefox and opera don't treat <object> as scoping

[01:51:01.0000] <Hixie>
er, 86400

[01:51:02.0000] <zcorpan>
yet

[01:51:03.0000] <hsivonen>
Hixie: naively, one would think laws were more malleable than the way Earth's orbit time divides by cesium pulses

[01:52:00.0000] <tthorsen>
ok, If this is the intended behaviour of the spec, then I'll happily change our test cases

[01:52:01.0000] <Hixie>
hsivonen: like i said, let me know when you get the laws changed, and we can take it to the ITU. :-)

[01:53:00.0000] <tthorsen>
after all this means that I get to do nothing, and the guy who's responsible for the tests needs to do all the work :-)

[01:53:01.0000] <zcorpan>
tthorsen: fwiw we've changed to match the spec (for our next major release)

[01:55:00.0000] <tthorsen>
zcorpan: sounds good to me. Fwiw we will also match the spec for our next major release.

[01:59:00.0000] <tthorsen>
zcorpan: The thing we talked about yesterday - about ignoring cdata in select tags. Do you know whether the spec will be changed or not?

[01:59:01.0000] <jgraham>
hsivonen: Per wikipedia leap seconds will be needed much more frequently in the future, although it seemes there is a plan to vote on the issue in 2011

[02:00:00.0000] <tthorsen>
I that case too, it doesn't really matter if the spec changes or not, but I'd like some kind of confirmation before I go ahead and get all our tests changed.

[02:00:01.0000] <tthorsen>
s/doesn't really matter/doesn't really matter to me/

[02:00:02.0000] <jgraham>
tthorsen: What usually happens is that we wait for Hixie to read the feedback, he makes a judgement call and then we complain again if we think he used faulty logic :)

[02:01:00.0000] <jgraham>
If it's a high priority for you then I suggest that you make him aware of that so he can schedule his priorities accordingly

[02:02:00.0000] <tthorsen>
Nah, the normal procedure sounds fine to me.

[02:03:00.0000] <jgraham>
(I should note that there is sometimes a multi-year delay between the feedback being recieved and it being considered in detail, although I guess the parsing section will be ouched much sooner than that)

[02:04:00.0000] <jgraham>
touched

[02:04:01.0000] <takkaria>
fwiw, NetSurf (browser using Hubbub) has had no reported problems with the current HTML5 algorithm, but it has a very small number of users and no scripting support

[02:07:00.0000] <tthorsen>
well, a couple of the things I mentioned were mere clarification issues. A couple of things turned out to be the intended behaviour, and not a bug, after all. So in the end we don't see a lot of real problems either.

[02:08:00.0000] <Hixie>
oldest feedback still pending a reply is from unix time_t 1073954713

[02:08:01.0000] <Hixie>
whenever that is

[02:09:00.0000] <Hixie>
oldest feedback that's going to get a reply any time soon is from unix time_t 1101133613

[02:09:01.0000] <Hixie>
(the oldest feedback is about the rendering section)

[02:09:02.0000] <takkaria>
13th January

[02:09:03.0000] <takkaria>
2004

[02:09:04.0000] <Hixie>
wow, that's older than the whatwg

[02:10:00.0000] <takkaria>
and 22 November 2004 for the "reply any time soon" bit

[02:10:01.0000] <Hixie>
that makes more sense

[02:10:02.0000] <tthorsen>
takkaria: But, for instance the nested forms issue on http://bankrate.com, i think you'd be able to see in the NetSurf browser too. It's certainly very visible if you parse the html with html5lib and open it in firefox.

[02:10:03.0000] <takkaria>
yeah, there's definitely something going wrong there

[02:11:00.0000] <tthorsen>
the page is laid out in two floated columns, and when the two columns don't end up with the same parent, they don't end up next to each other.

[02:12:00.0000] <takkaria>
/me nods, it looks a mess

[02:13:00.0000] <Hixie>
ok bed time

[02:13:01.0000] <Hixie>
nn

[02:53:00.0000] <tthorsen>
zcorpan: You say that the next version of Opera will deal with scoping like in the html5 spec; how does the new Opera like this markup:

[02:54:00.0000] <tthorsen>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cp%3EAudio1%3C%2Fp%3E%0A%3Cp%3E%3Cobject%3E%3Cp%3EX%3C%2Fp%3E%3C%2Fp%3E%0A%3Cp%3EAudio2%3C%2Fp%3E%0A%3Cp%3E%3Cobject%3E%3Cp%3EY%3C%2Fp%3E%3C%2Fp%3E%0A%0A

[02:55:00.0000] <tthorsen>
The current Firefox and Opera makes sense of this - IE produces something really strange, and the html5 parser puts the two objects inside one another

[03:23:00.0000] <hsivonen>
http://www.opensource.apple.com/darwinsource/10.5.5/autozone-77.1/README.html interesting

[03:32:00.0000] <Philip`>
hsivonen: Their explanation of "conservative" seems completely unrelated to what I'd usually understand conservative GC to be

[03:33:00.0000] <hsivonen>
Philip`: the usual conservative being related to guessing whether a word holds a pointer?

[03:33:01.0000] <Philip`>
(where I'd usually understand to mean that it doesn't always know whether some piece of memory is a pointer or an integer, so it conservatively assumes it's a pointer)

[03:34:00.0000] <hsivonen>
how does a garbage collector looking at the heap of a C++ process ever know that something is an integer and not a pointer?

[03:36:00.0000] <Dashiva>
Philip`: That's just half of it

[03:37:00.0000] <Dashiva>
The other half is that you can't move memory, because then you're either not updating pointers, or you are updating integers that aren't pointers, either way it's a loss

[03:38:00.0000] <Philip`>
hsivonen: In general, it can't; but you could restrict the language a bit, and make the compiler cleverer so it retains all the necessary type information at runtime, and then it could work

[03:39:00.0000] <Philip`>
(e.g. Java)

[03:40:00.0000] <hsivonen>
Philip`: sure, I get that it works with Java, but how does e.g. Boehm ever know what to collect?

[03:40:01.0000] <Philip`>
Dashiva: That seems to me like it's a consequence of conservatism, and not part of what conservatism itself actually means

[03:40:02.0000] <Philip`>
but I suppose that's just a matter of definition :-)

[03:41:00.0000] <hsivonen>
I guess I should read up on C++ garbage collection some time

[03:41:01.0000] <Philip`>
hsivonen: Boehm doesn't, so it's conservative

[03:41:02.0000] <hsivonen>
Philip`: so why does it ever collect anything?

[03:42:00.0000] <hsivonen>
ooh. never mind

[03:42:01.0000] <hsivonen>
/me thought about things the wrong way round

[03:42:02.0000] <Dashiva>
I'm more confused by how they manage to do useful data flow analysis in c/c++ compilers

[03:43:00.0000] <Philip`>
I guess it just scans the stack for pointer-like things, and then scans the pointed-at memory regions for pointer-like things, until it's run out of things to scan, or something like that

[03:43:01.0000] <Philip`>
Dashiva: Usually they don't, because aliasing breaks everything :-)

[03:43:02.0000] <Dashiva>
Philip`: Pretty much what you said

[03:50:00.0000] <Philip`>
C++ is a bit crazy when simply adding a "restrict" to a function argument can make your function go twice as fast

[03:50:01.0000] <takkaria>
I hope everyone's seen the closure syntax for C++0x

[03:51:00.0000] <takkaria>
([](int x) { cout << x; })(3); will be valid C++

[03:52:00.0000] <takkaria>
or maybe it's [](int x){ cout << x; }(3);, I can't remember

[03:53:00.0000] <Philip`>
That doesn't seem excessively crazy

[03:54:00.0000] <Philip`>
/me wasn't previously aware that C++0x had closure syntax

[03:54:01.0000] <takkaria>
"int main(void) { int x = 2; [=](){ x++; cout << x;}(); cout << x; }" will print "32" (the '=' means "capture by value inside the closure")

[03:55:00.0000] <zcorpan>
tthorsen: i get http://tinyurl.com/5ahcw7

[03:55:01.0000] <takkaria>
they've also added semi-automatic typing, such that "auto x = 3;" will declare x as an int

[03:55:02.0000] <Dashiva>
... why

[03:56:00.0000] <takkaria>
http://blogs.msdn.com/vcblog/archive/2008/10/28/lambdas-auto-and-static-assert-c-0x-features-in-vc10-part-1.aspx

[03:56:01.0000] <Dashiva>
Have we learned nothing from fortran?

[03:56:02.0000] <takkaria>
if anyone fancies a read, that's got examples of a bunch of crazy syntax

[03:56:03.0000] <Philip`>
Hmm, I suppose they're not really closures, because they don't hang onto stack frames and you'll just end up crashing if you try to return them from functions and aren't careful

[03:57:00.0000] <Philip`>
I've wanted something like "auto" every time I've had to write "for (std::map<std::string, std::string>::iterator = m.begin(); ...)"

[03:57:01.0000] <zcorpan>
tthorsen: ie6 doesn't allow nested <object>s

[03:58:00.0000] <zcorpan>
not sure what ie8 does

[03:58:01.0000] <zcorpan>
tthorsen: firefox does the scoping thing with <applet> i think

[04:00:00.0000] <hsivonen>
Dashiva: isn't that auto thing more like JS than Fortran?

[04:01:00.0000] <Dashiva>
hsivonen: Looks to me like it only autotypes the declaration, you can't change types on assignment 

[04:01:01.0000] <Philip`>
Dashiva: It's static typing, not like dynamic var in JS

[04:01:02.0000] <Philip`>
Uh

[04:01:03.0000] <hsivonen>
oh

[04:01:04.0000] <Philip`>
hsivonen: ^

[04:01:05.0000] <hsivonen>
doesn't seem useful

[04:02:00.0000] <Philip`>
As far as I'm aware, it just saves you having to type out the (often quite long, if it's using templates) type name on the LHS of a declaration when the compiler already knows the RHS's type

[04:02:01.0000] <Dashiva>
Apparently it's for storing lambdas

[04:03:00.0000] <hsivonen>
Philip`: as far as templated types go, I think this is a sign of a pretty severe problem: http://www.bdsoft.com/tools/stlfilt.html

[04:04:00.0000] <Philip`>
hsivonen: That's just a compiler implementation quality issue - they are all rubbish and print horribly verbose error messages :-(

[04:04:01.0000] <tthorsen>
zcorpan: It's interesting that you get that result. That _is_ certainly the same as what the html5 algorithm outputs.

[04:04:02.0000] <Philip`>
I got an infinitely long template error message once

[04:04:03.0000] <hsivonen>
Philip`: "just"? :-)

[04:04:04.0000] <Philip`>
(or at least the compiler generated a few hundred megabytes of error output before I killed it)

[04:05:00.0000] <zcorpan>
tthorsen: why is it interesting?

[04:06:00.0000] <Philip`>
hsivonen: There's no reason they couldn't include functionality equivalent to what stlfilt does

[04:07:00.0000] <Philip`>
/me has been working with someone who's written a compiler that converts certain arbitrarily-complex algebraic expressions into executable code, which works by stringing together a whole load of templated library functionality into a single C++ type that performs the whole computation

[04:07:01.0000] <tthorsen>
zcorpan: well, it's not backwards compatible with the current behaviour of any of the browsers

[04:08:00.0000] <Philip`>
and the error messages are crazy, but otherwise it's actually pretty nice, and works much better than the other approaches that have been tried

[04:08:01.0000] <hsivonen>
Philip`: sounds like a shotgun aimed at foot

[04:09:00.0000] <zcorpan>
tthorsen: it's pretty broken in ie though so probably pages don't depend on any particular behavior. or have you found otherwise?

[04:09:01.0000] <tthorsen>
no. I only have a testcase. I'm not sure if that testcase is based on something that was found in the wild.

[04:09:02.0000] <zcorpan>
tthorsen: we found a site before that depended on this behavior for <applet> (the markup was something like <p><applet></p>foo</applet> where the "foo" wasn't expected to be rendered)

[04:12:00.0000] <Philip`>
hsivonen: But it works :-)

[04:13:00.0000] <tthorsen>
zcorpan: wow, you'd think ordering those tags sanely wouldn't be rocket science...

[04:14:00.0000] <zcorpan>
tthorsen: well, it's the web we're talking about, remember :)

[04:15:00.0000] <hsivonen>
/me finds http://mxr.mozilla.org/mozilla-central/source/content/base/public/nsIDocument.h#729

[04:16:00.0000] <Philip`>
Even sane people will do something crazy if you give them a billion chances to mess it up :-)

[04:20:00.0000] <tthorsen>
Philip`: btw, I looked at the thing you posted yesterday with the "<code><pre>...</code></pre>...". It's kinda related to a problem I'm seeing myself

[04:21:00.0000] <tthorsen>
I've got "<abbr><p>abbr</abbr></p><acronym><p>acronym</acronym></p>" which doesn't work out quite right

[04:22:00.0000] <tthorsen>
Both seem to be related to step 3 of the "any other end tag" handling in "in body"

[04:23:00.0000] <Philip`>
I guess it only works correctly if you've got misnested formatting elements, or something?

[04:23:01.0000] <Philip`>
(<b><i>...</b></i> etc)

[04:24:00.0000] <Philip`>
/me goes away

[04:33:00.0000] <tthorsen>
Philip`: Yes. Removing step 3 completely (or replacing it with a step that makes sure we don't go out of bounds on the stack of open elements) fixes these cases, but I don't feel very confident it's the right fix...

[04:47:00.0000] <zcorpan>
is <code> a formatting element in webkit?

[04:48:00.0000] <zcorpan>
if you use <span> instead then webkit will nest

[04:50:00.0000] <zcorpan>
hmm firefox just closes <code> at <pre>

[04:51:00.0000] <tthorsen>
yeah. Opera and IE do the same thing, though

[04:52:00.0000] <tthorsen>
(at least my version of Opera does)

[04:54:00.0000] <zcorpan>
not quite, but yeah

[04:56:00.0000] <zcorpan>
(try to put some text between the end tags)

[05:08:00.0000] <tthorsen>
I think Opera's result looks nice, but I'm not sure how to write the html5 spec to mimic that. When we get to the </code> we can't pop our way out to the <code> because that would remove the <pre>.

[05:10:00.0000] <tthorsen>
This actually reminds me of what I proposed for both the nested forms case and the script handling in "after head" case. When we see the </code>, if there's a <code> in the stack of open elements, we just remove it.

[05:10:01.0000] <zcorpan>
tthorsen: what we do can't be mimicked without changing css

[05:11:00.0000] <tthorsen>
really? How does css affect this?

[05:11:01.0000] <zcorpan>
try http://software.hixie.ch/utilities/js/live-dom-viewer/?%3Ccode%20style%3Dcolor%3Ared%3E%3Cpre%3Ex%3C%2Fcode%3Ey%3C%2Fpre%3E

[05:13:00.0000] <tthorsen>
the change I proposed just now, would do just that, wouldn't it?

[05:14:00.0000] <zcorpan>
note that only the "x" is red, not the "y"

[05:18:00.0000] <tthorsen>
oh, right. That's really weird. I would have thought that css should have been done after, and independently from, the parser.

[05:19:00.0000] <tthorsen>
ie: <code> has style color: red, y ends up inside <code>, so y should be red

[05:20:00.0000] <zcorpan>
right

[05:21:00.0000] <tthorsen>
is this behaviour in Opera considered a bug or a feature?

[05:22:00.0000] <zcorpan>
well per html5 it's a bug

[05:23:00.0000] <zcorpan>
i think the closest you can get rendering-wise is to make <code> a formatting element

[05:25:00.0000] <tthorsen>
yeah. That certainly does produce the same result in our browser

[11:28:00.0000] <Hixie>
/me doesn't think it likely that we'll change the spec for <code><pre></code>

[11:28:01.0000] <Hixie>
really our only option is making <code> one of the quirky tags

[11:28:02.0000] <Hixie>
formatting tags

[11:28:03.0000] <Hixie>
and i'd rather not add more of those

[11:36:00.0000] <Hixie>
where on earth is zcorpan's dom core draft

[11:40:00.0000] <hsivonen>
Hixie: http://simon.html5.org/specs/web-dom-core

[11:40:01.0000] <Hixie>
thanks

[11:42:00.0000] <Hixie>
hm, his appendChild() definition doesn't prevent elements appended to text nodes

[11:42:01.0000] <Hixie>
oh well

[12:05:00.0000] <jgraham>
Does actual badness occur is <code> is a formatting tag?

[12:11:00.0000] <hober>
Seems like people can just write <pre><code>

[12:11:01.0000] <hober>
how often does <code><pre> even happen?

[12:42:00.0000] <Philip`>
Hixie: Not fixing <code><pre></code> is bad PR because it breaks early adopters :-(

[12:43:00.0000] <Philip`>
(specifically http://planet.intertwingly.net/ in "REST APIs must be hypertext driven")

[12:43:01.0000] <Hixie>
um, early adopters shouldn't use bad markup? :-)

[12:44:00.0000] <Philip`>
The early adopter didn't use bad markup; he used html5lib to sanitise other people's markup

[12:47:00.0000] <Philip`>
(under the assumption that html5lib would parse HTML in a way that's largely compatible with browsers; and browsers all handle <code><pre></code></pre> and get the 'correct' output)

[14:00:00.0000] <Hixie>
Philip`: well, there's not much we can do there since four browsers handle that markup in four different ways

[14:02:00.0000] <Philip`>
Hixie: That means there's four ways that all result in the paragraph after the code being rendered properly (not monospaced text etc), so there's plenty of choices for HTML5 to copy :-)

[14:23:00.0000] <Hixie>
Philip`: maybe. of course, changing this will just result in other bugs, the question is which bugs do we want.

[14:25:00.0000] <jgraham>
Hixie: The ones that browsers already have :)

[14:25:01.0000] <Hixie>
they differ

[14:29:00.0000] <jgraham>
Hixie: Well what I actually mean is that we want to pick something that will result in the rendering observed in browsers in cases where they are in agreement and try to limit our differences to cases where browsers differ

[14:30:00.0000] <jgraham>
(in agreement for rendering if not for the DOM)

[14:36:00.0000] <Hixie>
that's what we want to do, yes

[14:37:00.0000] <Hixie>
my point is that it's not clear that we're not already at a global minimum for that goal.

[14:39:00.0000] <Philip`>
It's clear that we're not, because you could rewrite the algorithm to say "if the document is equal to this multi-kilobyte string then return this particular DOM, else continue with the standard parser algorithm", which would result in the algorithm handling more pages correctly and wouldn't make any more be handled incorrectly

[14:58:00.0000] <jgraham>
The rendering of <b><code>foo</b>bar</code> from HTML5 seems to match Firefox but not Safari, Opera or IE7

[14:59:00.0000] <jgraham>
I don't remember if someone already mentioned that

[15:00:00.0000] <Hixie>
yes, because Opera and IE7 make all tags use the AAA or their equivalent, and Safari has a different list of tags that Firefox.

[15:00:01.0000] <Hixie>
HTML5 uses the intersection of the set of tags from firefox and safari

[15:02:00.0000] <hsivonen>
hmm. there are JavaFX questions on StackOverflow http://stackoverflow.com/questions/tagged/javafx

[15:03:00.0000] <jgraham>
Hixie: It seems like copying Safari would be good here

[15:03:01.0000] <Hixie>
so it is claimed :-)

[15:04:00.0000] <Hixie>
any change we make changes the total number of bugs, sometimes up, sometimes down. we only know of the bugs that are reported. the judgement we have to make is determining whether there are more bugs introduced by the proposed change that we don't know about than there are left if we don't make the change.

[15:05:00.0000] <hsivonen>
many more questions tagged silverlight http://stackoverflow.com/questions/tagged/silverlight not surprising

[15:05:01.0000] <hsivonen>
and even more Flash http://stackoverflow.com/questions/tagged/flash

[15:05:02.0000] <hsivonen>
even more html http://stackoverflow.com/questions/tagged/html

[15:06:00.0000] <jgraham>
It seems pretty likely that changes that make us render more like IE and fix issues with actual sites will move the bug count down rather than up. Do you have any specific concerns about the proposed changes?

[15:07:00.0000] <Hixie>
they make us more like IE for certain snippets, and less like IE for others.

[15:08:00.0000] <jgraham>
What becomes less like IE?

[15:09:00.0000] <Hixie>
there start being more <code> elements per <code> token in the DOM than there are in IE, for example.

[15:09:01.0000] <Hixie>
so scripts depending on a 1:1 mapping start failing

[15:10:00.0000] <Hixie>
it also changes the performance characteristics to be more expensive, so pages that IE might handle cheaply become expensive

[15:10:01.0000] <Hixie>
and so on

[15:10:02.0000] <Hixie>
in general we want to keep the number of formatting elements to an absolute minimum

[15:10:03.0000] <Hixie>
if we didn't, we'd just remove the difference between formatting elements and phrasing elements and just have them all do the AAA thing

[15:11:00.0000] <jgraham>
Isn't that generally true with the AAA stuff? It seems like there wouln't be much content that relied on that otherwise it would likely break in Firefox (albeit with different elements)

[15:11:01.0000] <jgraham>
(I am of course aware of the /topic)

[15:11:02.0000] <jgraham>
(that === 1:1 mapping)

[15:11:03.0000] <Hixie>
yes, it is true, and that's why both firefox and safari try to keep the list to a minimum too

[15:12:00.0000] <Hixie>
firefox doesn't have <code> in its list, you'll note


2008-11-13
[16:51:00.0000] <Hixie>
aboodman2: yt?

[16:52:00.0000] <Hixie>
aboodman2: cross-domain importScripts(), opinion?

[20:59:00.0000] <aboodman>
Hixie: ping?

[21:32:00.0000] <Hixie>
hey

[21:32:01.0000] <Hixie>
i was wondering if you knew why we'd made importScripts() block cross-domain imports

[23:03:00.0000] <heycam>
Hixie, you can now use [IndexGetter=NoFunction] (and similarly for [IndexSetter], [NameGetter] and [NameSetter]) to cause the function not to exist in ES

[23:19:00.0000] <Hixie>
heycam: did you come up with a solution to the issue of the objects in question actually having the properties or whatever the issue was?

[23:57:00.0000] <othermaciej>
Hixie: the distinction of "really existing" is meaningless

[23:57:01.0000] <othermaciej>
but

[23:57:02.0000] <othermaciej>
it's worth specifying in what cases magic getter properties are enumerable

[23:57:03.0000] <othermaciej>
I don't think that can be specified in terms of delegating to a function in the IDL

[23:57:04.0000] <othermaciej>
nor is it really sound to specify behavior of the "in" operator that way

[00:00:00.0000] <Hixie>
right, the enumerability is what i meant

[01:58:00.0000] <heycam>
Hixie, not yet

[01:58:01.0000] <heycam>
othermaciej, is it really meaningless?  it would change what Object.prototype.hasOwnProperty would return.

[01:59:00.0000] <heycam>
/me checks how enumeration of properties is defined in es3

[01:59:01.0000] <othermaciej>
heycam: I mean, there's no spec-level concept of "real" and "not real" properties

[02:00:00.0000] <heycam>
othermaciej, in which spec?

[02:00:01.0000] <othermaciej>
ECMA-262

[02:00:02.0000] <heycam>
no it only talks about properties

[02:00:03.0000] <heycam>
but magic [[Get]] will look like there are properties

[02:00:04.0000] <othermaciej>
I'm not sure it allows [[Get]] to be out of sync with [[HasProperty]] or hasOwnProperty

[02:01:00.0000] <heycam>
i think you can have whatever host object [[Get]] behaviour that you want, even if it is inconsistent with [[HasProperty]] etc.

[02:01:01.0000] <jwalden>
I think it must, because [[Get]] can be a getter function

[02:02:00.0000] <othermaciej>
I guess host objects can do anything

[02:02:01.0000] <othermaciej>
anyway

[02:02:02.0000] <heycam>
i think garrett showed that the properties were "real" though (for collections), so i'll be changing it to that

[02:02:03.0000] <othermaciej>
please don't use the term "real" or "actual"

[02:02:04.0000] <othermaciej>
it makes my skin crawl

[02:02:05.0000] <heycam>
ok :)

[02:02:06.0000] <othermaciej>
it doesn't mean anything

[02:02:07.0000] <heycam>
he showed that the objects had properties, rather than magic [[get]]

[02:02:08.0000] <othermaciej>
that doesn't mean anything either

[02:02:09.0000] <heycam>
?

[02:03:00.0000] <othermaciej>
he showed that the dynamic properties are reflected by [[HasProperty]] (and hasOwnProperty) as well as by [[Get]]

[02:04:00.0000] <othermaciej>
another interesting question is whether they are enumerable (or conversely whether they are DontEnum or not); I don't know if he tested that

[02:05:00.0000] <othermaciej>
I don't think the IndexGetter IDL syntax is sufficient to capture these issues

[02:05:01.0000] <heycam>
no there'll likely need to be something additional to say whether the properties are enumerable

[02:05:02.0000] <othermaciej>
even the [[HasProperty]] part can't properly be captured

[02:06:00.0000] <othermaciej>
(as far as I know)

[02:07:00.0000] <heycam>
if html5 needs some objects to have non-standard [[HasProperty]] behaviour, then no that can't be captured yet either

[02:07:01.0000] <othermaciej>
because any value that an IDL method could return is in theory a possible valid value of a function

[02:07:02.0000] <heycam>
what do you mean by that?

[02:09:00.0000] <heycam>
anyway i'll do some testing tomorrow and see if i come up with something sensible

[02:09:01.0000] <heycam>
/me tv

[02:10:00.0000] <zcorpan>
Hixie: appendChild() prevents elements from being appended to text nodes in http://simon.html5.org/specs/web-dom-core#legal-hierarchy

[02:12:00.0000] <heycam>
btw zcorpan i like how you can click on the <dfn>s in that document and get an index of what refers to it

[02:12:01.0000] <zcorpan>
heycam: you can do the same in the html5 spec

[02:13:00.0000] <hsivonen>
zcorpan: looks like this is a case of sequential readability vs. random access readability :-(

[02:13:01.0000] <heycam>
zcorpan, i can't get it to do that in html5 for me

[02:14:00.0000] <zcorpan>
heycam: the whatwg single-page version?

[02:15:00.0000] <heycam>
no i was looking at the multipage version

[02:15:01.0000] <zcorpan>
hsivonen: i guess in due course i might spec it in the algorithms in a way that matches how you'd sanely implement it

[02:16:00.0000] <hsivonen>
zcorpan: "The Node is a Document node and has zero or one child Element node. " could be clearer if it somehow made it unambiguous that the sentence doesn't restrict other types of children

[02:16:01.0000] <heycam>
/me cries as the single page version makes his opera unusable

[02:17:00.0000] <heycam>
/me goes to watch True Blood instead

[02:19:00.0000] <hsivonen>
heycam: two years ago when the spec was smaller, it worked fine in Opera 7 on Nokia 770 with 64 MB of RAM.

[02:19:01.0000] <hsivonen>
/me doesn't understand the perf complaints about the spec. It loads faster than many MXR pages.

[02:19:02.0000] <zcorpan>
hsivonen: true. is s/zero or one/no more than one/ clearer?

[02:20:00.0000] <hsivonen>
zcorpan: yeah

[02:23:00.0000] <zcorpan>
hsivonen: fixed

[02:24:00.0000] <Philip`>
The spec still works very badly for me in Opera (freezing for tens of seconds after loading); Firefox is much better

[02:26:00.0000] <hsivonen>
yeah. the situation with current spec and current opera isn't quite smooth

[02:33:00.0000] <Philip`>
Some Opera developer should look into it and work out how to fix the spec / Opera so that they cooperate nicely and smoothly, because it's bad PR if their browser can't render specifications :-)

[02:34:00.0000] <wilhelm>
/me looks.

[02:38:00.0000] <MikeSmith>
I heard that Opera is preconfigured now to do secret user-data collection about any user who accesses the HTML5 spec

[02:39:00.0000] <Philip`>
I don't mind it collecting user data

[02:39:01.0000] <Philip`>
(The only problem is if it starts sending that collection to someone)

[02:39:02.0000] <MikeSmith>
I was kidding of course

[02:40:00.0000] <MikeSmith>
the truth is that Firefox is configured to optimize the user experience for any user accessing the HTML5 spec

[02:40:01.0000] <Philip`>
Anyway they could just look at their web server access logs for the Opera icon that's used in the stability annotations

[02:41:00.0000] <MikeSmith>
I was kidding about Firefox too

[02:42:00.0000] <MikeSmith>
the truth is that the HTML5 spec is going secret user-agent sniffing and varying its user experience depending on UA

[02:45:00.0000] <Philip`>
Maybe Opera has some high-tech AI algorithm that detects you're reading the latest version of the spec, parses it for changes since the browser was initially programmed, and then automatically patches the browser so it's consistently up-to-date

[02:45:01.0000] <Philip`>
and that's what makes it freeze for a bit while loading the spec

[02:45:02.0000] <krijnh>
MikeSmith: you're full of kidding today :)

[02:53:00.0000] <Philip`>
/me resists the temptation to make a joke about what else MikeSmith might be full of, because that would be entirely unfair

[03:06:00.0000] <MikeSmith>
Philip`: I'm full of good will for humanity

[04:03:00.0000] <MikeSmith>
can any opera folk tell me if Arve is around somewheres today?

[04:07:00.0000] <wilhelm>
MikeSmith: He's not in his office.

[04:07:01.0000] <MikeSmith>
wilhelm: ok, thanks

[04:25:00.0000] <zcorpan>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cvideo%20src%3Dimage%3E

[04:25:01.0000] <zcorpan>
safari shows the image

[04:30:00.0000] <zcorpan>
hsivonen: http://livedom.validator.nu/?%3C!DOCTYPE%20html%3E%0D%0A%3Cp%3E%3Csvg%3E%3Cforeignobject%3E%3C%2Fp%3E is weird

[04:34:00.0000] <zcorpan>
in fact just <svg><foreignobject></p> is weird

[05:17:00.0000] <hsivonen>
zcorpan: isn't that per spec, though?

[08:00:00.0000] <mookid>
Is there any chance of having content-type negotiation in hyperlinks in html5?

[08:01:00.0000] <mookid>
i.e. a way of indicating what the Accept header should be for a link..

[08:01:01.0000] <mookid>
that might encourage browsers to be better HTTP clients :)

[08:02:00.0000] <mookid>
being able to indicate PUT and DELETE methods in html would be great too..

[08:03:00.0000] <mookid>
I guess you've already discussed this - what is your approach to this right now?

[08:13:00.0000] <Philip`>
Why not link to a URL that is specific to the desired content-type, instead of trying to modify Accept?

[08:14:00.0000] <Philip`>
(e.g. link to foo.txt or foo.html, instead of to foo with Accept:text/html)

[08:15:00.0000] <Philip`>
HTML5 allows <form method="PUT"> and <form method="DELETE">

[08:51:00.0000] <mookid>
Philip`: why? becuase that's how HTTP was designed :)

[08:52:00.0000] <mookid>
URL = resource locator, a resource has nothing to do with content-type

[08:53:00.0000] <mookid>
one resource can have multiple content-types, which one is returned should be evaluated according to the Accept header.. 

[08:54:00.0000] <mookid>
by specifying .html or .txt you are adding content-type to a URL, which was never the intention for HTTP

[08:54:01.0000] <Dashiva>
Yet it works

[08:55:00.0000] <mookid>
HTTP has been misused because of the failures of HTML and brwosers to conform to HTTP

[08:55:01.0000] <mookid>
HTML less so

[08:55:02.0000] <mookid>
but browsers use HTML as an excuse for not being proper HTTP clients

[08:56:00.0000] <Dashiva>
What would the benefit be?

[08:56:01.0000] <mookid>
benefit of conforming to HTTP? Oh I don't know.. consistency?

[08:56:02.0000] <takkaria>
sadly, whilst HTTP might be geared up for content negotiation and this entire "one resource, multiple representation" thing, the rest of the world just isn't

[08:56:03.0000] <mookid>
because HTML is crap

[08:57:00.0000] <mookid>
as are the browsers

[08:57:01.0000] <mookid>
you can change this with HTML5

[08:57:02.0000] <Dashiva>
Heh

[08:57:03.0000] <takkaria>
I meant more the operating systems, and the servers

[08:57:04.0000] <mookid>
rubbish lol

[08:57:05.0000] <mookid>
that's nonsense

[08:57:06.0000] <mookid>
no offence

[08:57:07.0000] <Dashiva>
They're violating HTTP left and right

[08:57:08.0000] <takkaria>
and the way most people use file extensions, and think of URLs as mapping to a directory tree on a server

[08:57:09.0000] <Philip`>
I like it when people tell me how to access all their data via URLs, rather than requiring me to wildly guess content-types until it happens to send me the right one :-)

[08:58:00.0000] <mookid>
so you're giong to leave out trivial.. potentially game changing improvements to HTML5 on teh basis that no one has been doing it properly?

[08:58:01.0000] <mookid>
superb.

[08:58:02.0000] <Dashiva>
Like what?

[08:58:03.0000] <mookid>
like an Accept tag in hyperlinks

[08:58:04.0000] <Dashiva>
What are these game changing improvements?

[08:58:05.0000] <Philip`>
Why is it game-changing? It's trivial for your server to map a single resource with multiple content-types onto multiple distinct URLs

[08:59:00.0000] <takkaria>
well, I'd sugget you're better trying to convince authors to see the world your way before you try and add features to a language which currently no-one wants to use

[08:59:01.0000] <mookid>
it's completely fundamental to enabling browsers to be good HTTP clients

[08:59:02.0000] <mookid>
what are you talking about HTTP is already established?

[08:59:03.0000] <mookid>
it's you guys and the brwoser guys not playing the game

[08:59:04.0000] <Dashiva>
http://www.example.com/resource with Accept:X can be mapped to http://www.example.com/resource?accept=X so I'm not seeing the big loss

[09:00:00.0000] <mookid>
there's a HUGE loss

[09:00:01.0000] <mookid>
If I tell you there's a resource

[09:00:02.0000] <mookid>
at say

[09:00:03.0000] <mookid>
example.com/something

[09:00:04.0000] <Dashiva>
There's a huge win: You don't have to specify anything for the huge majority of cases with only one type

[09:00:05.0000] <mookid>
you can put that into a browser.. you get HTML

[09:00:06.0000] <mookid>
you put that into iTunes you get the mp3

[09:00:07.0000] <mookid>
you put it into a feed reader you get the RSS

[09:00:08.0000] <mookid>
that's amazing..

[09:00:09.0000] <mookid>
if you do it your way the URL only serves one content-type

[09:01:00.0000] <mookid>
that's poop

[09:01:01.0000] <Dashiva>
Meanwhile the rest of the world has hyperlinks

[09:01:02.0000] <mookid>
what?

[09:01:03.0000] <Philip`>
How would I know that I should put that URL into iTunes or a feed reader?

[09:01:04.0000] <mookid>
you could do all or any

[09:01:05.0000] <mookid>
that's the point.

[09:01:06.0000] <mookid>
it's jsut a RESOURCE

[09:01:07.0000] <mookid>
URL..

[09:01:08.0000] <mookid>
a resource can have many representations

[09:01:09.0000] <mookid>
depending on what UA you use to access it

[09:01:10.0000] <Dashiva>
That wasn't one resource, that was two resources

[09:02:00.0000] <Philip`>
If I tried putting random URLs into iTunes in the hope that they were music, I'd be continually disappointed and would give up; so I'd only try it once you explicitly told me that this URL could be accessed as an MP3, in which case you might as well tell me an MP3-specific URL

[09:02:01.0000] <mookid>
a resource is not a specific thing it's an abstract concept

[09:02:02.0000] <takkaria>
argh, I hate discussions about ontologies and the Web, it never ends well

[09:02:03.0000] <mookid>
but this is nonsense

[09:02:04.0000] <mookid>
why are you rejecting this?

[09:02:05.0000] <mookid>
you could include an accept tag in a hyperlink

[09:02:06.0000] <Dashiva>
Because we can already do everything you have suggested

[09:03:00.0000] <takkaria>
and do it easier

[09:03:01.0000] <mookid>
oh that's just onsense

[09:03:02.0000] <mookid>
no you dont

[09:03:03.0000] <Dashiva>
Why bolt on extra complexity for no gain other than some abstract concept's purity?

[09:03:04.0000] <mookid>
it's not EXTRA complexity

[09:03:05.0000] <mookid>
EXTRA complexty is using a URL for somethign it wans't designed for

[09:03:06.0000] <Dashiva>
No, that's existing functionality

[09:03:07.0000] <mookid>
why don't you ask Roy Fielding he knows a thing or 2 abuot HTTP..

[09:03:08.0000] <Philip`>
The default stance towards features has to be rejection, because otherwise HTML would be more horrendously bloated than it is, and so convincing arguments are required before a feature will be accepted

[09:03:09.0000] <mookid>
no it's NOT

[09:04:00.0000] <mookid>
that's not functionality at all that's a dirty hack that's becoma  standard

[09:04:01.0000] <Dashiva>
Also known as functionality

[09:04:02.0000] <takkaria>
that's what most functionality in the world is, oddly enough :)

[09:04:03.0000] <Philip`>
That's what HTML is :-)

[09:04:04.0000] <mookid>
when you type in example.com/flibblwe.html - there is no gurantee whatsoever that html is returned..

[09:04:05.0000] <takkaria>
I think most states are built on dirty hacks, too

[09:04:06.0000] <mookid>
when you type in example.com/flibblwe.html - there is no gurantee whatsoever that html is returned..

[09:04:07.0000] <Dashiva>
There's no guarantee the world will exist tomorrow either

[09:04:08.0000] <takkaria>
but it's a reasonable guess

[09:04:09.0000] <mookid>
oh good argument.

[09:04:10.0000] <takkaria>
but if you type in example.com/flibblew then you really have no idea what it could be

[09:05:00.0000] <mookid>
yeah so is putting example.com/bla into ituens..

[09:05:01.0000] <mookid>
yes exatly

[09:05:02.0000] <mookid>
so you put it into whatever UA you have..

[09:05:03.0000] <Dashiva>
Why would I put bla into a music player? There's nothing telling me it's a song

[09:05:04.0000] <mookid>
and it tries to get the appropriate content type for you

[09:05:05.0000] <mookid>
in the case of a browser this is quite important because you're supposed to be able to GET lots of different content types

[09:05:06.0000] <Dashiva>
Now you're mixing two completely separate issues

[09:06:00.0000] <mookid>
no I'm not

[09:06:01.0000] <Dashiva>
User agents use accept all the time to get the right content

[09:06:02.0000] <Dashiva>
But putting accept into HTML elements is a different beast altogether

[09:06:03.0000] <Dashiva>
Besides, such accept info wouldn't travel with the URL when you enter it into iTunes

[09:07:00.0000] <mookid>
so you can't download a pdf.. or rss.. or xhtml.. or json with a brwoser?

[09:07:01.0000] <Dashiva>
(Oh look, invisible metadata)

[09:07:02.0000] <mookid>
header?

[09:07:03.0000] <mookid>
Accept is a header

[09:07:04.0000] <mookid>
it's a standard component of an HTTP transaction

[09:07:05.0000] <Dashiva>
As an attribute on a link, it is metadata

[09:07:06.0000] <mookid>
to tell the browser how to compose the request..

[09:08:00.0000] <mookid>
it's fundamental

[09:08:01.0000] <takkaria>
there are advantages to having the pdf and html versions at different urls

[09:08:02.0000] <takkaria>
because humans use urls

[09:08:03.0000] <Dashiva>
And humans usually use two different files

[09:08:04.0000] <takkaria>
and if you want to pass soeone the url to the PDF version, you paste "http://blah/file.pdf" at them

[09:08:05.0000] <takkaria>
without .pdf you'd be hard pushed to do that

[09:09:00.0000] <mookid>
I'm not saying that you souldn't be able to specify - I'm implementing an API that negotiates with either the Accept header, or if it's specified a ?type=application/pdf

[09:09:01.0000] <Dashiva>
So what's the problem then?

[09:09:02.0000] <mookid>
but browsers wont be able to be good HTTP clients until you provide markup that allows them to do this properly

[09:10:00.0000] <Dashiva>
Use type=whatnot in HTML, you support it

[09:10:01.0000] <mookid>
that doesn't change the accept header

[09:10:02.0000] <takkaria>
HTTP is broken in that regard, then, since the accept header is invisible to users when really it should be visible to then

[09:10:03.0000] <takkaria>
so HTML is just working around a broken transport layer

[09:10:04.0000] <mookid>
HTTP ISNT BROKEN

[09:10:05.0000] <mookid>
HTTP and browsers are broken

[09:10:06.0000] <mookid>
HTML^

[09:10:07.0000] <takkaria>
it's certainly no platonic form

[09:11:00.0000] <Philip`>
(HTTP servers are broken too)

[09:11:01.0000] <Dashiva>
Don't take it personally, it's just a protocol :)

[09:11:02.0000] <mookid>
HTTP servers arent broken?

[09:11:03.0000] <mookid>
it's an important protocol

[09:11:04.0000] <Philip`>
(I particularly like the one that returns four random bytes for Content-Type when you send a HEAD request)

[09:11:05.0000] <Dashiva>
Philip`:  How about the one that returns text/plain on unknown content-types :)

[09:11:06.0000] <mookid>
and it's completely miss-represented and miss-understood because people get all political and dont want to admit that stuff like HTML and browsers are crap at speaking HTTP

[09:12:00.0000] <mookid>
so you've got an opporunity to bridge the gap..

[09:12:01.0000] <mookid>
and instead of taking stuff like this onboard

[09:12:02.0000] <mookid>
you get defensive

[09:12:03.0000] <Philip`>
Dashiva: That's just boring brokenness, not fun brokenness

[09:12:04.0000] <mookid>
and nothign gets fixed

[09:12:05.0000] <Dashiva>
Who is defensive again?

[09:12:06.0000] <mookid>
but you get to feel good about yourself because there's a new version of HTML

[09:12:07.0000] <mookid>
I'm trying to make the web work better

[09:12:08.0000] <Dashiva>
No you're not

[09:12:09.0000] <mookid>
you're trying to be on a document for HTTP5

[09:13:00.0000] <Dashiva>
You're trying to make HTTP look better

[09:13:01.0000] <mookid>
what?

[09:13:02.0000] <mookid>
I have no affiliation with HTTP whatsoever

[09:13:03.0000] <Dashiva>
You haven't mentioned a single thing that would benefit the web

[09:13:04.0000] <mookid>
of HTTP?

[09:13:05.0000] <mookid>
are you insnae?

[09:13:06.0000] <mookid>
isnane^

[09:13:07.0000] <Dashiva>
Just pragmatic

[09:13:08.0000] <mookid>
no you're misguided

[09:14:00.0000] <mookid>
are there project leads here because this is pretty important and from what I can gather not one of you appreciates any of these points

[09:14:01.0000] <mookid>
I'm in the real world developing real world solutions and HTTP works perfectly.. brwosers and HTML are attrocious

[09:15:00.0000] <Dashiva>
You could send an email to the mailing list

[09:15:01.0000] <Dashiva>
Hixie will be sure to read it

[09:15:02.0000] <Philip`>
I think most people here would agree that browsers and HTML have significant problems; the disagreement is usually about how to best move forwards from where we are

[09:16:00.0000] <mookid>
well I would say that looking into HTTP  - respecting it and according to it

[09:16:01.0000] <mookid>
would be a good start

[09:16:02.0000] <mookid>
and "oh well HTTP can still work and we dont need to provide a mechanism for setting an Accept header" is frankly not good enough

[09:16:03.0000] <mookid>
considering how important HTML is

[09:17:00.0000] <mookid>
I really hope your approach is not reflective of this project

[09:18:00.0000] <mookid>
Hixie: what are your thoughts on this?

[09:19:00.0000] <Dashiva>
He's probably still asleep

[09:20:00.0000] <mookid>
Do you not think an optional tags for appropriate HTTP requests is even a possibility?

[09:20:01.0000] <mookid>
-an

[09:20:02.0000] <Dashiva>
I predict Hixie will say something along the lines of "What are the use cases?"

[09:21:00.0000] <mookid>
great..l so he can nerd me into position

[09:21:01.0000] <Philip`>
The HTML5 approach is (in theory) to start by considering use cases, like "person X wants to send a link to a music file to a friend who can listen to it" or whatever, and then considering various solutions (e.g. send a multi-type URL and make UAs use Accept correctly, or send a URL specific to the MP3, or whatever) and choosing the 'best' (usually simplest, given the cost of adding features to browsers) that satisfies the use cases

[09:21:02.0000] <mookid>
but what if the benefits are mostly external to HTML itself

[09:21:03.0000] <mookid>
then it's not considered

[09:22:00.0000] <mookid>
my argument is that it's better for HTTP and ultimately better for the web

[09:22:01.0000] <mookid>
how do I make that a use case for HTML itself?

[09:22:02.0000] <mookid>
your argument will alwasy be 'well you can put the content type in the URL'

[09:22:03.0000] <mookid>
if you look at the definition of URL though.. it's pretty clear that shouldn't be the approach

[09:23:00.0000] <mookid>
the fact that you can doesnt make it right

[09:23:01.0000] <mookid>
e.g. I can spam aload of ranting jibberish at you - that doesnt mean I'm right

[09:23:02.0000] <Philip`>
Whenever a feature is added to the spec, the browser developers will have to be convinced that it's worth implementing, and they're unlikely to do that if the only justification is some abstract "it's better for the web" and it doesn't directly make their users happier

[09:23:03.0000] <Dashiva>
Say, mookid

[09:23:04.0000] <Dashiva>
Did you read the discussion about HTML URLs?

[09:23:05.0000] <mookid>
it's not HTML urls

[09:23:06.0000] <mookid>
it's just URLs

[09:23:07.0000] <Dashiva>
Oh boy

[09:23:08.0000] <mookid>
what? it's not..

[09:24:00.0000] <Dashiva>
Philip`: Can I point him there in good conscience?

[09:24:01.0000] <mookid>
is that not allowed because it makes you feel less important?

[09:24:02.0000] <takkaria>
HTML has its own concept of how to map and URI expressed in HTML to a real URI

[09:25:00.0000] <Philip`>
Dashiva: You mean the discussion about whether HTML5 should call its URL-ish syntax "URLs" instead of something ugly like "HURLs"?

[09:25:01.0000] <mookid>
why?

[09:25:02.0000] <Philip`>
/me isn't quite sure why that's at all relevant

[09:25:03.0000] <mookid>
why not just follow what the HTTP protocol tells you

[09:25:04.0000] <mookid>
it's pretty specific

[09:25:05.0000] <mookid>
R = Resource

[09:25:06.0000] <Dashiva>
Philip`: It's another violation of the HTTP/URL purity :)

[09:25:07.0000] <mookid>
R != Resource + content-type

[09:26:00.0000] <mookid>
even if that is the way things are done now

[09:26:01.0000] <mookid>
and I'm not even saying 'stop people from doing it'

[09:26:02.0000] <mookid>
I'm just saying give them the ability to do it right

[09:26:03.0000] <mookid>
and let them choose

[09:26:04.0000] <Philip`>
Dashiva: Not really - it's just a syntactic issue about how you translate a stream of bytes in an HTML document into a requestable URI

[09:26:05.0000] <mookid>
a few optional tags isn't exactly rocket science

[09:27:00.0000] <mookid>
I guess I'm not the first person to bring this up right?

[09:28:00.0000] <takkaria>
AFAICR, you are

[09:28:01.0000] <jmb>
mookid: why does html have to add an attribute to link elements? user agents know what content types they support, so they send an accept header. it's then up to the server to serve something up that's appropriate. the exact same thing happens right now with language negotiation

[09:28:02.0000] <mookid>
jmb that's not true for browsers actually

[09:28:03.0000] <mookid>
brwosers handle lots of different content types

[09:28:04.0000] <mookid>
not just HTML

[09:28:05.0000] <mookid>
*chock horror* !

[09:28:06.0000] <mookid>
shockl^

[09:28:07.0000] <mookid>
chock^

[09:28:08.0000] <mookid>
shock^

[09:28:09.0000] <takkaria>
that's exactly what he said...

[09:28:10.0000] <jmb>
mookid: yes, and they send an appropriate Accept header

[09:28:11.0000] <mookid>
no they don't

[09:29:00.0000] <Philip`>
My browser says it accepts */*

[09:29:01.0000] <mookid>
they send a generic header

[09:29:02.0000] <mookid>
that's what I'm saying

[09:29:03.0000] <mookid>
that's not how HTTP was intended

[09:29:04.0000] <mookid>
that's how it's been implemented

[09:29:05.0000] <Philip`>
(Specifically: "text/html, application/xml;q=0.9, application/xhtml+xml, image/png, image/jpeg, image/gif, image/x-xbitmap, */*;q=0.1")

[09:29:06.0000] <mookid>
because HTML and browsers are crap

[09:29:07.0000] <mookid>
..

[09:29:08.0000] <mookid>
that's exaclty my point.

[09:29:09.0000] <mookid>
!

[09:29:10.0000] <jmb>
what has html to do with this?

[09:29:11.0000] <mookid>
oh deasr lord.

[09:30:00.0000] <mookid>
<a> = GET request to a URL - it shoudl indicate exactly how to construct that request

[09:30:01.0000] <mookid>
but at the moment it doesnt

[09:30:02.0000] <jmb>
no

[09:30:03.0000] <mookid>
and apparently the new version wont either

[09:30:04.0000] <mookid>
which is sad

[09:30:05.0000] <mookid>
because you have an oppotunity

[09:30:06.0000] <jmb>
the page author has utterly no way of knowing what the user agent supports

[09:30:07.0000] <mookid>
and a protocol to guide you

[09:31:00.0000] <mookid>
the Accept HEADER

[09:31:01.0000] <mookid>
that's the point

[09:31:02.0000] <jmb>
therefore, they have no reason whatsoever to be given a way to change the user agent's accept header

[09:31:03.0000] <mookid>
yes there is

[09:31:04.0000] <mookid>
if I request from my browser example.com/report

[09:31:05.0000] <mookid>
and I want that page to link to the pdfversion

[09:31:06.0000] <mookid>
and not the html

[09:32:00.0000] <mookid>
how am I supopsed to tell the brwoser to ask for pdf ?

[09:32:01.0000] <mookid>
your sugegstion is to us ethe URL

[09:32:02.0000] <mookid>
URL's arent for that

[09:32:03.0000] <mookid>
YES you can do it

[09:32:04.0000] <jmb>
if the browser can't handle PDF, then why should it ask for it?

[09:32:05.0000] <Philip`>
mookid: Just to be clear: Are you suggesting that browsers should support something like <a href="..." accept="audio/mp3, audio/wav;q=0.9">? and do you just want Accept, or all the other HTTP request headers that you currently can't access?

[09:32:06.0000] <mookid>
yes exactly

[09:32:07.0000] <mookid>
exactly Philip` 

[09:32:08.0000] <mookid>
so generic if not specified

[09:33:00.0000] <mookid>
for that browser/OS

[09:33:01.0000] <Philip`>
mookid: Just for Accept, not anything else?

[09:33:02.0000] <mookid>
if specified it tells brwoser to specify header

[09:33:03.0000] <mookid>
well Accept is the only one I can think of that has practical use to the user

[09:33:04.0000] <mookid>
there may be others

[09:34:00.0000] <mookid>
are you talking to the HTTP community abuot this?

[09:34:01.0000] <mookid>
it's quite important actually

[09:34:02.0000] <mookid>
Philip`: are you seeign where I am coming from?

[09:34:03.0000] <mookid>
or not?

[09:34:04.0000] <Philip`>
mookid: Should it only apply to <a href>, and not to <img src> or <script src> or <link href> etc?

[09:35:00.0000] <mookid>
right.. any hyperlink

[09:35:01.0000] <mookid>
that's not a hyperlink is it

[09:35:02.0000] <mookid>
I dont know the word

[09:35:03.0000] <mookid>
anythign containg an URL refrence

[09:36:00.0000] <mookid>
<a> would be the most comomon though obviously

[09:36:01.0000] <mookid>
and if developers didn't use it - it would make no difference to them anyway

[09:37:00.0000] <mookid>
+ it's not a binding contact for the server, the protocol doesn't require the server to actually respect the Accept header - in the same way that I can make my server respond to index.html with a jpeg

[09:39:00.0000] <takkaria>
the main point here, is that it provides no tangible benefit to allow people to do that over other methods that are already widely used, and so it would be a very low priority in specification development

[09:39:01.0000] <mookid>
*shakes head*

[09:39:02.0000] <mookid>
lordy lordy lordy

[09:40:00.0000] <mookid>
it provides massive benefits

[09:40:01.0000] <Philip`>
mookid: I do see what you're proposing, but I don't currently see that the cost of implementing it would be worth the value (particularly since it seems it only has any value once it's implemented in nearly every deployed web browser, because otherwise people won't be able to safely use it)

[09:40:02.0000] <mookid>
do you know who Roy Fielding is?

[09:40:03.0000] <takkaria>
yes

[09:40:04.0000] <takkaria>
do you know who Tony Blair is?

[09:40:05.0000] <mookid>
he agrees with me.

[09:41:00.0000] <mookid>
Roy Fielding is pretty authorative guy when it comes to the web

[09:41:01.0000] <takkaria>
appeals to authority never work in the WHAT WG, I'm afraid

[09:41:02.0000] <Dashiva>
Argument from authority?

[09:41:03.0000] <mookid>
er

[09:41:04.0000] <mookid>
I'm sorry but

[09:41:05.0000] <mookid>
how can you make sweeping statements like "this isnt high enough priority"

[09:41:06.0000] <mookid>
you must assume some kind of authority within the realm of HTML to make that assertion

[09:42:00.0000] <mookid>
oops? did you just make an internet boo boo ?

[09:42:01.0000] <Dashiva>
The design principles?

[09:42:02.0000] <mookid>
There is just so much potential with a new version of HTML

[09:42:03.0000] <mookid>
yuor gonna ruin it if you ignore this

[09:43:00.0000] <mookid>
why don't you start a dialogue with the HTTP guys and see what the feedback is?

[09:43:01.0000] <Philip`>
The HTTP guys' feedback has mostly been objecting to aspects of HTML5 that can't be changed because of compatibility requirements on the web, if I remember correctly

[09:44:00.0000] <mookid>
well this isn't a compatablity requirements

[09:44:01.0000] <mookid>
issue

[09:44:02.0000] <mookid>
wahtever coming back later this is insane

[09:44:03.0000] <mookid>
:D

[09:45:00.0000] <Dashiva>
Philip`: Also refusing to change things that clash with the web

[09:52:00.0000] <Philip`>
(From #html-wg: http://www.w3.org/html/wg/markup-spec/ )

[09:53:00.0000] <Dashiva>
obsolete...

[09:55:00.0000] <aboodman>
Hixie: not sure if you got last message... I would lean toward doing what the web already does, but that is not based on any sort of analysis

[10:04:00.0000] <Philip`>
Dashiva: Oh, didn't see it on the list

[10:07:00.0000] <Hixie>
aboodman: k

[10:08:00.0000] <Hixie>
aboodman: (i remember that we added the cross-site limitations for a reason, i just don't remember if the reason was "to do the right thing" or "because it prevents attack X")

[11:32:00.0000] <Lachy>
my presentation at pubcon is over

[11:33:00.0000] <MikeSmith>
ぁ

[11:33:01.0000] <MikeSmith>
Lachy: お疲れさま

[11:34:00.0000] <Lachy>
surprisingly I stuttered a lot more than usual this time

[11:34:01.0000] <Lachy>
MikeSmith, wtf?

[11:34:02.0000] <MikeSmith>
Lachy: how did it go? how many people in the audience?

[11:34:03.0000] <MikeSmith>
otsukare-sama

[11:34:04.0000] <Lachy>
I'd guess around 50 to 60

[11:34:05.0000] <Lachy>
though I'm not sure. Could have been more

[11:34:06.0000] <MikeSmith>
this was an HTML5 presentation?

[11:35:00.0000] <Lachy>
it was a panel on HTML and CSS. 3 of us gave 15 min presentations each

[11:36:00.0000] <Lachy>
I should have spoken more about forms instead of things like <section>, since forms would have been more relevant to this business and marketing oriented crowd

[11:36:01.0000] <Lachy>
but I didn't have any decent slides about forms to reuse

[11:36:02.0000] <Lachy>
I will have to make them

[11:37:00.0000] <Lachy>
anyway, I went and saw Steve Wyrick, a magician, last night.

[11:38:00.0000] <Lachy>
he was good. I really have no idea how he managed to get from the stage to a seat in the middle of the audience

[11:39:00.0000] <Lachy>
but the person he was sitting next to got a huge surprise :-)

[11:41:00.0000] <Philip`>
Maybe it was his identical twin

[11:41:01.0000] <Philip`>
or an illegal cloning experiment

[11:44:00.0000] <Lachy>
I also found a cool magic shop, and learned how to make a levitating playing card

[11:47:00.0000] <MikeSmith>
Lachy: Pubcon is the same organizers as Webmaster World?

[11:47:01.0000] <Lachy>
yes

[11:48:00.0000] <MikeSmith>
anybody else from Opera was speaking there? or around for the event?

[11:49:00.0000] <Lachy>
Gier from the web apps dept is here somewhere, but not presenting

[11:52:00.0000] <MikeSmith>
Lachy: please tell him I said hi. I've not seem him since Xtech a couple years ago

[12:51:00.0000] <jgraham>
Being able to set HTTP headers used in hyperlinks from HTML seems to have a very poor backward compaibility story

[12:59:00.0000] <Dashiva>
jgraham: No, no. HTTP has supported headers for many years ;)

[13:05:00.0000] <jgraham>
:-p

[13:14:00.0000] <gsnedders>
It opens up all kinds of attack vectors

[13:18:00.0000] <jcranmer>
is scheme:path?query;param#anchor the correct format, or is it scheme:path;param?query#anchor ?

[13:19:00.0000] <Dashiva>
I believe it's the latter

[13:37:00.0000] <hsivonen>
gsnedders: I think Macrovision on VHS isn't *DRM*, because it lacks the D

[13:37:01.0000] <gsnedders>
hsivonen: Pedant :)

[13:38:00.0000] <hsivonen>
gsnedders: it seems to me that the argument is:

[13:39:00.0000] <hsivonen>
 1) Font foundries want to make infringing people aware that they are infringing

[13:40:00.0000] <hsivonen>
 2) People grab a font using view source / Web Inspector / Firebug don't realize they are infringing

[13:41:00.0000] <hsivonen>
so the argument doesn't hinge on any technical prevention

[13:41:01.0000] <hsivonen>
but on what hoops make people aware that they are infringing when they infringe

[13:41:02.0000] <gsnedders>
Then any obstruction or DRM is overkill

[13:42:00.0000] <hsivonen>
s/grab/who grab/

[13:43:00.0000] <gsnedders>
Would it not be simpler for browsers to include a small note in all save boxes, along the lines of, "This file may be subject to copyright/licence restrictions."

[13:44:00.0000] <jgraham>
I don't see why that is needed

[13:45:00.0000] <jgraham>
or how it would help, really

[13:45:01.0000] <gsnedders>
Nor do I really, but it seems to address the needs of the foundries

[13:46:00.0000] <hsivonen>
jgraham: the whole discussion is basically about what rituals would make one group of believe that another group of people is lead to think in a certain way

[13:47:00.0000] <jgraham>
hsivonen: Yeah, but putting randon pseudo-legalese in browsers is really not nice

[13:49:00.0000] <jgraham>
I still don't understand why fonts deserve more protection than photographs

[13:49:01.0000] <hsivonen>
jgraham: I agree. Particularly when the pseudo-legalese is a truism that should be assumed to be known by default.

[13:49:02.0000] <Hixie>
"<gsnedders> Nor do I really, but it seems to address the needs of the foundries" assumes that we need to address those needs

[13:49:03.0000] <Hixie>
which i'm pretty sure we don't

[13:50:00.0000] <gsnedders>
Hixie: Just ignore the font foundries?

[13:50:01.0000] <gsnedders>
Hixie: Surely that's no different to ignoring implementers?

[13:50:02.0000] <jgraham>
gsnedders: That seems like an option.

[13:50:03.0000] <gsnedders>
(in terms of getting stuff used)

[13:52:00.0000] <Hixie>
implementors can veto stuff because if they don't implement, the spec is fiction

[13:52:01.0000] <Hixie>
font foundries have no veto ability here, and their needs directly conflict with the users'

[13:52:02.0000] <Hixie>
users win

[13:52:03.0000] <hsivonen>
Hixie: what happened with Google's EOT patch for WebKit?

[13:53:00.0000] <Hixie>
we decided we couldn't justify introducing DRM given our "do no evil" principles

[13:53:01.0000] <gsnedders>
Hixie: No, users don't win. They can't use the majority of fonts.

[13:53:02.0000] <Hixie>
sure they can

[13:54:00.0000] <gsnedders>
(and if you just send OTF/TTF anyway, who's betting the foundries will just change their licences to disallow that)

[13:54:01.0000] <Hixie>
of the fonts available to a random author, the vast majority are free

[13:54:02.0000] <hsivonen>
Hixie: it seems to me that the "need" of foundries is to believe that users know what they are infringing. I'm not convinced that in would be against the users' needs to know what behaviors are infringing. However, I'd prefer a way to establish the belief and awareness without using technology as a intermediate.

[13:55:00.0000] <Hixie>
users are fully aware that copying a font from another site is a violation of copyright

[13:55:01.0000] <Hixie>
so that need is already met

[13:55:02.0000] <jgraham>
hsivonen: That is hardly a unique need. Why do fonts get special consideration over all other media

[13:58:00.0000] <hsivonen>
Hixie: yeah, according to a recent study conducted in Finland, 95% of school kids believe that distributing movies and music without authorization is illegal.

[13:59:00.0000] <Hixie>
people know it's illegal, they just don't care

[14:04:00.0000] <Dashiva>
hsivonen: You say "believe", makes it sound like it isn't :)

[14:06:00.0000] <hsivonen>
Dashiva: well, I'm pretty sure it's illegal here, but these studies also include BS about using the word that approximates "felony" for things that are illegal but not punishable

[14:06:01.0000] <hsivonen>
(compare with riding a bike without a helmet)

[14:07:00.0000] <Dashiva>
The jail/no jail distinction?

[14:08:00.0000] <hsivonen>
Dashiva: for example, I'm pretty sure that downloading movies / music without authorization is in the illegal but not punishable category here.

[14:08:01.0000] <hsivonen>
Dashiva: but the people who get prosecuted also shared the content

[14:08:02.0000] <Philip`>
If it's not punishable, what would stop anyone from doing it?

[14:09:00.0000] <hsivonen>
Philip`: knowing that it illegal. And for all things not copyright-related, Finns will generally obey the law on the law's sayso

[14:10:00.0000] <hsivonen>
Also, over here the police can do pretty much what it pleases and can use equipment confiscation as a de facto punishment on a presumed guilty-basis.

[14:13:00.0000] <Philip`>
Given how many people infringe copyright even when they know it's illegal, I can't imagine many are put off by it being considered legally naughty but not punishable

[14:13:01.0000] <hsivonen>
I would guess that for people who are very familiar with the law and practice, the risk of equipment confiscation without judicial oversight is a stronger deterrent than the law

[14:19:00.0000] <hsivonen>
(oh. there's another thing that's illegal but not punishable and that people do: lying about campaign finance to the ministry of justice)

[14:20:00.0000] <hsivonen>
(the politicians had the good sense to make their own illegalities non-punishable)

[14:22:00.0000] <Dashiva>
I don't get that

[14:22:01.0000] <Dashiva>
I learned in class that illegal was defined as there being a law against it

[14:24:00.0000] <hsivonen>
Dashiva: right. But the law doesn't have to set punishments for all things that are illegal.

[14:25:00.0000] <hsivonen>
Dashiva: and over here if an illegal thing doesn't have an entry in the Criminal Code, it's a different type of illegality than a "crime"

[14:27:00.0000] <Dashiva>
So what happens if the police find you committing one of those crimes? They take your computer to gather evidence, and then throw the evidence away?

[14:29:00.0000] <Hixie>
copyright violation is punishable

[14:29:01.0000] <Hixie>
it's just a civil offense rather than a criminal one

[14:29:02.0000] <Hixie>
which means financial damages and reparations, generally

[14:30:00.0000] <hsivonen>
Hixie: Finland doesn't have punitive damages, so in theory, if it's a civil case, it's not a punisment but just damages

[14:30:01.0000] <hsivonen>
*punishment

[14:30:02.0000] <Hixie>
and what's the nominal "damage" of violating copyright?

[14:32:00.0000] <hsivonen>
Dashiva: they can take the computers and all inessential related equipment away for investigation, be slow with the investigation, and turn over the evidence to support a trial brought by a RIAA/MPAA/IFPI-connected civil association

[14:32:01.0000] <hsivonen>
Hixie: the retail price of the same piece of content on CD/DVD

[14:33:00.0000] <hsivonen>
Hixie: assuming a theory that every copy is a lost sale, which is of course bogus

[14:33:01.0000] <Hixie>
wow, that's low

[14:33:02.0000] <jcranmer>
hsivonen: better than the RIAA's theory

[14:33:03.0000] <Hixie>
the us has insanely high numbers

[14:33:04.0000] <Hixie>
like, ever violation is 5 digits usd

[14:33:05.0000] <hsivonen>
Hixie: plus legal fees!

[14:33:06.0000] <jcranmer>
i.e., every copy represents thousands of lost sales

[14:33:07.0000] <Hixie>
every, even

[14:34:00.0000] <hsivonen>
so if you lose a case to the enforcement association, the court orders you to pay their legal fees

[14:34:01.0000] <hsivonen>
so for the downloader case, the main risk is equipment confiscation and legal fees

[14:35:00.0000] <hsivonen>
for the uploader case, there's also a potentially high supposed damage

[14:35:01.0000] <hsivonen>
none of these are "punishments" in the legal sense although de facto, they are

[14:38:00.0000] <hsivonen>
not having punitive damages is good. If something requires more punishment, a fine to the state leads to much more healthy behavior than awarding more than the actual damages to the supposedly injured party

[14:39:00.0000] <hsivonen>
(the reason why we don't have punitive damages is a principle that the injured party shouldn't end up better off than where they started by suffering a damage)

[14:41:00.0000] <Hixie>
you'll note i didn't mention punitive damages :-)

[14:41:01.0000] <Hixie>
i don't think the us has them for copyright violations either

[14:41:02.0000] <Hixie>
but what do i know

[14:41:03.0000] <hsivonen>
Hixie: well, the U.S. has crazy statutory damages

[14:41:04.0000] <Dashiva>
It's not like the actual law matters to the RIAA :P

[14:41:05.0000] <hsivonen>
I'm not sure if those are considered "punitive" legally in this case

[14:44:00.0000] <jruderman>
how about cruel and unusual?

[14:44:01.0000] <hsivonen>
anyway, damages in excess of the actual damages whether statutory or made up be a judge on the spot suck, and it's better to put the excess into the state treasure than to award it to another litigant, IMO

[14:45:00.0000] <Hixie>
statutory damages kinda makes sense in that for non-financial stuff there has to be some sort of going rate for pain

[14:45:01.0000] <Hixie>
i mean, it's not like you can easily measure the pain cause by slander in USD, for example

[14:46:00.0000] <jruderman>
/me slanders Hixie's cat

[14:46:01.0000] <Hixie>
poor mowmow

[14:46:02.0000] <Hixie>
they were at the vet this morning

[14:46:03.0000] <jcranmer>
/me has a cat who likes to sit and sleep

[14:46:04.0000] <jcranmer>
very good paperweight

[14:46:05.0000] <hsivonen>
jruderman: it seems to me that in RIAA cases damages are cruel but unfortunately not unusual

[14:48:00.0000] <hsivonen>
s/treasure/treasury/


2008-11-14
[17:42:00.0000] <Hixie>
81 e-mails on the subject of the video looping attributes??!

[17:42:01.0000] <Hixie>
are you kidding me?

[17:42:02.0000] <Hixie>
man you people need to stop repeating yourselves. :-P

[17:43:00.0000] <Dashiva>
It's not our fault nobody specified a loop attribute (header) for email

[17:45:00.0000] <Hixie>
use cases:

[17:45:01.0000] <Hixie>
looping a background audio track

[17:45:02.0000] <Hixie>
playing a fragment when treating a media file as a "sprite" to avoid the overhead of multiple Audio objects and HTTP connections

[17:46:00.0000] <Hixie>
jumping to a starting point without having to wait for the video file to load

[17:47:00.0000] <MikeSmith>
+ a background video (similar to a background image); e.g., that just asks as the "idle screen" for kiosk or whatever

[17:47:01.0000] <Hixie>
read "audio or video" whenever i say "audio" or "video"

[17:47:02.0000] <MikeSmith>
OK

[17:48:00.0000] <Hixie>
(they're the same api)

[22:50:00.0000] <Hixie>
/me summons roc

[23:11:00.0000] <Hixie>
sicking: yeah i don't remember what the use case was either... hopefully aaron will clarify

[23:38:00.0000] <zcorpan>
Hixie: thanks for dropping those looping features... it makes my job a bit easier

[00:09:00.0000] <jwalden>
Hixie: excellent pun about people repeating themselves on the looping attributes!  :-D

[01:07:00.0000] <hsivonen>
Can some explain me why software dealing with TV-originating data should crop away overscan as suggested in the pixel aspect ratio thread?

[01:08:00.0000] <hsivonen>
(I, for one, don't see the point of content getting cropped when exporting from EyeTV. If the uncropped version didn't look "right", why does EyeTV support displaying the incoming TV feed with all the pixels visible?)

[03:00:00.0000] <famicom>
hsivonen i actually knew that

[03:00:01.0000] <hsivonen>
famicom: what are you referring to?

[03:01:00.0000] <famicom>
<hsivonen> Can some explain me why software dealing with TV-originating data should crop away overscan as suggested in the pixel aspect ratio thread?

[03:02:00.0000] <hsivonen>
famicom: so what's the point of cropping these days when displays don't need overscan for CRT tech?

[03:05:00.0000] <famicom>
software or hardware decoder

[03:08:00.0000] <hsivonen>
famicom: either way. what's the point?

[03:12:00.0000] <famicom>
eh

[03:12:01.0000] <famicom>
ask me some time when i've had some rest

[03:24:00.0000] <BenMillard>
jgraham, any help you can offer with Action-85 (http://www.w3.org/html/wg/tracker/actions/85) would be much appreciated :)

[03:25:00.0000] <BenMillard>
at the moment I'm documenting the most general similarities they have, such as applying header cells downwards or rightwards and limiting this to <th> cells

[03:27:00.0000] <BenMillard>
take your time...I've given myself about 2 weeks to produce the comparison and will be tidying the 2008 Collection during that time as well

[03:42:00.0000] <jgraham>
BenMillard: I've started implementing the HTML 5 spec algorithm

[03:42:01.0000] <jgraham>
No eta or promises :)

[03:44:00.0000] <BenMillard>
jgraham, that's cool

[03:44:01.0000] <Philip`>
/me gets confused trying to work out where 'eta' fits on the alpha/beta/RC/release scale

[03:45:00.0000] <jgraham>
between zeta and theta?

[03:46:00.0000] <BenMillard>
hmm, what about the proposed extensions to the current algorithms which aren't in a prototype but would support a bit more existing content?

[03:46:01.0000] <jgraham>
Pointer?

[03:47:00.0000] <BenMillard>
jgraham, the requests for <td headers> pointing at <td id> to make the association, and for <th> to become associated with another <th>

[03:47:01.0000] <BenMillard>
/me is looking for a suitable link...

[03:48:00.0000] <jgraham>
I thought the <th> pointing at another <th> was a major difference between smart headers and HTML5

[03:49:00.0000] <jgraham>
/me might have forgotten all about this in the last month

[03:49:01.0000] <BenMillard>
yeah, I wanted Smart Headers to do that but can't remember if it does (I think you're right, actually)

[03:50:00.0000] <BenMillard>
http://esw.w3.org/topic/HTML/Action72Headers captures some of the requests

[03:50:01.0000] <BenMillard>
jgraham, yep <th> associates with <th> in Smart Headers: http://james.html5.org/cgi-bin/tables/table_inspector.py?input_type=type_uri&uri=http%3A%2F%2Fprojectcerbera.com%2Fblog%2F2008%2F09%2Funtangle&source=&algorithm=smartheaders

[03:52:00.0000] <jgraham>
I could easilly make <td headers> pointing at <td id> associate as an option

[03:52:01.0000] <BenMillard>
that could be worth doing to help test compatibility with real tables

[03:53:00.0000] <BenMillard>
I guess <th headers> pointing at <td id> should work as well for consistency? not sure if any content actually does that, though

[03:55:00.0000] <jgraham>
/me wishes he had a personal bug tracking system

[03:56:00.0000] <jgraham>
(not for tracking bugs in my person you understand)

[03:58:00.0000] <Philip`>
So why not just set one up? :-)

[03:59:00.0000] <Philip`>
As long as you don't do something silly like try to install Bugzilla, it shouldn't be hard to set something up or find some online bug tracking service

[04:00:00.0000] <BenMillard>
jgraham, I sometimes have a block of bullet-pointed comments at the start of my (extremely badly written) code for that :)

[04:00:01.0000] <jgraham>
Philip`: No really good reason although they generally seem to be slightly more than trivial to install (even when they are not bugzilla)

[04:01:00.0000] <jgraham>
In fact I think I have a half-finished installation of roundup somewhere

[04:02:00.0000] <Philip`>
Could you use a personal Google Code project, or do you need more privacy than that allows?

[04:03:00.0000] <jgraham>
I guess that would work although I am not so enamoured with google code

[04:03:01.0000] <jgraham>
/me will probably just install something for himself eventually

[04:04:00.0000] <Philip`>
At least it's more enamourable than Sourceforge :-)

[04:05:00.0000] <Philip`>
/me installed a personal Trac some time ago, then used it for two days before realising he was too lazy to organise his life even when there's technology to help

[04:06:00.0000] <Philip`>
so now I just avoid ever having more than one responsibility at a time, so I don't need to track anything

[04:07:00.0000] <jgraham>
I installed trac once but I don't really like trac

[04:11:00.0000] <yecril71>
A script should be able to recognize the current display medium by querying the DOM.

[04:12:00.0000] <yecril71>
There is no need to restrict a script using a declarative attribute.

[04:13:00.0000] <yecril71>
A script can just do nothing if it detects a different medium than its target.

[04:59:00.0000] <BenMillard>
I'm off to give blood now, bye all.

[05:07:00.0000] <yecril71>
A hidden input can be placed in the last item.

[05:07:01.0000] <yecril71>
Closing items and rows is optional.

[05:07:02.0000] <yecril71>
Internet Explorer ignores </LI > and </TR > completely.

[05:10:00.0000] <yecril71>
The content that follows goes into the latest LI or TD.

[05:21:00.0000] <Hish>
hi. is there a way to define html fragments in html5? so, is <div template="#xxx"></div> <datatemplate id="xxx">content</datatemplate> --> <div template="#xxx">content</div> valid. or is there another/better way?

[05:27:00.0000] <Philip`>
Hish: <datatemplate> has been removed from HTML5

[05:29:00.0000] <Philip`>
But ignoring that, I don't quite understand the question

[05:30:00.0000] <Hish>
Philip: oh ok. Well, think about a presentation where we have a bunch of slides which will all refer to one master slide with a background image and maybe some svg graphics. is there a way to 'implement' this masterslide into all slides? pretty much as I described above?

[05:31:00.0000] <Hish>
to keep going with this example: currently I have to clone the masterslide and append it to each slide.

[05:35:00.0000] <takkaria>
I'm not sure what you mean, given that HTML doesn't natively support slides in any way

[05:37:00.0000] <Philip`>
Hish: Ah, like including an external file into the middle of another HTML document?

[05:38:00.0000] <Philip`>
or is this all in a single file?

[05:38:01.0000] <Hish>
both

[05:40:00.0000] <Hish>
but in general, that's what I mean: including html fragments n times into a html document. 

[05:40:01.0000] <Hish>
think about 10 divs and all should contain the same background image (ok, can be done by CSS) and the same svg graphic (can not be done by CSS). to get this graphic on all pages, I would need to clone them with javascript or put 10 grapics into each div. 

[05:41:00.0000] <Philip`>
There's <iframe> which lets you include an external file (and <iframe seamless> in HTML5, to make its size adjust automatically to fit its content), but I'm not sure that'd work here

[05:41:01.0000] <Hish>
no. must be in the same DOM.

[05:42:00.0000] <Hish>
I mean, it's ok to do the javascript clone. I just saw this template attribute in html5 and I was thinking that'll do the job. but maybe this tag is also removed or meant in another way.

[05:43:00.0000] <Philip`>
There has been lots of discussion about something like an <include src="..."> to insert code from elsewhere into the middle of a document, but (if I remember correctly) that has been rejected since there are too many problems

[05:44:00.0000] <Hish>
ok. thx for your help.

[05:44:01.0000] <Philip`>
The data templates were intended partly for forms, but I guess they could be used as a more generic templating system for this kind of thing

[05:45:00.0000] <Philip`>
but it turned out that they weren't a good enough solution to deserve being put in the spec, so the template feature was removed

[05:45:01.0000] <Hish>
I know these data templates from old XUL times and they were a nightmare

[05:47:00.0000] <Philip`>
I've never used XUL, but I tried writing one example using the HTML5 data templates and it sort of worked out in the end but it wasn't the most intuitive thing ever

[06:17:00.0000] <Philip`>
/me learns that jQuery doesn't like you doing $(...).slice("1")

[06:17:01.0000] <Philip`>
Uh

[06:17:02.0000] <Philip`>
/me learns that jQuery doesn't like you doing $(...).eq("1")

[06:18:00.0000] <Philip`>
because it implements it with Array.slice(i, i+1) which is slice("1", "11") which is not good

[07:01:00.0000] <yecril71>
You need XSLT: XML->HTML for templates.

[07:01:01.0000] <yecril71>
This is quite heavyweight.

[07:01:02.0000] <yecril71>
Or you can do JavaScript: JSON->HTML.

[07:02:00.0000] <hsivonen>
yecril71: what kind of templates? what's heavy?

[07:03:00.0000] <yecril71>
Templates to insert the same piece of HTML into various places in the same document.

[07:03:01.0000] <yecril71>
Hish asked a question about that recently.

[07:04:00.0000] <yecril71>
The XSLT engine is heavy.

[07:07:00.0000] <jgraham>
"Need" is a  rather strong word. You could also use server-side templates or js bsed templates

[07:18:00.0000] <Philip`>
Just write your whole web page with document.write(), then it's easy to repeat sections of markup

[07:22:00.0000] <jgraham>
Philip`: I thought that's how most webpages were wrotten

[07:22:01.0000] <jgraham>
written

[07:22:02.0000] <jgraham>
or rotten I guess

[07:26:00.0000] <yecril71>
Yes, that is the right word for the activity :-)

[07:27:00.0000] <yecril71>
I have recently tried a JSON->HTML framework for the purpose.

[07:27:01.0000] <yecril71>
Seems promising, although not having closing tags can be frustrating sometimes.

[07:45:00.0000] <Hish>
is it allowed to use customs tags like <tab> or <mynamespace:tab> in html5 or will they be rejected by a validator?

[07:46:00.0000] <Philip`>
Hish: They will be rejected

[07:46:01.0000] <Hish>
thx

[07:47:00.0000] <Philip`>
The usual extensibility solution is something like <div class="tab">

[07:47:01.0000] <Hish>
that's how it works right now ;-)

[07:47:02.0000] <Philip`>
Custom tags are particularly bad for backward compatibility, since most browsers don't parse them nicely

[07:48:00.0000] <Hish>
am I wrong or do most of the browsers parse unknown tags as inline elements?

[07:49:00.0000] <Philip`>
and they're bad for forward compatibility (if you're not using special syntax like namespaces), since a future version of HTML might want to start using that name

[07:51:00.0000] <Philip`>
Hish: You are partly wrong - IE parses unknown tags as empty elements (so <foo>bar</foo> is an empty FOO element, then "bar" text, then an empty /FOO element)

[07:51:01.0000] <Philip`>
and other browsers don't all handle unknown tags exactly like inline tags like <span>

[07:52:00.0000] <Hish>
well, for I you need this stupid document.createElement("foo") statement

[07:52:01.0000] <Philip`>
though they're close enough that it works in most cases if you pretend they're just inline

[07:52:02.0000] <Hish>
for IE

[07:53:00.0000] <Philip`>
Ah, yes - if you do that in IE, then they get parsed like pseudo-XML elements, which sort of does what you'd expect most of the time

[07:54:00.0000] <Hish>
html5 tags work fine with

[07:54:01.0000] <Hish>
<!--[if IE]><script>

[07:54:02.0000] <Hish>
new function (){

[07:54:03.0000] <Hish>
 var a=["section","header","footer"............];

[07:54:04.0000] <Hish>
 for(var i=0; i<a.length; i++)

[07:54:05.0000] <Hish>
  document.createElement(a[i]);

[07:54:06.0000] <Hish>
}();

[07:54:07.0000] <Hish>
</script><![endif]-->

[07:54:08.0000] <mookid>
have you implemented my Accept tag yet?

[07:55:00.0000] <Philip`>
Hish: Depends on what you mean by "fine" :-)

[07:55:01.0000] <Philip`>
IE will parse "<section/>" into an empty element, whereas every other browser will just treat it as a start tag that contains whatever follows

[07:56:00.0000] <Philip`>
and crazy things will happen when you have misnested tags, which won't match the crazy things that HTML5 says browsers must do

[07:57:00.0000] <Hish>
<section /> doesn't make sense for me at all

[07:58:00.0000] <Hish>
but I agree. with IE you can only get the basics to work.

[09:11:00.0000] <Philip`>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Cdiv%20style%3D%22position%3A%20absolute%22%3EOne%3C%2Fdiv%3E%0D%0A%3Cdiv%20style%3D%22margin-top%3A%202em%22%3E%26nbsp%3B%20%26nbsp%3B%20%26nbsp%3B%20%26nbsp%3BTwo%3C%2Fdiv%3E

[09:12:00.0000] <Philip`>
Opera 9.62 and IE6 put line Two below line One; Firefox 2/3 and Safari 3.0 put them on the same line

[09:12:01.0000] <Philip`>
so I guess someone is buggy?

[13:49:00.0000] <Hixie>
man it annoys me when people say that content labelled as xml that isn't well-formed xml isn't xml.

[13:51:00.0000] <takkaria>
all depends on your ontology, really

[13:52:00.0000] <Philip`>
Why is it that the word "ontology" has always filled me with loathing, even before I had any idea at all what it meant?

[13:52:01.0000] <takkaria>
me too, actually

[13:52:02.0000] <takkaria>
I guess "definitions" is equally as good a word

[13:53:00.0000] <Philip`>
Whenever someone says "I'm working on developing ontologies", I think "aaargh! run away!" despite having no understanding that would rationally justify that response

[13:53:01.0000] <takkaria>
I think my distrust of the word came from seeing it used by people who talked about RDF

[13:54:00.0000] <takkaria>
but being on a philosophy degree, it just becomes part of everyday language use

[13:59:00.0000] <Philip`>
Why do people always talk about the semantics of markup, and seem to ignore that nearly all of the semantics of a document comes from its natural language textual content?

[13:59:01.0000] <Philip`>
(Am I just misunderstanding what people mean by "semantics"?)

[14:00:00.0000] <takkaria>
I think that people misunderstand what "semantics" means, given that people use the phrase "semantic meaning"

[14:01:00.0000] <takkaria>
I like to read "meaning" where other people "semantics" when it comes to HTML, so I always read it as "meaning meaning", which is a sure sign someone's not thinking clearly

[14:03:00.0000] <othermaciej>
I think some people think of "semantics" as meaning "machine-extractable meaning"

[14:03:01.0000] <othermaciej>
so it would not apply to natural language

[14:03:02.0000] <othermaciej>
even though machines can in fact in an operational sense often extract meaning from natural language

[14:04:00.0000] <othermaciej>
Hixie: it seems not completely irrational to say that - I'd certainly say content labeled as ASCII that isn't well-formed ASCII isn't ASCII

[14:27:00.0000] <Hixie>
othermaciej: i didn't say it wasn't true

[14:27:01.0000] <Hixie>
othermaciej: just that it was annoying

[14:29:00.0000] <Hixie>
hey, noah is happy that we're listening to him

[14:29:01.0000] <Hixie>
sweet

[14:30:00.0000] <takkaria>
when Rob Burns and Philip Taylor agree about a topic though, that's a sure indicator of something...

[14:46:00.0000] <Hixie>
still no roc, eh

[14:56:00.0000] <deltab>
Philip`: semantics of markup vs. semantics of content?

[15:45:00.0000] <Hixie>
/me considers changing canPlayType() to returning "probably", "maybe", and "no"

[15:53:00.0000] <Philip`>
That will be fun for everyone who doesn't read the documentation and then writes "if (canPlayType(...)) { ... }"

[15:53:01.0000] <Hixie>
http://www.whatwg.org/specs/web-apps/current-work/multipage/video.html#media-types

[15:54:00.0000] <Hixie>
well, they'll hit that error pretty soon, right

[15:54:01.0000] <Hixie>
it's not like the -1, 0, 1 case where it would take each codepath in some cases

[15:54:02.0000] <Hixie>
and the return values are at least self-documenting here


2008-11-15
[16:02:00.0000] <Hixie>
is philipj ever on IRC?

[16:16:00.0000] <Hixie>
afk

[16:30:00.0000] <Hixie>
i don't understand yecril71's change to the faq

[17:09:00.0000] <Hixie>
yt?

[17:09:01.0000] <Hixie>
er

[17:09:02.0000] <Hixie>
wc

[17:15:00.0000] <Hixie>
heycam: looks like we need three mechanisms in webidl to make enumeration things work right

[17:16:00.0000] <Hixie>
a mechanism to list properties that are exposed (name, value)

[17:16:01.0000] <Hixie>
a mechanism to list properties that aren't exposed but that can be accessed anyway (name, value)

[17:16:02.0000] <Hixie>
and a mechanism for how to handle setting of those properties (when present or not)

[17:17:00.0000] <Hixie>
and each mechanism needs a way to handle whether the IDL should override these mechanisms or be shadowed by them

[17:17:01.0000] <Hixie>
and we also need a way to mark the item, namedItem, etc methods as hooking into this

[17:19:00.0000] <othermaciej>
Hixie: I would say that WebIDL needs a way to mark up an interface as having dynamic properties; whether they show up to the "in" operator and hasOwnProperty; whether they are enumerable; and whether they take priority over IDL-defined properties or not

[17:20:00.0000] <othermaciej>
Hixie: I don't think it is necessary to represent at the IDL level that certain IDL interfaces do lookups against the same namespace as dynamic properties

[17:20:01.0000] <othermaciej>
Hixie: that can be expressed in spec prose just fine

[17:20:02.0000] <othermaciej>
(and sometimes there isn't even such a method)

[17:21:00.0000] <othermaciej>
it's kind of like the case of two properties having the same value or two methods doing the same thing; there's probably such cases, but there is no need for IDL to point them out

[17:21:01.0000] <othermaciej>
Hixie: I guess you also might want to express the traits I described separately for index and non-index properties

[17:21:02.0000] <Hixie>
i just don't want to have to say the same thing a bazillion times

[17:22:00.0000] <Hixie>
if i can just say "the item() method is a _foobarmagicmethod_." then that's fine too

[17:22:01.0000] <Hixie>
i don't care if it's in idl or prose so long as it is terse and non-redundant

[17:22:02.0000] <Hixie>
still means the webidl spec needs something for me :-)

[17:23:00.0000] <othermaciej>
Hixie: I think in prose you would define the name lookup algorithm once, and say that algorithm applies to dynamic properties, and is used by nameGetter

[17:24:00.0000] <Hixie>
right, i'm saying that the name lookup algorithm should be in webidl

[17:24:01.0000] <Hixie>
so that different specs don't need to redefine it

[17:24:02.0000] <othermaciej>
sorry?

[17:24:03.0000] <Hixie>
webidl the spec

[17:24:04.0000] <Hixie>
not the language

[17:24:05.0000] <othermaciej>
the name lookup algorithm is the thing that's specific to each object

[17:24:06.0000] <othermaciej>
in the sens eI was using the term

[17:24:07.0000] <Hixie>
oh

[17:24:08.0000] <Hixie>
i see what you're saying

[17:24:09.0000] <Hixie>
hm, yeah, maybe there's not much to specify

[17:25:00.0000] <othermaciej>
(and really it should be an algorithm that gives a current list of dynamic property names and values)

[17:25:01.0000] <Hixie>
well either way, i'm sure we can figure something out

[17:25:02.0000] <othermaciej>
(though namedItem and the like would use it for lookup purposes rather than listing ever)

[17:25:03.0000] <othermaciej>
right

[17:25:04.0000] <othermaciej>
it will be good to have this straight

[17:25:05.0000] <othermaciej>
I would like it to be possible for WebKit to use spec WebIDL directly as our IDL files with as little modification as possible

[17:26:00.0000] <Hixie>
that'd be nice

[17:26:01.0000] <othermaciej>
so I tend to think of things from that design angle

[17:26:02.0000] <othermaciej>
I know that is not the primary goal

[17:26:03.0000] <othermaciej>
but it would help validate both WebIDL itself and any spec using it that WebKit supported

[17:26:04.0000] <othermaciej>
if that worked out

[17:28:00.0000] <Hixie>
yeah

[18:08:00.0000] <aboodman2>
sicking: it seems like your example with relative urls would actually work.

[18:08:01.0000] <aboodman2>
maybe I am missing some subtle detail of how urls are resolved.

[18:09:00.0000] <sicking>
aboodman: the url would be resolved against google.com though, no?

[18:09:01.0000] <sicking>
aboodman2, rather than against example.html

[18:09:02.0000] <aboodman2>
it seems like it would resolve to http://docs.google.com/documents/report.doc

[18:09:03.0000] <sicking>
right

[18:10:00.0000] <aboodman2>
which is the same thing it would resolve to relative to index.html, assuming index.html was also hosted somewhere on http://docs.google.com/

[18:10:01.0000] <sicking>
index.html is hosted on myserver.com

[18:10:02.0000] <sicking>
sorry, should have said that

[18:10:03.0000] <aboodman2>
cross-origin workers aren't allowed in html5

[18:10:04.0000] <aboodman2>
i thought

[18:10:05.0000] <sicking>
cross-origin loading of js files are

[18:10:06.0000] <sicking>
i hope, no reason not to

[18:11:00.0000] <sicking>
they will execute in the security context of the loading page though

[18:11:01.0000] <sicking>
same as for <script>

[18:11:02.0000] <aboodman2>
huh

[18:11:03.0000] <aboodman2>
/me consults the spec

[18:11:04.0000] <sicking>
i know the original spec didn't allow it, but i thought that was a bug

[18:12:00.0000] <sicking>
i don't see why we would disallow cross-site loading of the script file since <script> already allows it

[18:12:01.0000] <aboodman2>
"If the origin of the resulting absolute URL is not the same as the origin of the script that invoked the constructor, then throw a security exception."

[18:12:02.0000] <sicking>
bah

[18:12:03.0000] <sicking>
that sucks

[18:12:04.0000] <aboodman2>
workers aren't really the same thing as <script> tags

[18:12:05.0000] <aboodman2>
i think that is part of where our mental models differ

[18:13:00.0000] <sicking>
i agree :)

[18:13:01.0000] <sicking>
for dedicated workers they are very similar

[18:13:02.0000] <aboodman2>
workers are like pages

[18:13:03.0000] <sicking>
just out-of-process

[18:13:04.0000] <aboodman2>
importScripts() is like <script> tags

[18:13:05.0000] <aboodman2>
and you can't communicate with windows that you open that are in other origins

[18:14:00.0000] <sicking>
so i see dedicated workers as simply an out-of-process <script>

[18:14:01.0000] <sicking>
where you offload stuff you don't want to execute on the UI thread

[18:14:02.0000] <aboodman2>
yes, you see them like threads.

[18:14:03.0000] <aboodman2>
i am not sure how to resolve this difference in seeingness

[18:14:04.0000] <sicking>
i think it's a difference in use cases

[18:15:00.0000] <sicking>
bah, gotta take off for bus, i'll be back on in a few mins

[18:15:01.0000] <aboodman2>
ok

[18:15:02.0000] <aboodman2>
i will keep typing as i will probably have to go soon too

[18:16:00.0000] <sicking>
quicky first though: you had suggested that we support appCodeName way back when. Is that one really useful? everyone returns 'Mozilla'

[18:16:01.0000] <sicking>
heh

[18:16:02.0000] <aboodman2>
you mean on the navigator object that is exposed to workers? I just figured that we should support the same API that windows support. no reason to make things incompatible just for the heck of it.

[18:16:03.0000] <aboodman2>
there is already code that works with the navigator object.

[18:16:04.0000] <sicking>
windows support a loooong list

[18:17:00.0000] <sicking>
https://developer.mozilla.org/En/DOM/Window.navigator

[18:17:01.0000] <aboodman2>
right, I guess I assumed that exposing things to workers was not a lot of extra work.

[18:17:02.0000] <sicking>
i was hoping to not do the whole thing

[18:17:03.0000] <aboodman2>
it is hard to know what people are using today.

[18:17:04.0000] <aboodman2>
if you take a look at the code that libraries use to figure out what browser is which, it is crazy.

[18:18:00.0000] <aboodman2>
since everybody lies about who they are, it kind of has to be.

[18:18:01.0000] <sicking>
ah, i was hoping you had some data on that when you originally posted 5 of them :)

[18:18:02.0000] <sicking>
ok, rally taking off

[18:18:03.0000] <aboodman2>
ok.

[18:19:00.0000] <sicking>
have a good weekend

[19:04:00.0000] <Hixie>
the navigator.* stuff the spec now says to support only includes the ones that all the browsers support and return different things for

[19:04:01.0000] <Hixie>
so appCodeName isn't included

[19:04:02.0000] <Hixie>
(it is in fact commented out)

[19:04:03.0000] <Hixie>
regarding cross-origin, i don't mind allowing it if people want to allow it, in fact i'd like to know the use case for not allowing it

[19:07:00.0000] <aboodman2>
Hixie: in Gears, workers run in the origin they are loaded from

[19:07:01.0000] <aboodman2>
it's a feature.

[19:08:00.0000] <aboodman2>
it fills a similar role to postMessage() across iframes. But previously you said that you didn't want to bite that off immediately.

[19:09:00.0000] <Hixie>
ah hm yes, allowing cross-origin loads would make it confusing to know what origin to give the script execution context

[19:10:00.0000] <aboodman2>
maybe you could load workers across origins if you say that workers run in the caller's origin and have the caller's base URI.

[19:10:01.0000] <aboodman2>
but then shared workers don't work as well.

[19:30:00.0000] <Hixie>
i guess we could allow importScripts() cross-origin then

[19:30:01.0000] <Hixie>
without allowing workers themselves

[19:36:00.0000] <aboodman2>
that seems reasonable

[00:33:00.0000] <yecril71>
My edit to <http://wiki.whatwg.org/index.php?title=FAQ&oldid=3433#Should_I_top-post_or_reply_inline.3F> was intentional.

[00:34:00.0000] <yecril71>
I have explained it at <http://wiki.whatwg.org/wiki/Talk:FAQ>.

[00:36:00.0000] <yecril71>
The URL "/documents/report.doc" is local absolute.

[00:36:01.0000] <yecril71>
It will not cause the worker to load the wrong document.

[00:53:00.0000] <Hish>
hi, can anybody help me with a line height issue? <p style="font-size:36pt;">abc<span style="font-size:10pt;"> some long text </span></p> results in a paragraph with a line height of 36pt even in the lines which should only have 10pt. this happens in xhtml 5 as well as html 5. html 4 transitional works fine.

[00:56:00.0000] <yecril71>
I might be able to, although not on this channel.

[00:58:00.0000] <Hish>
yecril71: what other channel?

[00:59:00.0000] <yecril71>
What do you mean by "in html 5"?

[01:00:00.0000] <Hish>
when I set the doctype to <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd"> it looks fine and when I set it to <!DOCTYPE HTML> I have this problem.

[01:01:00.0000] <Hish>
and I assumed that <!DOCTYPE HTML> is used for html 5. isn't it?

[01:01:01.0000] <yecril71>
Not that way.  HTML 5 uses it, but the rendering engine cannot infer it is HTML 5 from it.

[01:02:00.0000] <yecril71>
HTML 5 does not have a way to name its version.

[01:02:01.0000] <Hish>
ok. then maybe I'm wrong here. thx

[01:19:00.0000] <yecril71>
"pixel float" is a state of pixels that float.

[01:19:01.0000] <yecril71>
So "pixelfloat" for video aspect ratio is rather comical.

[02:45:00.0000] <Philip`>
Hish: What browser(s) do you see that behaviour in?

[02:46:00.0000] <Philip`>
Hish: Firefox 3 seems to do 36pt spacing with both <!DOCTYPE HTML> and <!...that html4/strict one...> (but does narrower spacing if it's in quirks mode if you don't give any doctype at all)

[02:46:01.0000] <Hish>
firefox, safari opera. in opera you can fix it when the outer tag is a span instead of a p

[02:47:00.0000] <Hish>
firefox only displays is correctly with html4 loos dtd

[02:48:00.0000] <Hish>
but I need xhtml to display inline svg

[02:48:01.0000] <Hish>
also I don't give a great future to html 4 loose

[02:48:02.0000] <Philip`>
Hish: Sounds like a quirks vs standards mode issue

[02:49:00.0000] <Philip`>
https://developer.mozilla.org/en/Mozilla_Quirks_Mode_Behavior mentions line height behaviour, so it's probably that

[02:49:01.0000] <Hish>
yep. but in this case, the standard behaviour doesn't make sense. also opera shows me that it can be displayed properly

[02:49:02.0000] <BenMillard>
Hish, maybe you could upload some test cases so others can see what's happening themselves?

[02:50:00.0000] <Hish>
do you have a pastebin here?

[02:50:01.0000] <Philip`>
If it's short, you can use the permalink feature on http://software.hixie.ch/utilities/js/live-dom-viewer/

[02:51:00.0000] <takkaria>
I think either behaviour makes sense, really, there are pros and cons to either

[02:55:00.0000] <Philip`>
Can't you just do <!DOCTYPE html><p style="font-size:10pt"><span style="font-size:36pt">abc</span> some long text </p> ?

[02:56:00.0000] <Philip`>
(You probably don't want to use quirks mode (which is what you get with the HTML4 Transitional DTD, or with no doctype at all) because that makes browsers even more weird and buggy and uninteroperable)

[02:57:00.0000] <Hish>
http://pastebin.com/m5d342c7e

[02:59:00.0000] <BenMillard>
using <ol> for code samples is an anti-pattern...you can't copy and paste them :(

[02:59:01.0000] <Hish>
there's a download button

[03:00:00.0000] <Philip`>
There's a textarea you can copy from

[03:00:01.0000] <BenMillard>
Philip`, aha! it was below the fold

[03:00:02.0000] <Lachy>
WOW! I can't believe the stupidity in this video http://failblog.org/2008/09/08/conspiracy-fail/

[03:01:00.0000] <Philip`>
Also you could use Opera, which doesn't insert annoying "#" lines if you copy-and-paste from the <ol>

[03:02:00.0000] <BenMillard>
Hish, does your test expect a specific browser width?

[03:03:00.0000] <Hish>
try 500px so that the small lines wrap over 2-3 lines

[03:05:00.0000] <BenMillard>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20HTML%3E%0A%3Chtml%3E%0A%3Chead%3E%0A%3Cmeta%20http-equiv%3D%22Content-Type%22%20content%3D%22text%2Fhtml%3B%20charset%3Dutf-8%22%3E%0A%3Ctitle%3EUntitled%20Document%3C%2Ftitle%3E%0A%3Cstyle%20type%3D%22text%2Fcss%22%3E%0Abody%20%7Bfont-size%3A6px%3B%7D%0A.style1%20%7B%0A%09font-size%3A36px%3B%0A%7D%0A.style2%20%7B%0A%20%20%20%20font-size%3A10px%3B%20%0A%7D%0

[03:05:01.0000] <BenMillard>
%2Fhead%3E%0A%0A%3Cbody%3E%0A%3Cp%3E%3Cspan%20class%3D%22style2%22%3Ethis%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small

[03:05:02.0000] <BenMillard>
20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20%3C%2Fspan%3E%3C%2Fp%3E%0A%3Chr%3E%0A%3Cp%3E%3Cspan%20class%3D%22style1%22%3Ebottomline%3A%20inline%20elements%20with%20a%20large%20font%20size%20must%20be%20closed%20before.%20Currently%20only%

[03:05:03.0000] <BenMillard>
s%20correctly.%20Now%20a%20small%20text%3C%2Fspan%3E%3Cspan%20class%3D%22style2%22%3Ethis%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20i

[03:05:04.0000] <BenMillard>
his%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20%3C%2Fspan%3E%3Cspan%20class%3D%22style1%22%3EOk.%20now%20back%20to%20large.%20%3C%2Fspan%3E%3C%2Fp%3E%0A%3Chr%3E%0A%3Cp%20class%3D%22style1%22%3E%3Cspan%20style%3D%22font-size%3

[03:05:05.0000] <BenMillard>
is%20a%20test.%3C%2Fspan%3E%3Cspan%20class%3D%22style2%22%3Ethis%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%

[03:05:06.0000] <takkaria>
fail!

[03:05:07.0000] <BenMillard>
20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20%3C%2Fspan%3Enow%20back%20to%20large.%20And%20also%20here%20a%20few%20lines.%20And%20also%20here%20a%20few%20lines.%20And%20also%20here%20a%20few%20lines.%20And%20also%20here%20a%20few%20lines.%20And%20a

[03:05:08.0000] <BenMillard>
20lines.%20And%20also%20here%20a%20few%20lines.%20And%20also%20here%20a%20few%20lines.%20%3C%2Fp%3E%0A%3Chr%3E%0A%3Cp%3E%3Cspan%20class%3D%22style1%22%3EThis%20part%20is%20only%20displayed%20correctly%20by%20opera.%20%3Cspan%20class%3D%22style2%22%3Ethis%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%2

[03:05:09.0000] <BenMillard>
%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20this%20is%20very%20small.%20%3C%2Fspan%3E

[03:05:10.0000] <BenMillard>
20And%20also%20here%20a%20few%20lines.%20And%20also%20here%20a%20few%20lines.%20And%20also%20here%20a%20few%20lines.%20And%20also%20here%20a%20few%20lines.%20And%20also%20here%20a%20few%20lines.%20And%20also%20here%20a%20few%20lines.%20%3C%2Fspan%3E%3C%2Fp%3E%0A%3C%2Fbody%3E%0A%3C%2Fhtml%3E

[03:05:11.0000] <BenMillard>
yikes, that was long

[03:07:00.0000] <gsnedders>
uh…

[03:07:01.0000] <gsnedders>
flooder!

[03:08:00.0000] <gsnedders>
/me agreess with takkaria 

[03:09:00.0000] <Philip`>
I think you should pass it through hugeurl.com first

[03:09:01.0000] <gsnedders>
would that not shorten it?

[03:10:00.0000] <Philip`>
Uh, no, because it's HugeURL and its aim is not to shorten URLs

[03:13:00.0000] <BenMillard>
http://xrl.us/oxesv seems to work

[03:14:00.0000] <BenMillard>
"(0% of the original length)" lol

[03:17:00.0000] <Hish>
I don't see a difference to my doc

[03:17:01.0000] <Hish>
the line-heights are still too large

[03:18:00.0000] <BenMillard>
by "works" I mean "loads the correct page without flooding this channel" :)

[03:18:01.0000] <Hish>
ok...

[03:19:00.0000] <BenMillard>
the demo is more about CSS than HTML

[03:20:00.0000] <Hish>
it's more about different DTDs

[04:47:00.0000] <Hish>
just a quick thanks to all who helped me with this line-height thing. I solved it using a .class>* selector.

[05:55:00.0000] <BenMillard>
I see the phrase "spec lawyering" being used in documents and IRC logs but no clear explanation of what it means.

[05:56:00.0000] <BenMillard>
anyone care to enlighten me? :)

[06:05:00.0000] <hsivonen>
BenMillard: spec lawyering means reading a spec in very careful and pedantic way in order to extract precise meaning (as opposed to a more casual reading)

[06:05:01.0000] <BenMillard>
hsivonen, thanks!

[06:05:02.0000] <BenMillard>
does it imply that's a good or bad thing?

[06:07:00.0000] <hsivonen>
BenMillard: spec lawyering is good when implementing a spec, but it's not appropriate to expect e.g. Web authors to be spec lawyers

[06:11:00.0000] <BenMillard>
hsivonen, so if a web author does read a spec very carefully to extract the precise meaning of a particular feature, are they doing a good thing?

[06:12:00.0000] <hsivonen>
BenMillard: good thing, but we shouldn't write spec in such a way that they have to do that

[06:12:01.0000] <BenMillard>
I see

[07:27:00.0000] <jgraham>
https://listserv.heanet.ie/cgi-bin/wa?A2=ind9509&L=html-wg&D=0&P=32830

[07:31:00.0000] <takkaria>
nice!

[07:35:00.0000] <takkaria>
DanC in his reply to that says, "But a more important question is: which will get us to reliable interoperability quicker: getting the document base to be SGML-conforming, or getting all the deployed software to agree on a specification derived from HTMLparse.c?

[07:39:00.0000] <jgraham>
It's interesting that his following assumption "We documents have a short half life"has tuned out to be fairly inaccurate

[07:51:00.0000] <Philip`>
jgraham: If you consider "We[b] documents" to be the collection of all documents on the web, and you reasonably assume that that collection is expanding exponentially, then it seems quite accurate to say that the half-life is short

[07:51:01.0000] <Philip`>
If the size of the web doubles every year, then the proportion of documents older than any given date will decay with a half life of one year

[07:56:00.0000] <takkaria>
also, back in '95, they only had a few years of data

[07:56:01.0000] <Dashiva>
And no geocities

[07:56:02.0000] <takkaria>
so everything was fairly recently edited, really

[07:57:00.0000] <Philip`>
"just stick around a while. Things will settle down." - and now, 13 years later...

[07:57:01.0000] <Philip`>
Did anyone back then expect they'd be sticking around for so long a while? :-)

[08:00:00.0000] <takkaria>
that a strategy has been pursued for thirteen years and has failed seems to count as good reasoning for taking a different approach

[08:08:00.0000] <Philip`>
You're willing to wave the white flag of surrender and throw out all that work when we may be on the cusp of success?

[08:09:00.0000] <jgraham>
Philip`: It seems unfair to look at the original documents in fractional terms. A more interesting question would be given a fixed sample of documents how many of them are still useful/avaliable n years later

[08:10:00.0000] <jgraham>
s/terms/terms relative to the current web/

[08:10:01.0000] <takkaria>
Philip`: sounds suspiciously like the arguments against drug law reform

[08:12:00.0000] <jgraham>
/me really doesn't understand why people get hung up about "clean" markup anyway

[08:13:00.0000] <jgraham>
There are obvious advantages to the original author

[08:14:00.0000] <Philip`>
Maybe because people wish the world was nice and clean and sensible and understandable, not just a giant collection of hacks, and so they want to promote moving in that direction

[08:16:00.0000] <jgraham>
But I wonder why they wish that. Is it just a built in sense of aesthetics or is there a practical benefit that can actually be realied

[08:16:01.0000] <jgraham>
realised

[08:16:02.0000] <jgraham>
I guess there might be some correlation between having conformant markup and people being able to use tools like greasemonkey although it's not obvious that that is always true

[08:17:00.0000] <Philip`>
Why do you say "just"? Most of human art and mathematics is about striving for aesthetics, so it seems a fairly important goal :-)

[08:18:00.0000] <jgraham>
/me is going to ask for arts council funding next time he writes a html document

[08:20:00.0000] <Philip`>
It's not about the aesthetics of a single HTML document, it's about the aesthetics of the entire system and associated culture

[08:36:00.0000] <jgraham>
Philip`: In that case we alreasy lost and the esthetics argument should die :)

[08:36:01.0000] <jgraham>
aesthetics

[08:45:00.0000] <takkaria>
I dunno, you could probably do some funky ASCII art that was also valid HTML

[09:56:00.0000] <Philip`>
jgraham: Just because perfection is unattainable, that doesn't mean you should stop striving for it :-)

[10:49:00.0000] <Hixie>
a lot of geeks, to some extent or another, suffer from a series of mental conditions that have in common a fundamental desire for tidyness

[10:49:01.0000] <Hixie>
it's not necessarily a rational thing, though geeks, given their usually high level of intelligence, will often find rationalisations for it

[10:50:00.0000] <Hixie>
being of that mindset myself, i would quite like markup to be more widely conforming myself :-)

[10:50:01.0000] <Dashiva>
I have that obsession with code and other digital things, but my apartment is a right mess

[10:50:02.0000] <Hixie>
yeah it varies from person to person

[10:51:00.0000] <Hixie>
i'm similar

[10:51:01.0000] <Hixie>
i know people who are the opposite

[12:46:00.0000] <othermaciej>
I think the root geek desire is not necessarily for tidiness as such but for consistency

[12:46:01.0000] <othermaciej>
and for fitting into a system

[12:46:02.0000] <othermaciej>
at least as to technical things

[13:39:00.0000] <hsivonen>
I wonder what kind of client-side software Roy Fielding would ship

[13:41:00.0000] <takkaria>
bad client-side software? :)

[13:42:00.0000] <Dashiva>
What's the context now?

[13:43:00.0000] <jmb>
that rather depends if the html his software is dealing with has come from the web or not

[13:45:00.0000] <hsivonen>
Dashiva: public-html

[13:46:00.0000] <takkaria>
it surprises me a bit that he doesn't think most web authors don't author for browsers


2008-11-16
[01:24:00.0000] <jgraham>
". They have clearly told us several times that the reason they use our products is to view pages on the web" heh

[01:34:00.0000] <erlehmann>
jgraham, wat?

[01:35:00.0000] <hsivonen>
erlehmann: see sicking's email to public-html

[01:47:00.0000] <jgraham>
BenMillard: I have some semi-working code for the HTML5 table header lgorithm but I won't get to finish it before Wednesday at the earliest

[01:48:00.0000] <BenMillard>
jgraham, that's OK. Don't bust a gut over it. :)

[01:48:01.0000] <jgraham>
No guts will be busted

[01:48:02.0000] <BenMillard>
lol

[03:19:00.0000] <gsnedders>
MikeSmith's doc breaks ISO 2145!

[03:20:00.0000] <hsivonen>
gsnedders: does ISO 2145 have two interoperable implementations?

[03:20:01.0000] <gsnedders>
hsivonen: LaTeX, MediaWiki, and Anolis at least

[03:22:00.0000] <Philip`>
I think you should be much more concerned that it breaks XHTML 1.0, rather than a boring numbering spec that nobody cares about :-p

[03:23:00.0000] <gsnedders>
Philip`: Nobody is an exaggeration. We have at least one case of a person caring.

[03:23:01.0000] <Philip`>
But he's a nobody :-p

[03:23:02.0000] <gsnedders>
http://lists.w3.org/Archives/Public/www-html-editor/2005JulSep/0003.html

[03:24:00.0000] <hsivonen>
/me wonders how long will it take for someone to complain about the browser-relatedness of the default style sheet bits

[03:27:00.0000] <hsivonen>
gsnedders: is ISO 2145 the same for all locales?

[03:27:01.0000] <gsnedders>
hsivonen: Yes

[03:28:00.0000] <hsivonen>
what a great opportunity for a country who wants to feel important to deviate from the standard

[03:36:00.0000] <Philip`>
/me notes that Markus Kuhn seems to be quite unhappy with the US's deviation from the standard A4-style paper sizes

[03:39:00.0000] <hsivonen>
I'd take the decimal separator and spelling from the U.S. but I'd take paper sizes and electricity sockets from Germany

[03:40:00.0000] <gsnedders>
Philip`: ISO 216 paper sizes, not A4-style

[03:40:01.0000] <gsnedders>
/me ducks

[03:40:02.0000] <hsivonen>
(seriously, the EU should do something about British/Irish and Italian electricity sockets)

[03:40:03.0000] <takkaria>
I have a suspicion gsnedders has most of the ISO specs in his bedroom. :)

[03:41:00.0000] <gsnedders>
takkaria: I'm not rich enough :P

[03:41:01.0000] <takkaria>
hsivonen: aren't the european sockets only two-pronged?

[03:41:02.0000] <gsnedders>
takkaria: No

[03:41:03.0000] <gsnedders>
Well, depends where

[03:41:04.0000] <gsnedders>
Some earted sorkets are three; most are two

[03:41:05.0000] <gsnedders>
/me can't type

[03:42:00.0000] <hsivonen>
takkaria: no. Both the German and French sockets have 3-wire extensions and the common plug design fits both French and German 3-wire sockets

[03:42:01.0000] <takkaria>
sounds useful

[03:42:02.0000] <takkaria>
otoh, I don't see that the EU can do anything about the UK's sockets now really

[03:43:00.0000] <hsivonen>
The unearthed plug on my magsafe charger fits pretty much any EU socket except British/Irish

[03:43:01.0000] <hsivonen>
the earthed plug doesn't fit into Italian sockets

[03:43:02.0000] <Philip`>
/me has a Danish desk lamp, which has a two-pronged plug, but it fits fine in a three-prong socket if he uses another plug to temporarily stick into the plug's earth socket to open up the other two holes, and he hopes that's not too dangerous a thing to do

[03:44:00.0000] <takkaria>
heh, I have done that before

[03:44:01.0000] <hsivonen>
the kind of unearthed plug that fits into Italian sockets can be crammed onto a British one, too, but not well

[03:44:02.0000] <takkaria>
what with the UK generally being a bit Eurosceptic, the general public would get very pissed off with any attempt to change plug sockets

[03:46:00.0000] <gsnedders>
BS 1363 is also widely regarded as safer

[03:46:01.0000] <Philip`>
How would it be possible to migrate the plug/socket format without suddenly breaking backward compatibility and forcing everyone to buy dozens of adapter thingies?

[03:46:02.0000] <hsivonen>
I'm not so concerned about using adapters when traveling in Britain/Ireland. What bothers me is the waste of all the UK-style cords that I've received just in case in various hardware packages

[03:47:00.0000] <hsivonen>
Philip`: I think it should be possible to construct a socket that has the UK and German earthing features

[03:49:00.0000] <gsnedders>
The problem is fitting circular and rectangular plugs in

[03:55:00.0000] <gsnedders>
BS 546 wouldn't have that problem, but is only used in lighting applications in the UK now :P

[04:07:00.0000] <gsnedders>
Also, my understanding is the main issue is most British houses use 30A ring circuits, and thus the plugs really need to be fused

[04:09:00.0000] <Philip`>
We should just move to a wireless power infrastructure

[04:09:01.0000] <hsivonen>
gsnedders: is there a fuse in the *plug*??

[04:10:00.0000] <Philip`>
hsivonen: Yes, like in http://www.leadsdirect.co.uk/images/technical/uk%20mains%20plug%20wiring.png

[04:11:00.0000] <hsivonen>
wow. that's a design choice I didn't think about at all

[04:12:00.0000] <Philip`>
Where else would you put the fuses? :-)

[04:13:00.0000] <hsivonen>
Philip`: at the point where the apartment's wiring connects to the building wiring

[04:14:00.0000] <Philip`>
Ah

[04:16:00.0000] <gsnedders>
Philip`: That's the case with BS 546, which is why they are used in theatres a lot

[04:16:01.0000] <gsnedders>
Philip`: If you have a fuse inaccessible on a lighting rig…

[04:17:00.0000] <Philip`>
I suppose that's less sensitive (and therefore less safe?) than having a fuse in each plug

[04:17:01.0000] <gsnedders>
Philip`: It's just as safe

[04:17:02.0000] <gsnedders>
Philip`: It's the same amperage as in the socket

[04:18:00.0000] <gsnedders>
Philip`: There's a separate wire from the connector box to each plug

[04:18:01.0000] <takkaria>
hsivonen: wikipedia has some history of why it was put in the plug

[04:19:00.0000] <gsnedders>
High amperage ring circuits to lose less copper post-war

[04:19:01.0000] <gsnedders>
s/lose/use/

[04:19:02.0000] <takkaria>
that could be a headline

[04:19:03.0000] <hsivonen>
gsnedders: my apartment has 6 fuses: one for stove, two for lighting and 3 for regular sockets

[04:20:00.0000] <Philip`>
If I have a 3A fuse in the plug for some device that usually uses 2A, and then that device gets a bit shorted or something so that it starts drawing as much current as a kettle, then it'd be nice for it to blow the fuse rather than to happily pass through a 13A fuse that's installed miles away in some cupboard where I don't know what device is connected to what socket

[04:20:01.0000] <hsivonen>
and I have more than 3 regular sockets

[04:20:02.0000] <hsivonen>
so if a fuse burns, several devices lose power

[04:21:00.0000] <gsnedders>
And the plug is less than half the amperage of the ring circuit, so for safety reasons that it has to be in the plug/socket. Having it in the plud allows smaller fuses in low power devices

[04:21:01.0000] <gsnedders>
hsivonen: But do several sockets just share 16A?

[04:22:00.0000] <hsivonen>
gsnedders: yes

[04:22:01.0000] <gsnedders>
hsivonen: That's not the case in the UK so much

[04:24:00.0000] <Philip`>
gsnedders: So we can blame the Nazis for our electric socket design?

[04:24:01.0000] <gsnedders>
Philip`: In part :P

[04:24:02.0000] <Philip`>
They're a soft target, so I'm happy to wholly blame them

[04:26:00.0000] <gsnedders>
/me has no idea what the wiring like in this Edwardian house is like

[04:26:01.0000] <gsnedders>
It has had several systems in it

[04:28:00.0000] <gsnedders>
We have BS 546 plugs still

[04:28:01.0000] <gsnedders>
Still live, too

[04:30:00.0000] <hsivonen>
Ah. Wikipedia confirms my suspicion of Denmark not being a Nordic team player when it comes to electricity sockets

[04:38:00.0000] <Philip`>
/me sees someone complaining about the limited range of fonts in Google Docs, and supposes that's a good opportunity for using downloadable fonts so it's not limited to the intersection of what all users have installed by default

[04:48:00.0000] <takkaria>
uploadable fonts would be useful too

[04:55:00.0000] <Philip`>
Hmm, can you upload fonts from web browsers?

[04:56:00.0000] <Philip`>
Windows Explorer handles its fonts directory really weirdly and seems to try to prevent you from ever seeing that there's actual files in there

[05:00:00.0000] <takkaria>
you probably could

[05:00:01.0000] <takkaria>
just browse to c:\windows\fonts or wherever, I would have thought it would let you select font files that wya

[05:03:00.0000] <Philip`>
But that's a magic folder

[05:03:01.0000] <hsivonen>
does IE allow EOT fonts for contenteditable parts of HTML?

[05:04:00.0000] <Philip`>
(I'm not sure quite how magic it is, though)

[05:09:00.0000] <yecril71>
If any equivalence between mouse position and sound balance

[05:09:01.0000] <yecril71>
for the blind should be invented,

[05:10:00.0000] <yecril71>
it should be invented by the operating system GUI

[05:10:01.0000] <yecril71>
or an assistive technology add-on.

[05:10:02.0000] <yecril71>
HTML has nothing to do with that.

[05:15:00.0000] <takkaria>
ever thought of publishing a book of philosophical remarks on HTML? :)

[05:19:00.0000] <yecril71>
Never in my life.

[05:20:00.0000] <yecril71>
HTML is a very practical thing to me.

[05:20:01.0000] <yecril71>
That is, since Yucca explained to me not to expect too much.

[05:27:00.0000] <hsivonen>
what's Yucca working on these days?

[05:35:00.0000] <yecril71>
I have seen his post to www-html quite recently.

[10:09:00.0000] <hsivonen>
I'm aware of problems wih html5.validator.nu. validator.nu works. sorry about the inconvenience. I'll fix.

[10:42:00.0000] <hsivonen>
html5.validator.nu is back up

[10:42:01.0000] <hsivonen>
sorry about the outage

[10:51:00.0000] <gsnedders>
/me wonders where he could get a job for three months next summer

[10:54:00.0000] <jcranmer>
intern for some company

[10:55:00.0000] <wilhelm>
gsnedders: http://www.opera.com/company/jobs/opening.dml?id=211

[10:55:01.0000] <Philip`>
Your local supermarket might have some vacancies

[10:56:00.0000] <gsnedders>
wilhelm: "Enrollment in a 3-5 year university program, or completing such a program in 2009." — not me

[10:57:00.0000] <hsivonen>
looks like the AI meme is continuing: http://www.cynicalturtle.net/kame/index.php/2008/11/14/487-liens-du-14-11-2008

[10:57:01.0000] <jcranmer>
gsnedders: ask politely?

[10:57:02.0000] <gsnedders>
wilhelm: And May is too soon. My last exam is first Friday of June

[10:57:03.0000] <gsnedders>
Philip`: That won't pay so well :P

[10:58:00.0000] <jcranmer>
GSoC?

[10:58:01.0000] <hsivonen>
does GSoC not require university enrollment?

[10:58:02.0000] <jmb>
not strictly, no

[10:58:03.0000] <gsnedders>
jcranmer: see hsivonen, and I'll be < 18

[10:59:00.0000] <wilhelm>
gsnedders: If it looks interesting, apply anyway. I didn't fulfil the formal requirements when I applied.

[10:59:01.0000] <jmb>
the <18 thing may be more of an issue there

[10:59:02.0000] <jcranmer>
starts May 26

[10:59:03.0000] <jcranmer>
too early

[11:00:00.0000] <gsnedders>
I know someone who tried to get exempt from that, as he was 17 at the application deadline but was 18 at the start of the time

[11:00:01.0000] <gsnedders>
He didn't get exempted.

[11:01:00.0000] <jcranmer>
Yes. You must be 18 years of age or older by April 14, 2008 to be eligible to participate in Google Summer of Code in 2008.

[11:01:01.0000] <gsnedders>
/me will be 18 on 20100420

[11:01:02.0000] <jcranmer>
trying to get something under 18 is a tad difficult

[11:01:03.0000] <Philip`>
You could lie about your age

[11:02:00.0000] <jcranmer>
any government-based summer internships?

[11:02:01.0000] <jcranmer>
I did a SEAP program (through the U.S. Navy) before my junior year

[11:02:02.0000] <gsnedders>
jcranmer: In Scotland age of majority is 16 :P

[11:03:00.0000] <gsnedders>
anyhow, supper

[11:18:00.0000] <takkaria>
wilhelm: hmm, that's an interesting link, thanks. :)

[11:25:00.0000] <wilhelm>
(c:

[11:58:00.0000] <Hixie>
Philip`: you can make magic folders non-magic by deleting the hidden file in them that makes them magic

[11:58:01.0000] <Hixie>
something .ini iirc

[12:04:00.0000] <gavin>
desktop.ini

[12:50:00.0000] <Philip`>
/me wonders if someone should tell bz that arguing with Rob Burns is probably not going to be very productive

[15:29:00.0000] <Hixie>
http://www.w3.org/mid/5AB52079-5CDC-4AA5-BD60-458A6A8ACD0E⊙gc

[15:29:01.0000] <Hixie>
i really have no idea what to make of that

[15:31:00.0000] <Philip`>
You could print it out and make a paper plane of it


2008-11-17
[16:03:00.0000] <nessy>
I'd call that web fundamentalism

[16:04:00.0000] <Hixie>
i don't even know what he means

[16:05:00.0000] <Hixie>
which i guess is his point

[16:05:01.0000] <Hixie>
hopefully he'll clarify

[16:35:00.0000] <jwalden>
Philip`: it's a character flaw; bz argues too much with people he should ignore, even when he knows better (and he'll be the first to admit it) :-)

[16:39:00.0000] <roc_>
everyone I know succumbs to that at times

[16:42:00.0000] <jwalden>
I tend to think bz succumbs more often than others

[17:00:00.0000] <Hixie>
I used to do that

[17:00:01.0000] <Hixie>
I designed mail filters to avoid it

[17:01:00.0000] <Hixie>
(the mail is filtered such that it takes enough effort for me to see it that I don't see it until after the thread is a few days old, at which point I don't feel the need to jump in anymore)

[18:34:00.0000] <Hixie>
should placeholder="" apply to type=password? what about type=date? type=number?

[18:35:00.0000] <Hixie>
yes no yes looks like the better answer

[18:40:00.0000] <Hixie>
/me decides yes no no

[18:58:00.0000] <jcranmer>
********* yesterday 7 ?

[19:17:00.0000] <Hixie>
hm?

[19:18:00.0000] <Hixie>
rb is such a troll it's hilarious

[19:18:01.0000] <Hixie>
his lst 6 or 7 e-mails have all said the exact same thing and avoided answering any of the actual questions he was asked

[01:00:00.0000] <yecril71>
The wording "static object" is ambiguous indeed, new or not new.

[01:00:01.0000] <yecril71>
I would rather say "snapshot" to avoid misunderstandings.

[01:03:00.0000] <Hixie>
"static" is far more correct english than "snapshot" for this case

[01:07:00.0000] <hsivonen>
is comatose politically incorrect? The kind of NodeLists the Selectors API returns are comatose, not live.

[01:08:00.0000] <hsivonen>
or a copy of the value

[01:08:01.0000] <Hixie>
"static" is the right word, people's misunderstandings notwithstanding

[01:30:00.0000] <takkaria>
7B7B7B7B7B7B7B7BThen the operations could still handle that common case of being the

[01:30:01.0000] <takkaria>
setter/getter methods.Then the operations could still handle that common case of being the

[01:30:02.0000] <takkaria>
oops, sorry

[01:30:03.0000] <takkaria>
Then the operations could still handle that common case of being the

[01:30:04.0000] <takkaria>
ack

[01:31:00.0000] <takkaria>
mouse is being a bit screwy.  and you get no points for guessing what I'm reading

[01:37:00.0000] <jruderman>
i've been to that erotic-stories site

[01:38:00.0000] <yecril71>
Tell it to K&R.

[01:39:00.0000] <yecril71>
The effect of their technical vocabulary cannot be undone, I am afraid.

[01:41:00.0000] <Hixie>
?

[01:41:01.0000] <Hixie>
i rarely understand you, yecril71 :-)

[01:42:00.0000] <yecril71>
The meaning of "static" is engraved in stone by Kerrighan and Ritchie.

[01:43:00.0000] <yecril71>
It need not be perfect English but you cannot get rid of that.

[01:43:01.0000] <yecril71>
Especially in the context of a programming language.

[01:49:00.0000] <Hixie>
the meaning of static as defined by C-like languages is pretty much exactly what we're talking about here

[01:49:01.0000] <Hixie>
so that's fine

[02:01:00.0000] <zcorpan>
Hixie: perhaps you should define "static" alongside "live" and xref it

[02:01:01.0000] <Hixie>
well it now says "new"

[02:01:02.0000] <Hixie>
which seems clear to me

[02:01:03.0000] <Hixie>
since why would you return a new object each time if it was live

[02:02:00.0000] <zcorpan>
ok

[02:05:00.0000] <hsivonen>
http://blog.jclark.com/2008/11/what-allowed-in-uri.html

[02:08:00.0000] <Hixie>
not sure what to make of that

[02:17:00.0000] <yecril71>
See e.g. the documentation of "itoa", or "localtime".

[02:17:01.0000] <yecril71>
It is not reentrant *because* it returns a static object.

[02:17:02.0000] <yecril71>
(a pointer to one)

[02:18:00.0000] <yecril71>
Your argument about "new" is valid but:

[02:18:01.0000] <yecril71>
- it can be easily overlooked

[02:19:00.0000] <yecril71>
- the statement seems contradictory.

[02:26:00.0000] <Hish>
hi. regarding the <output for"id1 id2 id3> element: is there a way for the #id1 element to find out in which output element it is used for calculation?

[02:27:00.0000] <hsivonen>
yay for placeholder

[02:30:00.0000] <Hixie>
Hish: not currently

[02:30:01.0000] <Hish>
ok. thx.

[02:41:00.0000] <takkaria>
wow, WF2 integration work is pretty much done

[02:46:00.0000] <Hixie>
getting there

[02:47:00.0000] <Hixie>
i need to go through the wf2 spec at one point line by line 

[02:47:01.0000] <Hixie>
making sure i haven't missed anything

[02:47:02.0000] <hsivonen>
Hixie: do you have plans on rewriting the meta sniffing algorithm as a state machine?

[02:48:00.0000] <Hixie>
which one?

[02:48:01.0000] <hsivonen>
Hixie: the prescan on bytes

[02:48:02.0000] <Hixie>
"8.2.2.1 Determining the character encoding"?

[02:49:00.0000] <hsivonen>
Hixie: yes

[02:49:01.0000] <Hixie>
no, no intention of rewriting that section.

[02:49:02.0000] <Hixie>
why?

[02:49:03.0000] <Hixie>
does it need it?

[02:49:04.0000] <hsivonen>
Hixie: it might be useful to implment it as a state machine at least

[02:50:00.0000] <hsivonen>
unless waiting for the first n bytes has a trivial perf hit even when the meta occurs slighly before n

[02:50:01.0000] <hsivonen>
Hixie: do you have plans on prescribing speculative script tag prescanning?

[02:50:02.0000] <Hixie>
i imagine waiting for bytes will take far longer than reading them

[02:50:03.0000] <Hixie>
no

[02:50:04.0000] <Hixie>
other than network traffic, it has no observable side-effects

[02:51:00.0000] <hsivonen>
ok

[03:48:00.0000] <mookid>
Hixie: someone said you would be the person to talk about with regards to an informal proposal

[03:48:01.0000] <Hixie>
sure

[03:48:02.0000] <mookid>
did you read any of it previously?

[03:48:03.0000] <mookid>
It was a discussion abuot an optional Accept attribute

[03:49:00.0000] <mookid>
for indicating content-type of a request

[03:49:01.0000] <mookid>
in the Accept header

[03:50:00.0000] <Hixie>
just saw an e-mail about it, filed it along with other forms-related feedback for future consideration (it can take years for me to get to replying to a particular e-mail, there's 2545 e-mails in my feedback pile right now)

[03:50:01.0000] <mookid>
it's not forms related :P

[03:50:02.0000] <mookid>
an optional tag for <a> or <img> - any hypermedia request

[03:51:00.0000] <mookid>
to indicate to the browser what Accept header to send

[03:51:01.0000] <mookid>
having this as part of HTML5 will stop encourageing developers to misuse URI's and include content-type in them

[03:52:00.0000] <Hixie>
yeah i guess it should have been put in the links folder

[03:52:01.0000] <Hixie>
no biggie

[03:52:02.0000] <Hixie>
to be honest though personally i think the whole http content negotiation feature is a big failure and we'd be better off removing the whole thing from http

[03:53:00.0000] <mookid>
it's a failure because HTML and Browser support for the rfc is pathetic? -_-

[03:53:01.0000] <Hixie>
i haven't seen many people using it, i've seen very few people needing it, and users seem to get confused by it whenever they are exposed to it

[03:53:02.0000] <mookid>
you can't use it because the markup is insufficient

[03:54:00.0000] <Hixie>
no i think it's a failure because it solves a problem that doesn't exist in a way that isn't compatible with the way most people (authors and users) think

[03:54:01.0000] <mookid>
they only think that way because the HTML they've been working with is crap.!

[03:54:02.0000] <mookid>
that's not a fair criticism of HTTP

[03:54:03.0000] <mookid>
that's a problem with HTML and browsers

[03:55:00.0000] <mookid>
content-type doesn't belong in a resource identifier, is what I'm getting at

[03:58:00.0000] <Hixie>
what problem does this solve?

[03:58:01.0000] <Hixie>
content-type is a waste of time too, we should just be using unique identification strings for content types like the PNG header

[04:04:00.0000] <mookid>
right - you're suggesting that content negotiation belongs in the URI

[04:05:00.0000] <mookid>
reading Roy Fielding's dissertation it becomes clear that this is not the intention of a URI at all

[04:06:00.0000] <mookid>
I'm suggesting a way that HTML can allow for both - I don't see how that's a bad thing

[04:06:01.0000] <mookid>
unless, of course, you understand HTTP better than Roy..! :P

[04:08:00.0000] <Hixie>
i'm suggesting that there not be any content negotiation because people don't need it or care for it

[04:11:00.0000] <Hixie>
anyway, bed time for me now

[04:11:01.0000] <Hixie>
nn

[04:14:00.0000] <hsivonen>
http://mxr.mozilla.org/mozilla-central/source/parser/htmlparser/public/nsIParser.h#89

[04:14:01.0000] <hsivonen>
that's a lot of different cases...

[04:14:02.0000] <hsivonen>
now I need to figure out which ones are tentative and which ones are confident

[04:18:00.0000] <hsivonen>
I guess < 9 are tentative and >= 9 are confident

[04:42:00.0000] <mookid>
Hixie: content negotiation is one of the main components of RESTful APIs - it allows your URI's to serve JSON/XML for application interaction and HTML for user interaction

[04:43:00.0000] <mookid>
if you take out content-negotiation that means one URI can only serve one content-type

[04:43:01.0000] <mookid>
one URI = one resource

[04:44:00.0000] <mookid>
one resource can have many representations

[04:47:00.0000] <hsivonen>
mookid: what problem is solved if to dereference a link you need a URI and an accept value and then you stick these into separate places in the HTTP request instead of baking the format identification into the URI?

[04:47:01.0000] <mookid>
because that's not the purpose of a URI

[04:47:02.0000] <hsivonen>
mookid: that's not a statement of solving a problem.

[04:48:00.0000] <mookid>
no, but the approach that - well everyone uses URIs to negotiate content so therefore it's the best way

[04:48:01.0000] <mookid>
is equally insufficient

[04:48:02.0000] <hsivonen>
why?

[04:48:03.0000] <mookid>
because it's purely the inadequacy of html and browsers that has led to that

[04:48:04.0000] <hsivonen>
sounds like dogma to me

[04:49:00.0000] <hsivonen>
not like an actual problem

[04:49:01.0000] <mookid>
have you read Roy Fielding's thesis? there are benefits from having one URI for a resource

[04:50:00.0000] <hsivonen>
mookid: I have read parts of it, but I haven't read it from start to finish.

[04:50:01.0000] <hsivonen>
mookid: can you point me to a benefit?

[04:50:02.0000] <mookid>
yes

[04:50:03.0000] <mookid>
So I can send a colleague a message; 'you can get the report at http://example.com/report', and they can use that URL in any user agent that is appropriate. A browser is a special case in which many different content-types are dealt with. The same benefit is not achieved if the content is negotiated via the URL, since the user would have to know the type their user agent required and modify the URL accordingly

[04:51:00.0000] <mookid>
it means that you can use one URI to serve several user agents

[04:51:01.0000] <hsivonen>
what other user agent would the colleague use?

[04:51:02.0000] <mookid>
they could maybe use excel

[04:51:03.0000] <mookid>
or.. word.. or wahtever

[04:51:04.0000] <mookid>
this is the direction web applications are going

[04:52:00.0000] <mookid>
I'm not making this up for nothing!

[04:52:01.0000] <hsivonen>
so isn't the accept header bit then something that the agent should generate instead of something that the the hypertext reference should contain?

[04:52:02.0000] <mookid>
browsers are different case really

[04:53:00.0000] <mookid>
they, generally,  pass the content to the OS if they can't understand it

[04:54:00.0000] <mookid>
most UA's just handle one filetype

[04:54:01.0000] <mookid>
browsers do lots - whcih is why it would be nice to have that ability reflected in html

[04:54:02.0000] <mookid>
does that make sense?

[04:55:00.0000] <hsivonen>
so if you feed them general Web URLs, you are likely to end up frustrated, which is why the user probably wants to know a priori that the URL works with a particular kind of non-browser HTTP agent

[04:55:01.0000] <hsivonen>
mookid: no, I don't think your argumentation makes sense

[04:55:02.0000] <mookid>
they will, generally, have some idea - the worst that can happen is there UA will say "nope, no good here"

[04:55:03.0000] <mookid>
well it makes sense because I'm suggesting the best solution is to provide a way to use HTTP as it was intended

[04:56:00.0000] <hsivonen>
If you send me an URL to a report, why would I try it in Excel if you didn't tell me that it has been specifically crafted to work usefully in Excel?

[04:56:01.0000] <mookid>
if anyone should be justifying themselves with a use case it should be the guys saying "content negotiation is done in URIs"

[04:57:00.0000] <mookid>
well the idea is that as the application frameworks for these kinds of APIs grow.. all user agents will be able to interact with the data

[04:57:01.0000] <mookid>
e.g. opening it in iTunes streams a voice-translated version for the blind

[04:57:02.0000] <mookid>
that type of thing

[04:57:03.0000] <hsivonen>
no, only user agents that support the particular flavor(s) of data availabale will work

[04:58:00.0000] <hsivonen>
it's useless to try a URL that deferences into KML in Excel, for example.

[04:58:01.0000] <mookid>
right.. so the worst that can happen is a UA says 'nope sorry I dont know what to do with that'

[04:58:02.0000] <hsivonen>
if I have a zillion agents in addition to a browser, do I try each one when you send me and architecturally correct URL?

[04:59:00.0000] <mookid>
well you pick the one you know/guess is appropriate

[04:59:01.0000] <mookid>
you're assuming complete ignorance on behalf of the user

[05:00:00.0000] <hsivonen>
no, I'm assuming that a user who gets an HTTP URL by email without additional instruction dereferences the URL using a browser

[05:00:01.0000] <mookid>
I would send an email out saying "here's the report: example.com/report, you can open it in excel, powerpoint, and word" - how would you do it with URIs? :)

[05:01:00.0000] <hsivonen>
Here's the report in various formats http://example.com/report.xls, http://example.com/report.ppt and http://example.com/report.doc

[05:01:01.0000] <mookid>
but if they're all the same resourcee

[05:02:00.0000] <mookid>
just different representations

[05:02:01.0000] <mookid>
then that's a misuse of the term URI

[05:02:02.0000] <mookid>
I've spent a fair amoutn of time studying this

[05:02:03.0000] <mookid>
I'm not suggesting this for fun..!

[05:02:04.0000] <hsivonen>
I posit that the case where the alternative representations of a resource are truly equivalent is so rare that it is pointless to design for it

[05:03:00.0000] <mookid>
like RSS feed and index.html ?

[05:03:01.0000] <mookid>
:)

[05:04:00.0000] <hsivonen>
an RSS feed and index.html are rarely equivalent in the sense that one didn't contain some information that the other contains

[05:04:01.0000] <mookid>
I think my report example makes sense.. the only reason these kinds of interactions arent the case is because web application development is young.. and because html discourages peolpe from using URIs as they were intended

[05:04:02.0000] <mookid>
they are completely equivalent in the majority of cases

[05:04:03.0000] <mookid>
look at blogs..

[05:04:04.0000] <mookid>
look at technorati..

[05:05:00.0000] <hsivonen>
mookid: well, depends on at what point in time you cast "intended" in stone

[05:05:01.0000] <mookid>
well..

[05:05:02.0000] <hsivonen>
mookid: the original HTTP didn't even have content types

[05:05:03.0000] <mookid>
I'd call an rfc.. an intention

[05:05:04.0000] <mookid>
HTTP 1.1 did

[05:05:05.0000] <mookid>
and does

[05:05:06.0000] <mookid>
so that's a mute point

[05:05:07.0000] <hsivonen>
so one might write Accept off as later day astronauting ;-)

[05:05:08.0000] <mookid>
?

[05:06:00.0000] <mookid>
you aren't doing a very good job of convincing me I have to say

[05:06:01.0000] <hsivonen>
mookid: HTTP conneg is younger than URLs

[05:06:02.0000] <mookid>
I've spent a fair while looking at this

[05:06:03.0000] <hsivonen>
I could say the same thing.

[05:06:04.0000] <mookid>
well I'm looking at how it is defined now.. and how people are using it moving forwards (which is presumably what HTML5 should be looking at)

[05:06:05.0000] <mookid>
well you admitted that you havent read that much into it

[05:07:00.0000] <mookid>
why would you not want to give developers the option at least?

[05:07:01.0000] <hsivonen>
I haven't read the canonical theory in its entirety. I have considered the practical side quite a bit.

[05:07:02.0000] <mookid>
how can you do that if you dont have a full grasp of the concepts?

[05:07:03.0000] <hsivonen>
mookid: giving the option is not free

[05:07:04.0000] <hsivonen>
mookid: also, conneg sucks for seachability

[05:07:05.0000] <mookid>
free?

[05:08:00.0000] <mookid>
does it?

[05:08:01.0000] <hsivonen>
mookid: implementing the feature has an opportunity cost: the developers wouldn't be doing something else

[05:09:00.0000] <mookid>
these kinds of functionality is what helps the framework developers

[05:09:01.0000] <mookid>
if you can standardise this stuff it will assist tooling for proper HTTP transactions

[05:09:02.0000] <hsivonen>
mookid: yes, without representation-specific URIs and with a countably infinite set of potential Accept values, how do you index all representations of a resource?

[05:10:00.0000] <mookid>
that's an issue of producing a format for spiders

[05:10:01.0000] <mookid>
that's a non issue if all it requires is another content type built specifically for crawlers

[05:10:02.0000] <hsivonen>
well, doing that isn't costless, either

[05:10:03.0000] <hsivonen>
afk, I'll be back

[05:10:04.0000] <mookid>
depends how your aplpication is built

[05:10:05.0000] <mookid>
if you're just providing an extra view to a model

[05:10:06.0000] <mookid>
it's a 2 second job

[05:11:00.0000] <mookid>
across the board

[05:12:00.0000] <mookid>
content negotiation is bad right now for searchability because it *is* done in the URI

[05:12:01.0000] <mookid>
that's the point I'm trying to get across here

[05:13:00.0000] <mookid>
google are championing REST for a reason

[05:13:01.0000] <mookid>
:)

[05:14:00.0000] <takkaria>
relying on authors to provide an index of their representations to spiders seems very prone to spamming attempts

[05:15:00.0000] <Philip`>
Why is it any more prone to that than providing a separate HTML document which links to lots of other URIs that spiders will follow?

[05:17:00.0000] <Philip`>
(Unintentional errors seem a more important problem, because people would forget about their spider-specific index and it would be buggy or out-of-date)

[05:17:01.0000] <takkaria>
fair point

[05:20:00.0000] <hsivonen>
mookid: is google really championing REST according to Fielding or POX over HTTP Web services with a lower case s?

[05:21:00.0000] <hsivonen>
I'd argue that AtomPub-based services are their own class of services with Atom-aware clients. They aren't totally generic conneg

[05:44:00.0000] <hsivonen>
zcorpan: I believe I've now addressed the problems that blocked building v.nu on Windows

[05:59:00.0000] <MikeSmith>
/me wonders why/whether "using HTTP as it was intended" should be a goal

[06:00:00.0000] <MikeSmith>
I would doubt that most people would say we are now using HTML as it was originally intended

[06:03:00.0000] <hsivonen>
looks like I need to expand my fragment parsing API to deal with fragments rooted at foreign content...

[06:19:00.0000] <hsivonen>
when a document has been created with document.open() and everything is document.written, does the first level of document.write count as being equivalent to a network stream or does it count as the first level of document write?

[06:19:01.0000] <hsivonen>
I assume the former considering the behavior of Hixie's live dom viewer

[06:24:00.0000] <zcorpan>
hsivonen: that's nice

[06:27:00.0000] <hsivonen>
/me finds

[06:27:01.0000] <hsivonen>
  // Hack to pass on to the dtd the caller's desire to

[06:27:02.0000] <hsivonen>
  // parse a fragment without worrying about containment rules

[06:27:03.0000] <hsivonen>
  if (aMode == eDTDMode_fragment)

[06:27:04.0000] <hsivonen>
    mCommand = eViewFragment;

[06:28:00.0000] <hsivonen>
Hixie: have you studied the purpose/impact of Gecko having some kind of magic fragment mode for containment rules?

[06:54:00.0000] <pergj>
hsivonen: I don't remember seeing anything in the HTML5 spec that would give different behaviour for document.write immediately after document.open

[06:56:00.0000] <hsivonen>
pergj: but if document.write after document.open counted as the first level of document.write, how could Hixie's live dom viewer have the right document.write semantics?

[06:58:00.0000] <pergj>
hm... Not sure...

[08:58:00.0000] <BenMillard>
takkaria, you just sent 5 empty messages. So now who's "fail"! :P

[08:59:00.0000] <Dashiva>
They're not empty, they're images encoded with lenpeg v2

[09:00:00.0000] <BenMillard>
oh, they show up empty in Opera and the logs

[09:03:00.0000] <Dashiva>
http://www.dangermouse.net/esoteric/lenpeg.html

[09:03:01.0000] <BenMillard>
yeah, I've just reached the bottom of that :(

[09:05:00.0000] <takkaria>
ah, oops

[09:05:01.0000] <takkaria>
my mouse is doing weird things today

[09:06:00.0000] <takkaria>
mixed with X11's middle-click-is-paste thing, bad things happen

[09:16:00.0000] <MikeSmith>
takkaria: your mouse should meet Hixie's cats

[09:16:01.0000] <takkaria>
:)

[09:30:00.0000] <BenMillard>
MikeSmith, nice job on the TPAC 2008 minutes and overview page.

[09:33:00.0000] <MikeSmith>
BenMillard: thanks for your prompting. I still need to add some of the links you suggested

[09:41:00.0000] <jmb>
so, a really unscientific measure of how complex the html5 parsing algorithm is in comparison with a pre-existing html4 parser:

[09:41:01.0000] <jmb>
sloccount on libxml's HTML{parser,tree}.c => 5188 sloc

[09:42:00.0000] <jmb>
sloccount on hubbub's tokeniser and treebuilder source trees => 9127 sloc

[09:44:00.0000] <takkaria>
I imagine hubbub has more boilerplate than lixml though

[09:45:00.0000] <jmb>
could well. hence "unscientific" :)

[09:50:00.0000] <jmb>
oh, in case anyone cares, that's libxml 2.7.2, and hubbub svn head

[09:56:00.0000] <gsnedders>
First two thoughts when writing CV: 1) What do I put in it? 2) How do I mark it up?

[09:57:00.0000] <takkaria>
look at Tim Bray's

[09:57:01.0000] <gsnedders>
/me doesn't like the markup of Hixie's :)

[09:58:00.0000] <takkaria>
http://www.tbray.org/ongoing/When/200x/2005/11/12/Template.html

[09:59:00.0000] <Philip`>
Nobody cares about the markup, since they'll just print it out :-p

[09:59:01.0000] <gsnedders>
Is it really a table though?

[10:04:00.0000] <takkaria>
it's not really a table, I don't think, but I don't think it matters 

[10:04:01.0000] <takkaria>
you evidently do :)

[10:04:02.0000] <gsnedders>
http://alexking.org/site/projects/html-resume-template/demo/resume.php semms a lot closer

[10:10:00.0000] <BenMillard>
gsnedders, mine is here and I was told "your CV rocks" by 1 potential employer: http://sitesurgeon.co.uk/

[10:10:01.0000] <gsnedders>
How ugly.

[10:10:02.0000] <gsnedders>
/me ducks

[10:10:03.0000] <BenMillard>
yeah, the visual aspects aren't my strong point :)

[10:10:04.0000] <BenMillard>
the markup and content is what you seemed interested by

[10:11:00.0000] <BenMillard>
anyway, dinnertime now, cya

[10:38:00.0000] <gsnedders>
Prince doesn't support meta@charset seemingly

[10:42:00.0000] <takkaria>
no, it doesn't

[10:51:00.0000] <jcranmer>
I presume if I had <input type="xxxinvalid">, would getting .type return "text" ?

[10:51:01.0000] <jcranmer>
s/would/then/

[11:05:00.0000] <gsnedders>
jcranmer: yes

[11:05:01.0000] <jcranmer>
ah, good :-)

[11:27:00.0000] <Dashiva>
public-hmtl is back in high gear again, I see

[11:29:00.0000] <takkaria>
lots of revs but the break is on

[12:15:00.0000] <Lachy>
Hixie, Thanks for finally adding placeholder :-)

[15:34:00.0000] <Hixie>
man i need to eat before i can reply to the crazy stuff on the w3c html lists

[15:35:00.0000] <Hixie>
roy fielding in particular appears to live on a different planet

[15:37:00.0000] <roc>
the one email I read of his suggested that testing your HTML in a browser is bad because you might be tempted to make it less pure

[15:40:00.0000] <jcranmer>
bwahuh?

[15:40:01.0000] <gavin>
But now that Firefox is getting just as buggy and complex as

[15:40:02.0000] <gavin>
the other major browsers, they have no reason to switch at all.

[15:40:03.0000] <gavin>
Firefox usage hasn't increased since it decided to be no better

[15:40:04.0000] <gavin>
than the others.

[15:41:00.0000] <gavin>
that's an interesting assertion!

[15:41:01.0000] <jcranmer>
numerous factual errors in ther

[15:41:02.0000] <jcranmer>
there

[15:41:03.0000] <jcranmer>
shall I pull out the mandatory xkcd reference?

[15:42:00.0000] <Lachy>
jcranmer, xkcd references are always funny

[15:45:00.0000] <Philip`>
jcranmer: XKCD references are getting a bit tired now :-p

[15:46:00.0000] <jcranmer>
it's the one where the character responds to trolls with long, multi-paragraph replies detailing why the poster should just die

[15:49:00.0000] <Lachy>
http://xkcd.com/493/

[15:49:01.0000] <Lachy>
is that the one you meant?

[15:50:00.0000] <jcranmer>
no, not that one

[15:51:00.0000] <Lachy>
ok, then I can't find it

[15:52:00.0000] <jcranmer>
I'm not finding it either

[15:52:01.0000] <svl>
This one: http://xkcd.com/406/ ?

[15:54:00.0000] <jcranmer>
yes!


2008-11-18
[16:20:00.0000] <sicking>
Hixie, ping

[16:20:01.0000] <Hixie>
wassup?

[16:22:00.0000] <sicking>
nevermind, shift-reload fixed it :)

[16:22:01.0000] <sicking>
though found a bug in workers spec:

[16:23:00.0000] <sicking>
"If the close event that this algorithm just added hasn't yet been dispatched, then abort the script currently running in the worker."

[16:24:00.0000] <sicking>
somewhere in there it should say to dispatch the close event i would have thought

[16:43:00.0000] <Hixie>
sicking: send mail? (just to me if you want, you can just paste your irc comments)

[16:43:01.0000] <Hixie>
i'm about to be afk

[16:43:02.0000] <Hixie>
/me sends his reply to Roy

[16:43:03.0000] <Hixie>
hopefully we'll get some clarity

[16:44:00.0000] <sicking>
/me is about to give up on discussion with roy

[16:44:01.0000] <sicking>
or rather, i already have

[16:44:02.0000] <sicking>
the question is if that should mean giving up on the whole discussion with TAG or not

[16:44:03.0000] <sicking>
i hope it doesn't mean that

[16:45:00.0000] <Hixie>
i wish the tag people would give clear explanations of wtf they want

[16:45:01.0000] <Hixie>
they all seem to want something different

[16:45:02.0000] <Hixie>
yet they use the same terms

[16:45:03.0000] <Hixie>
and in half the cases, we already do what they want

[16:45:04.0000] <Hixie>
anyway

[16:45:05.0000] <Hixie>
afk

[16:50:00.0000] <roc>
wait, is Roy Fielding in TAG?

[16:50:01.0000] <Philip`>
http://junkyard.damowmow.com/284 sort of backs up the idea that most static content was written before IE6, if you assume that all pages with last-modified dates are probably static content (because nobody bothers generating that header in dynamic pages) and that ~52% counts as "most" and if you ignore the obvious bogosity that leads to 13% of pages being last modified in the future

[16:51:00.0000] <Philip`>
roc: http://www.w3.org/2001/tag/ says no

[16:52:00.0000] <roc>
ew

[16:52:01.0000] <roc>
phew

[17:04:00.0000] <sicking>
wait, he's not in the TAG?

[17:04:01.0000] <sicking>
well then

[17:07:00.0000] <othermaciej_>
Philip`: any mod dates before 1990 are also probably false

[17:08:00.0000] <othermaciej_>
in fact probably most of the 1992s are false, given how few pages were on the web then

[17:22:00.0000] <Hixie>
/me considers renaming HTML5 to Web5 and seeing what people think of _that_

[17:23:00.0000] <roc>
I recommend just "5"

[17:23:01.0000] <roc>
then people won't be able to complain about scope creep

[17:24:00.0000] <Hixie>
speaking of which, if we expand the scope we could include all of the DOM specs, CSS, HTTP, and XML into the same spec

[17:24:01.0000] <Hixie>
and SVG, too, while we're at it

[17:24:02.0000] <Hixie>
and MathML

[17:24:03.0000] <Philip`>
roc: Why bother with the version information at all? Just call it ""

[17:25:00.0000] <roc>
that's even harder to Google for

[17:25:01.0000] <Hixie>
that would solve that problem of multiple specs fragmenting the platform into different silos of technology

[17:34:00.0000] <sicking>
Hixie, well, so seems like this Roy guy is just a troll that we should ignore

[17:34:01.0000] <Hixie>
he's one the apache developers and vocal on the http wg

[17:35:00.0000] <Hixie>
i don't disagree that he's a troll, or at least, that he has opinions so far removed from what i see as reality that it has not been productive to communicate with him in general.

[17:35:01.0000] <sicking>
yep

[17:36:00.0000] <othermaciej>
he is considered an Important Web Architecture Person

[17:36:01.0000] <Hixie>
by whom?

[17:36:02.0000] <othermaciej>
see, that's why I used the passive voice

[17:37:00.0000] <sicking>
/me should do that more too

[17:37:01.0000] <othermaciej>
as a result of the acronym "REST" becoming popular, and him having invented it, many think therefore that he is an expert on Web architecture

[17:38:00.0000] <Hixie>
oh the REST nonsense is _his_ fault?!

[17:38:01.0000] <Hixie>
good to know

[17:38:02.0000] <othermaciej>
it was his PhD thesis

[17:38:03.0000] <othermaciej>
http://www.ics.uci.edu/~fielding/pubs/dissertation/top.htm

[17:39:00.0000] <Hixie>
i wish phd students were more like abarth and cjackson

[17:39:01.0000] <Hixie>
than roy and rburns :-)

[17:46:00.0000] <nessy>
we're not talking roy fielding, are we?

[17:46:01.0000] <gavin>
we are

[17:46:02.0000] <blooberry>
roy fielding? He's one of the main authors of the HTTP spec

[17:47:00.0000] <nessy>
http://www.ietf.org/rfc/rfc2396.txt

[17:47:01.0000] <nessy>
and of the original uri rfc

[17:47:02.0000] <nessy>
http://www.ietf.org/rfc/rfc2068.txt

[17:52:00.0000] <Hixie>
he's also the author of http://lists.w3.org/Archives/Public/public-html/2008Nov/0216.html

[17:55:00.0000] <roc>
ah

[17:55:01.0000] <roc>
gah

[17:55:02.0000] <Hixie>
that e-mail is fractally wrong, as far as i can tell

[17:55:03.0000] <Hixie>
that is, the more you study its flaws, the more flaws you find

[17:56:00.0000] <Hixie>
http://lists.w3.org/Archives/Public/public-html/2008Nov/0219.html was my reply

[17:56:01.0000] <roc>
yeah

[17:59:00.0000] <roc>
I'd be worried about responding ... it seems impossible to address all flaws, and therefore one risks appearing to tacitly endorse them

[18:00:00.0000] <Hixie>
did i miss any flaws? :-)

[18:03:00.0000] <roc>
"the original Firefox team has moved on"

[18:03:01.0000] <gavin>
bz corrected that one

[18:03:02.0000] <roc>
"in a year or two, there will be other fresh ideas on browsing

[18:03:03.0000] <roc>
implementations.  Such has been the case for over 15 years now"

[18:04:00.0000] <Hixie>
i didn't even know where to begin with those two

[18:04:01.0000] <Hixie>
bz addressed the first one

[18:04:02.0000] <Hixie>
the second... i didn't understand it

[18:04:03.0000] <Hixie>
i mean, it's not untrue

[18:04:04.0000] <Hixie>
and i couldn't tell what assumed disagreement made him want to specify it explicitly

[18:06:00.0000] <roc>
he seems to be talking about browser engines, and no new competitive browser engine has appeared in the last five years

[18:06:01.0000] <othermaciej>
I think WebKit is the most recent fresh idea in browser engines

[18:06:02.0000] <othermaciej>
that has achieved any serious level of market success

[18:06:03.0000] <Hixie>
it's the last one i know about for sure

[18:06:04.0000] <roc>
right

[18:06:05.0000] <othermaciej>
and certainly innovation in html parsing was not something we ever did or wanted to do

[18:07:00.0000] <Hixie>
and one major point of html5 is to help make competition easier

[18:07:01.0000] <Hixie>
so as to introduce more such browsers

[18:07:02.0000] <Hixie>
so who knows what he meant

[18:07:03.0000] <roc>
anyway, measuring the wrongness of Roy's position with great precision is not a good use of our time

[18:07:04.0000] <Hixie>
but it is somewhat entertaining

[18:08:00.0000] <roc>
it's as compelling as any car wreck

[18:08:01.0000] <Hixie>
(more seriously, i really do wish i could understand his position, as it would help us address his concerns and hopefully reduce the number of times he complains)

[18:13:00.0000] <blooberry>
has he brought forward any of these concerns in the past in the WG?

[18:13:01.0000] <blooberry>
or is he apparently having a Bad Day (eg: he has always had some disagreements and something finally made him "snap")

[18:14:00.0000] <Hixie>
he's always complaining about html5

[18:19:00.0000] <roc>
I think his position is quite clear. He has some false assumptions about Web browsers and Web developers, which you identified, and everything else flows from there.

[18:21:00.0000] <Hixie>
well hopefully we can correct those mistaken assumptions

[18:21:01.0000] <roc>
I'll bet money you can't

[18:24:00.0000] <roc>
you may impact lurkers who share some of those unexamined assumptions, though, so play on

[19:16:00.0000] <Hixie>
apparently i ignore feedback

[19:16:01.0000] <Hixie>
wow

[19:16:02.0000] <Hixie>
i wonder what these thousands of e-mails i've been replying to have been, if not feedback

[19:29:00.0000] <Hixie>
hmm

[19:29:01.0000] <Hixie>
window.resolveURL('test.html') or window.location.resolveURL('test.html') ?

[19:29:02.0000] <Hixie>
the former has more chance of clashing with scripts

[19:30:00.0000] <Hixie>
the latter might mislead authors when there's a <base> around

[19:30:01.0000] <sicking>
othermaciej, ping

[19:30:02.0000] <othermaciej>
sicking: yo

[19:30:03.0000] <othermaciej>
/me has a half-composed reply to Roy that he thinks better of

[19:30:04.0000] <sicking>
othermaciej, do you know where the ECMA spec for JSON lives?

[19:31:00.0000] <othermaciej>
sicking: you mean for the soon-to-be-standard JSON parsing API, or for the syntax of JSON?

[19:31:01.0000] <sicking>
othermaciej, both actually

[19:31:02.0000] <othermaciej>
sicking: the former is in ES3.1 drafts (see es3.x-discuss mailing list for latest draft link) and the latter is I believe an RFC

[19:31:03.0000] <sicking>
othermaciej, the latter more so

[19:32:00.0000] <othermaciej>
(IETF RFC)

[19:32:01.0000] <jcranmer>
http://www.ietf.org/rfc/rfc4627.txt

[19:32:02.0000] <sicking>
thanks!

[19:33:00.0000] <jcranmer>
thank awesomebar

[19:33:01.0000] <sicking>
othermaciej, is there a reason why it requires you to quote all attribute names? With double-quotes none the less?

[19:33:02.0000] <sicking>
othermaciej, took me quite a while to figure that one out

[19:34:00.0000] <othermaciej>
sicking: I am at a loss to explain any of the details of the JSON sytntax

[19:34:01.0000] <Hixie>
(fyi, s/none the less/no less/, probably)

[19:34:02.0000] <Hixie>
JSON syntax is Doug Crockford's brainchild

[19:34:03.0000] <Hixie>
as i understand it

[19:35:00.0000] <sicking>
bah, bummer

[19:35:01.0000] <Hixie>
ok window.location.resolveURL('foo'); will resolve the url 'foo' to an absolute URL

[19:35:02.0000] <Hixie>
then you can use that to pass the URL to a worker

[19:35:03.0000] <Hixie>
any objections?

[19:36:00.0000] <Hixie>
(sicking?)

[19:36:01.0000] <sicking>
Hixie, hmm.. the base thing seems like a problem :(

[19:36:02.0000] <sicking>
Hixie, or will that resolve against the base?

[19:36:03.0000] <Hixie>
it'll resolve against the base

[19:36:04.0000] <Hixie>
which is mildly confusing since Location doesn't have the base

[19:36:05.0000] <Hixie>
but oh well

[19:36:06.0000] <sicking>
right

[19:36:07.0000] <Hixie>
i'm scared of putting it on Window

[19:37:00.0000] <sicking>
don't feel strongly either way

[19:37:01.0000] <sicking>
right

[19:37:02.0000] <Hixie>
k

[19:37:03.0000] <Hixie>
well we'll try that

[19:37:04.0000] <Hixie>
and see who whines :-)

[19:40:00.0000] <sicking>
othermaciej, so i'd really like to be able to pass primitive values, such as 0 and false, through JSON.stringify/JSON.parse

[19:40:01.0000] <sicking>
othermaciej, i suppose that's not likely to happen at this point, right?

[19:40:02.0000] <othermaciej>
sicking: I haven't read up on what is spec'd for JSON

[19:40:03.0000] <othermaciej>
sicking: I think there is a desire to remain compatible with json2.js unless there is a major reason to do otherwise

[19:41:00.0000] <sicking>
othermaciej, ok

[19:41:01.0000] <othermaciej>
I think the root of a JSON structure has to be an array or object

[19:41:02.0000] <othermaciej>
so the strings "0" or "false" would not be valid JSON

[19:41:03.0000] <sicking>
othermaciej, the reason is that I want to allow JSON "values" to be passed through a bunch of the DOM APIs, such as postMessage or pushState

[19:42:00.0000] <sicking>
othermaciej, yeah, it seemed like that was what our JSON.stringify implementation accepted

[19:42:01.0000] <othermaciej>
sicking: that does seem like a good idea in general

[19:42:02.0000] <jcranmer>
   A JSON text is a serialized object or array.

[19:42:03.0000] <sicking>
othermaciej, how so?

[19:42:04.0000] <othermaciej>
sicking: your idea

[19:42:05.0000] <Hixie>
sicking: i think when we define what is passed through postMessage(), we'll use something that is wider than JSON

[19:42:06.0000] <othermaciej>
" I want to allow JSON "values" to be passed through a bunch of the DOM APIs, such as postMessage or pushState"

[19:42:07.0000] <othermaciej>
I think you would want to define an extended JSON that can be either a JS primitive value, or a JSON-compatible object

[19:42:08.0000] <sicking>
i know what i said :) i was wondering why you didn't think its a good idea

[19:43:00.0000] <othermaciej>
I do think it is a good idea!

[19:43:01.0000] <othermaciej>
I said: "sicking: that does seem like a good idea in general"

[19:43:02.0000] <sicking>
oh :)

[19:43:03.0000] <Hixie>
man, you're so negative

[19:43:04.0000] <Hixie>
always saying ideas are bad :-P

[19:43:05.0000] <sicking>
what can i say

[19:43:06.0000] <sicking>
i hate stuff

[19:43:07.0000] <Hixie>
i meant maciej :-P

[19:43:08.0000] <othermaciej>
clearly it is what people expect of me

[19:43:09.0000] <jcranmer>
A JSON parser MAY accept non-JSON forms or extensions.

[19:44:00.0000] <Hixie>
go interoperability

[19:44:01.0000] <jcranmer>
I'd be surprised to find a JSON parser that couldn't parse the string "0"

[19:44:02.0000] <othermaciej>
sicking: anyway, I would make postMessage and pushState eventually use a string conversion algorithm which accepts either primitives or JSON-compatible objects, and deserialize on the other end

[19:45:00.0000] <othermaciej>
sicking: actually I guess it is not necessary per spec for it to even stringify; the implementation could serialize and deserialize however it wants

[19:45:01.0000] <sicking>
right, that's what bens patch does

[19:45:02.0000] <sicking>
(well, it does stringify, though that's not technically needed)

[19:46:00.0000] <Hixie>
right, i would just define a set of things that must be supported, and everything else would just be stringified

[19:46:01.0000] <Hixie>
not sure how to define it yet

[19:46:02.0000] <Hixie>
it's one reason i haven't specced it yet

[19:47:00.0000] <othermaciej>
one question is what to do with non-JSON-compatible object graphs

[19:47:01.0000] <othermaciej>
otherWin.postMessage({"a": 1, "b": 2, "c": document.body})

[19:48:00.0000] <othermaciej>
I can think of at least 3 reasonable things to do

[19:48:01.0000] <othermaciej>
gotta head home

[19:48:02.0000] <othermaciej>
later folks

[19:53:00.0000] <Hixie>
null, exception, stringify?

[19:53:01.0000] <othermaciej>
null was not one I thought of

[19:53:02.0000] <othermaciej>
the two different ones were stringify generically at top level, vs stringify at JSON-compatibility failure point

[19:54:00.0000] <othermaciej>
so [object Object]

[19:54:01.0000] <othermaciej>
or something with [object HTMLBodyElement] somewhere in the middle

[19:54:02.0000] <othermaciej>
ciao

[19:55:00.0000] <deltab>
sicking: quoting names mean that there doesn't have to be a syntax rule for identifiers

[19:57:00.0000] <sicking>
deltab, how do you mean? {this: 0} should parse as {"this": 0} no?

[19:59:00.0000] <deltab>
yes, but {"foo:bar": 0} isn't equivalent to {foo:bar: 0}

[20:00:00.0000] <sicking>
sure

[20:00:01.0000] <sicking>
there can be parse errors

[20:01:00.0000] <sicking>
anyway, the discussion is moot since a decision has been made

[20:01:01.0000] <deltab>
I just thought you'd like to know why

[20:01:02.0000] <sicking>
sure, if you know why i'm curious :)

[20:02:00.0000] <deltab>
I don't

[20:02:01.0000] <sicking>
ah :)

[20:02:02.0000] <deltab>
I guess you're writing something

[20:03:00.0000] <sicking>
deltab, http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2008-November/017276.html

[20:03:01.0000] <sicking>
deltab, though unrelated to the quoting issue

[20:04:00.0000] <sicking>
deltab, the quoting issue came up because it took me about 5 times as long to test out the JSON support

[20:05:00.0000] <Hixie>
sicking: btw i definitely agree with the need to do this, and will be adding it in due course, though i'd really like to see more stability in those sections and implementation experience (thanks for sending this!) before i spec it out formally

[20:05:01.0000] <sicking>
Hixie, sounds good. Mostly wanted to check that people wasn't gonna get upset if we added support

[20:06:00.0000] <Hixie>
i don't think anyone thinks it's a bad idea, no. certainly i haven't heard anyone say it's a bad idea, and i've been floating around for a while.

[20:06:01.0000] <Hixie>
i'm a little apprehensive about how i'm gonna spec it, but we'll see

[20:22:00.0000] <sicking>
aboodman: you gotta stop top-quoting ;)

[20:23:00.0000] <jcranmer>
A: Because I said so.

[20:23:01.0000] <jcranmer>
Q: Why?

[20:24:00.0000] <aboodman2>
sicking: I usually bottom quote, not sure what was with me today.

[20:24:01.0000] <sicking>
aboodman2, cool :)

[20:24:02.0000] <aboodman2>
sicking: you top quoted too

[20:24:03.0000] <aboodman2>
sicking: were you just sticking to my convention or something?

[20:25:00.0000] <aboodman2>
I suppose I am getting lazy.

[20:25:01.0000] <sicking>
yeah, it's hard to reply to a top-quoted mail while bottom-quoting

[20:25:02.0000] <sicking>
i modified someone elses mail to bottom quote the other day and it seemed a bit rude

[20:25:03.0000] <sicking>
so i'll just be rude in here instead :)

[20:25:04.0000] <jcranmer>
bah, it's the internt

[20:25:05.0000] <jcranmer>
er, Internet

[20:26:00.0000] <jcranmer>
everything you're say is expected to be mangled and taken out of context

[20:26:01.0000] <sicking>
/me slaps jcranmer with a wet salmon

[20:26:02.0000] <sicking>
oh, sorry, i thought your point was that it's Internet, you're supposed to be rude :)

[20:26:03.0000] <aboodman2>
sicking: meh

[20:27:00.0000] <jcranmer>
oh, goodness gracious, no!

[20:27:01.0000] <jcranmer>
no one would be surprised if you were rude, though

[20:30:00.0000] <sicking>
/me slaps jcranmer with a pickled herring

[20:30:01.0000] <jcranmer>
/me eats the pickled herring

[20:30:02.0000] <sicking>
care for some sour-cream with it?

[20:31:00.0000] <jcranmer>
no, just ketchup

[21:40:00.0000] <Hixie>
sicking: i reformat and fix spelling and grammar mistakes in people's e-mails all the time :-)

[21:42:00.0000] <sicking>
Hixie, i'm a nicer guy than you :)

[21:42:01.0000] <sicking>
/me continues his new embrace of the way of the Internet to be rude

[21:42:02.0000] <Hixie>
sicking: :-P

[21:43:00.0000] <sicking>
sorry, in a grumpy mood today, backporting stupid patches to old stupid releases

[21:43:01.0000] <aboodman2>
the irc protocl makes me grumpy

[21:43:02.0000] <aboodman2>
protocol*

[23:26:00.0000] <Hixie>
hm the content model of <menu> is all wrong

[01:11:00.0000] <hsivonen>
As far as I can tell, the JSON practice towards errors in Draconianness

[01:11:01.0000] <hsivonen>
except when its not

[01:11:02.0000] <hsivonen>
which leads to interop trouble

[01:12:00.0000] <Hixie>
json is a mess

[01:14:00.0000] <hsivonen>
s/in/is/

[01:17:00.0000] <hsivonen>
Hixie: do you have document.write test cases that I should walk through both with the spec and with the idea I sent to the list?

[01:18:00.0000] <Philip`>
Has anyone tested IE8's JSON implementation to see how spec-compliant it is?

[01:18:01.0000] <Hixie>
hsivonen: don't think so

[01:19:00.0000] <hsivonen>
Hixie: ok. did what I sent to the list look superficially reasonable?

[01:20:00.0000] <hsivonen>
I'm expecting that the following properties are true of the spec:

[01:20:01.0000] <hsivonen>
 1) If a given script has not document.written before, the insertion point is where the tokenization position is

[01:21:00.0000] <hsivonen>
 2) When the tokenizer reaches an insertion point, nothing can write to that insertion point any longer

[01:21:01.0000] <hsivonen>
(or if something document.writes to that point, it degenerates into case #1)

[01:21:02.0000] <Hixie>
i cannot offhand tell if the implementation strategy you describe is equivalent to the spec

[01:22:00.0000] <hsivonen>
Hixie: OK. I guess I'll try it and see if there are issues that I didn't consider

[01:22:01.0000] <hsivonen>
Hixie: thanks

[01:23:00.0000] <Hixie>
i'd just make an object that exposes a string that represents the concept in the spec, personally

[01:23:01.0000] <Hixie>
i.e. an object that has an insertion point and a current position

[01:23:02.0000] <Hixie>
and that keeps track of what strings have been inserted where

[01:24:00.0000] <Hixie>
so that it is transparent to the reader

[01:24:01.0000] <Hixie>
it could even do the CRLF expansion stuff transparently

[01:25:00.0000] <hsivonen>
doing CRLF expansion strictly before tokenization seems inefficient

[01:25:01.0000] <Hixie>
yeah this might not be the most efficient solution, certainly

[01:26:00.0000] <hsivonen>
Considering that HTML parsing is perf critical to a browser, I think I'm going to tolerate abstraction leaks here

[01:26:01.0000] <hsivonen>
or rather, not having a proper abstraction for the UTF-16 stream at all

[01:27:00.0000] <hsivonen>
Specifically, once the Unicode decoder has written data to a range of memory, I don't want to move that data

[01:28:00.0000] <Hixie>
yeah i was going to say that in general i advise against premature optimisation but in this case it may be a case where we know that that part will need optimising anyway

[01:51:00.0000] <hsivonen>
clearly, Roy Fielding has a different view of who are experts in the Web standards community

[01:52:00.0000] <jwalden>
sicking: so one reason, I think, is that JSON is then also a subset of Python syntax, but I don't know if that's an actual reason; simplicity might be, tho (same way XML requires attributes be quoted or something)

[01:53:00.0000] <Philip`>
jwalden: It's not a subset - Python doesn't have true and false and null

[01:54:00.0000] <hsivonen>
Philip`: but one could do

[01:54:01.0000] <hsivonen>
true = 1

[01:54:02.0000] <hsivonen>
false = 0

[01:54:03.0000] <hsivonen>
null = None

[01:54:04.0000] <hsivonen>
as boilerplate

[01:55:00.0000] <hsivonen>
although in practice, parsing JSON using anything but a dedicated JSON parser leads to interop problems

[01:55:01.0000] <Philip`>
Also Python doesn't seem to understand "\u1234"

[01:55:02.0000] <Philip`>
(since you have to do u"\u1234" instead)

[01:55:03.0000] <hsivonen>
good point

[01:58:00.0000] <Philip`>
That's why JSON is dangerous - people think they can parse it easily just by twiddling a few bits and using "eval" in their favourite language, and it sorts of works but isn't quite interoperable

[01:59:00.0000] <hsivonen>
I wonder what classes of products Roy has in mind that can't comply to HTML5 as written

[01:59:01.0000] <jwalden>
er, hm

[01:59:02.0000] <hsivonen>
/me avoids getting involved in the thread, though

[01:59:03.0000] <jwalden>
I could have sworn I read somewhere about that being true

[02:01:00.0000] <Philip`>
Oh, and even if Python did understand "\u1234" then it probably wouldn't understand "\ud800\udc00 etc" for non-BMP characters

[02:04:00.0000] <Philip`>
"Those [authoring] tools are developed based on language specifications and generic templates, not by "designing for current browsers" ..." - is he failing to consider that people actually design and develop and extend the markup templates and fragments generated by those tools, and that's basically the same as hand authoring?

[02:06:00.0000] <Philip`>
People write e.g. blog templates by writing some HTML in a text editor, telling the blog software to use it, then repeatedly testing it in their current browser and tweaking it until it looks alright

[02:07:00.0000] <Philip`>
and when we talk about "authors" we mean mainly those people (as far as I'm aware) rather than the people writing the textual content

[02:11:00.0000] <Hixie>
/me replies to roy

[02:15:00.0000] <Hixie>
ok bed time

[02:15:01.0000] <Hixie>
nn

[02:16:00.0000] <roc>
you made me look, and now I'm unhappy again

[02:17:00.0000] <Hixie>
aww

[02:17:01.0000] <Hixie>
i'm sorry

[02:18:00.0000] <roc>
people who say things like "a resident memory size of 141M is ridiculous", with no context at all, are very annoying

[02:19:00.0000] <hsivonen>
I'm sure there's a non-contrived context where Apache 2 takes 141 MB of resident memory

[02:20:00.0000] <Hixie>
my apache instance is taking upwards of half a gig

[02:20:01.0000] <Hixie>
20mb per thread

[02:20:02.0000] <Hixie>
which is pretty insane too

[02:20:03.0000] <Philip`>
Resident, and not shared etc?

[02:21:00.0000] <Philip`>
/me tends not to worry until Opera's using ~1GB of memory, and then he restarts it and carries on

[02:22:00.0000] <Hixie>
my machine as a whole has 0kb swapped out, and about all that's running on it is apache

[02:22:01.0000] <Hixie>
and the total amount used is about 0.7gb

[02:22:02.0000] <Hixie>
so

[02:23:00.0000] <Philip`>
/me likes writing games, because you can get away with using half a gigabyte of RAM and nobody minds at all

[03:00:00.0000] <yecril71>
If an URI should mark a resource regardless of its form,

[03:01:00.0000] <yecril71>
the decision on which form to use should be entirely on the user agent.

[03:01:01.0000] <yecril71>
The author should have no say here.

[03:04:00.0000] <yecril71>
PUT and DELETE are outside the scope of user agents.

[03:04:01.0000] <yecril71>
User agents are for viewing content, not for publishing it.

[03:05:00.0000] <yecril71>
(except in the limited case of POST)

[03:08:00.0000] <yecril71>
An embedded device�s resources can be exhausted with setTimeout

[03:08:01.0000] <yecril71>
just as they can be exhausted with time update.

[03:09:00.0000] <yecril71>
If the user encounters a rogue site that does bad things to the device,

[03:09:01.0000] <yecril71>
he can just blacklist it so that it cannot do so any more.

[03:12:00.0000] <yecril71>
Oops, sorry, I was just repeating Adrian (unintentionally).

[03:17:00.0000] <Philip`>
URIs were a good idea because you could refer directly to concrete things - instead of saying "download this pretty document via FTP on example.org in /pub/stuff.pdf" you could say "download ftp://example.org/pub/stuff.pdf". If you have to start saying "download http://example.org/pub/stuff using a PDF reader or with Accept:application/pdf" then that seems like a step in the wrong direction

[03:21:00.0000] <hsivonen>
Hixie: do you remember if Gecko has the kind of cache-related document.write behavior that we want to avoid with HTML5?

[03:38:00.0000] <hallvors>
URIs are a reasonable compromise between computers and humans

[03:55:00.0000] <yecril71>
Resources can have various representations, and the user is free to choose whatever representation suits him best.

[03:56:00.0000] <yecril71>
So I would rather say "Download this pretty document via HTML; 

[03:56:01.0000] <yecril71>
oops, HTTP;

[03:57:00.0000] <yecril71>
the server will deliver your preferred format."

[03:57:01.0000] <hsivonen>
yecril71: with conneg, the user doesn't know what's available

[03:58:00.0000] <yecril71>
Well, if the preferred format is unavailable, the server can return the next preferred one.

[03:58:01.0000] <yecril71>
Why is the information of what is available needed?

[03:59:00.0000] <hsivonen>
yecril71: the automation is not needed

[03:59:01.0000] <hsivonen>
and the particular kind of automation suggested sucks

[04:03:00.0000] <hsivonen>
Somehow I'm not surprised: http://twitter.com/waka/statuses/1011080251

[04:12:00.0000] <Lachy>
oh crap. Dmitry Turin has emailed me off list now about HTML6 :-(

[04:12:01.0000] <Lachy>
/me will not respond

[04:19:00.0000] <hallvors>
/me responds, realising part of the reason he reads public-html is that weird people are interesting :-p

[04:19:01.0000] <hallvors>
(I just said "noone will reply to that question because it's impossible to estimate and commercially sensitive information")

[04:20:00.0000] <Lachy>
hallvors, did Dmitry email you too?

[04:21:00.0000] <Lachy>
I would have told him that no-one is interested in implementing HTML6 at all, but he's on my do-not-respond list

[04:21:01.0000] <hsivonen>
/me guesses he emailed all Opera and Mozilla reps to the HTML WG

[04:22:00.0000] <hallvors>
no, just saw him on public-html. I'll see if that bait makes him too much of a nuisance. :-p

[04:22:01.0000] <Philip`>
yecril71: The hypothetical document is only pretty in PDF; the HTML one has ugly fonts and bad layout, which is why I was telling the hypothetical person to download the PDF

[04:23:00.0000] <hallvors>
yecril71: what type of content I think I'm getting is important to me. If I can choose between a .html and a .pdf link, I am in control. Content negotiation and extension-free links actually take that control away from the user.

[04:26:00.0000] <Philip`>
/me prefers to think about "things" rather than about "resources" and "representations", and considers a PDF document to be a thing, and wants to have a simple uniform way to refer to such things, because then the world is nice and simple and makes sense

[04:28:00.0000] <hsivonen>
/me sees http://html60.euro.ru/site/html60/

[04:29:00.0000] <yecril71>
You can choose between .html and .pdf.

[04:29:01.0000] <yecril71>
You achieve that goal by setting your preferences.

[04:29:02.0000] <yecril71>
If the document in HTML is unusable, it should not be there in the first place.

[04:50:00.0000] <hallvors>
yecril71: I do NOT want to reconfigure my browser each time I need a specific variant of a resource. That is a rather silly suggestion. I happen to know how to do it (which most users don't) and even then it's a bad idea. 

[04:51:00.0000] <mookid>
hey hey! someone else understands HTTP!

[04:51:01.0000] <mookid>
rejoice!

[04:52:00.0000] <mookid>
Philip`: read what I just posted to the mailing list - that approach breaks caching

[04:53:00.0000] <mookid>
if you have a seperate URL for each representation and you update one of those URLs

[04:53:01.0000] <mookid>
then caching will only account for the update to the one specific representation

[04:53:02.0000] <mookid>
the others will be considered unchanged

[04:53:03.0000] <BenMillard>
mookid, what do you do for the websites you make?

[04:54:00.0000] <mookid>
I'm implementing both URI and protocol level conneg because I have to :)

[04:54:01.0000] <mookid>
If HTML was good enough I'd just use HTTP conneg

[04:54:02.0000] <Philip`>
mookid: Even if all representations shared the same URL, then when I update the resource it'll often also be updating totally separate resources that provide an index of all resources and indicate their modification times, and resources that list recent changes to the site, and all sorts of other things

[04:55:00.0000] <BenMillard>
mookid, what did you do on previous website(s)?

[04:55:01.0000] <mookid>
if they're seperate reasources why are they at the same URL ?

[04:55:02.0000] <mookid>
BenMillard: that's not an argument you're just pointing out that HTML is insufficient

[04:55:03.0000] <mookid>
I agree with you on that.

[04:56:00.0000] <BenMillard>
mookid, I'm not making an argument or commenting on HTML's cababilities. :/

[04:56:01.0000] <mookid>
If you read the original post on the mailing list I even pointed out that it would be best to have both options available.

[04:56:02.0000] <Philip`>
mookid: They aren't - there's e.g. http://example.com/exciting-document (which accepts GET and PUT as text/html or application/pdf), but also http://example.com/recent-changes-to-site (which is application/atom+xml or whatever it's called)

[04:57:00.0000] <Philip`>
mookid: so they're separate resources, at separate URLs, but updating one will cause changes to the other

[04:57:01.0000] <mookid>
right..

[04:57:02.0000] <Philip`>
mookid: so any caching solution has to take into account the fact that updates to one URL might invalidate caches of another URL

[04:57:03.0000] <mookid>
why?

[04:57:04.0000] <mookid>
your designign your applications badly then.

[04:57:05.0000] <Philip`>
mookid: How else would it be designed?

[04:58:00.0000] <mookid>
well why would recent changes to site make a difference to exciting document?

[04:58:01.0000] <mookid>
that's a nonsense example

[04:58:02.0000] <Philip`>
mookid: When you change the exciting-document, that's a recent change to the site, so it'll cause the list of recent changes to change

[04:58:03.0000] <mookid>
BenMillard: I assumed you were going to accuse me of being ahypocrite because I have used URI's for conneg

[04:59:00.0000] <mookid>
I have to.

[04:59:01.0000] <mookid>
what value is there in caching that resource?

[04:59:02.0000] <mookid>
if it's constantly changing

[04:59:03.0000] <mookid>
that's a stupid example

[05:00:00.0000] <Philip`>
mookid: It's not constantly changing; it's only changing when something else (at a different URL) changes, which might be quite rare

[05:00:01.0000] <mookid>
well ten you would change the caching settings to reflect that

[05:00:02.0000] <mookid>
...

[05:01:00.0000] <mookid>
it's not wise to cache documents that are beign altered on teh backend

[05:01:01.0000] <Philip`>
If the caching system is able to cope with updates to exciting-document invalidating the cache of recent-changes-to-site, then exactly the same system could cope with updates to report.pdf invalidating caches of report.html, so there wouldn't be the caching problem you mentioned

[05:02:00.0000] <mookid>
wow.. what a total and complete waste of developer time

[05:02:01.0000] <mookid>
my way is alot easier..

[05:02:02.0000] <mookid>
caching systems support it out of the box..

[05:03:00.0000] <mookid>
hey look - someone updated it - update the cache

[05:03:01.0000] <mookid>
they really are clever folks these HTTP guys

[05:03:02.0000] <hsivonen>
/me posits that according to REST lore, cacheability is regarded to be of higher importance than conneg

[05:04:00.0000] <mookid>
how many times do you hit a cache of some kind

[05:04:01.0000] <mookid>
do you know?

[05:04:02.0000] <mookid>
it's pretty often

[05:05:00.0000] <mookid>
and conneg works in conjunction with caching anyway.. if you use the HTTP conneg

[05:06:00.0000] <mookid>
that's the point I'm making..

[05:06:01.0000] <mookid>
!

[05:06:02.0000] <yecril71>
Each time you need a specific variant of the resource, you say

[05:07:00.0000] <BenMillard>
mookid, I was trying to figure out the conditions which made you feel existing features were insufficient for what you are doing.

[05:07:01.0000] <Philip`>
/me notes that his university has switched off its university-wide web cache because it wasn't worth continuing

[05:07:02.0000] <yecril71>
File.OpenIn AcrobatReader (for example).

[05:07:03.0000] <BenMillard>
which first requires me understand what you were doing before

[05:07:04.0000] <hsivonen>
mookid: my point is that if you need an additional datum to travel with the URI and you need the additional datum to derefence, the better design is to bake that datum in the URI itself

[05:07:05.0000] <mookid>
BenMillard: I've done a pretty good job describing it on the mailing list - I jeep having to repeat myself because people don't read my posts :(

[05:07:06.0000] <yecril71>
And the browser asks your favourite application

[05:07:07.0000] <yecril71>
to download the same resource in its favourite form.

[05:08:00.0000] <mookid>
hsivonen: that is complete nonsense..!

[05:08:01.0000] <hsivonen>
mookid: why?

[05:08:02.0000] <mookid>
additional datum?

[05:08:03.0000] <mookid>
the header is THERE in every request

[05:08:04.0000] <mookid>
HTML just provides no way of telling browsers how to do something intresting with it

[05:09:00.0000] <mookid>
the URI is for RESOURCE indentification.. not resource + representation identification

[05:09:01.0000] <hsivonen>
mookid: it's not there when you put the URI in a href. It's not there when you email an URI to someone. It's not there on IRC. It's not there in PDF, OOXML, ODF, etc.

[05:09:02.0000] <Philip`>
URIs are useful as a uniform way of referencing stuff, but if you now have to say "open http://{some URI} in Acrobat" then it's no longer a uniform way of referencing stuff

[05:09:03.0000] <mookid>
well shouldn't it be up to them how they decide to view the resource?

[05:10:00.0000] <mookid>
Philip`:  YES IT IS

[05:10:01.0000] <mookid>
that's a GET on that URI

[05:10:02.0000] <mookid>
that's completely uniform

[05:10:03.0000] <mookid>
what are you talking abuot?

[05:10:04.0000] <mookid>
seriously. :/

[05:10:05.0000] <hsivonen>
mookid: note that I don't accept as dogma that URIs should be used to identify abstract resources as opposed to be used for application-level network addressing

[05:10:06.0000] <yecril71>
Using File.OpenIn is uniform, it is even more uniform than using extensions in URLs.

[05:11:00.0000] <mookid>
hsivonen: that's because you incorrectly percieve HTTP as a transport protool

[05:11:01.0000] <mookid>
protocol

[05:11:02.0000] <mookid>
it's a TRANSFER ptorocol

[05:11:03.0000] <yecril71>
protocol

[05:11:04.0000] <mookid>
thanks.. :P

[05:12:00.0000] <mookid>
URI's are protocol level anyway

[05:12:01.0000] <hsivonen>
mookid: I said "application-layer" precisely to suggest the layer above the transport layer

[05:12:02.0000] <mookid>
and so..

[05:12:03.0000] <mookid>
ignoreing that..

[05:13:00.0000] <yecril71>
Content needs to be referenced, its particular representation does not.

[05:13:01.0000] <hsivonen>
Is the ISO C spec online with a paywall?

[05:13:02.0000] <yecril71>
Because it is content that matters.

[05:13:03.0000] <yecril71>
paywhat?

[05:14:00.0000] <mookid>
yecril71: you on WHATWG?

[05:14:01.0000] <hsivonen>
yecril71: a paywall is a thing where you have to pay in order to view documents behind the wall

[05:14:02.0000] <Philip`>
hsivonen: http://www.open-std.org/JTC1/SC22/WG14/www/docs/n1256.pdf is a recent (most recent?) draft, and drafts are free

[05:14:03.0000] <hsivonen>
Philip`: thanks

[05:14:04.0000] <BenMillard>
mookid, can you link me to the message which describes the use-case where users of your website the current features lacking? I'm unable to find it...

[05:15:00.0000] <BenMillard>
s/website the current/website found the current/

[05:15:01.0000] <mookid>
BenMillard: it's not a case of 'extra' ends.. it's better means.. 

[05:15:02.0000] <Philip`>
"Septermber 7, 2007" (sic) - hmm

[05:15:03.0000] <yecril71>
I review the current progress at WHATWG.

[05:15:04.0000] <mookid>
well never mind HTTP lets concentrate on video tags!

[05:15:05.0000] <mookid>
serious business

[05:16:00.0000] <yecril71>
Or rather follow, since nobody is interested in my review :-)

[05:17:00.0000] <yecril71>
The ISO standards are on the ISO Web site.

[05:17:01.0000] <hallvors>
"content needs to be referenced, its particular representation does not" is a statement that I disagree with based on personal experience. I *often* want to reference a particular representation where multiple exist.

[05:17:02.0000] <mookid>
BenMillard: I just think HTML should do more to make HTTP more accessible

[05:17:03.0000] <mookid>
and you CANT

[05:17:04.0000] <mookid>
because

[05:17:05.0000] <mookid>
...

[05:17:06.0000] <mookid>
HTML doesnt provide the mechanism for it

[05:17:07.0000] <mookid>
hello?

[05:17:08.0000] <mookid>
!

[05:17:09.0000] <BenMillard>
sure it does, you just include the file extension at the end of the href value :)

[05:18:00.0000] <mookid>
that's using a URL for something it was never initially intended to do

[05:19:00.0000] <Philip`>
The whole web is being used for things it was never initially intended to do

[05:19:01.0000] <mookid>
well..

[05:19:02.0000] <Philip`>
and that's good, because it's evolved to do things that people find useful

[05:19:03.0000] <mookid>
don't you think a new version of HTML is a good opporutity to fix that?

[05:19:04.0000] <mookid>
HTML has nothing to do with that

[05:19:05.0000] <mookid>
it's mor eof a hindrance than a help

[05:20:00.0000] <mookid>
look at the prevelance of javascript

[05:20:01.0000] <mookid>
moslty down to the failures of html/browsers

[05:20:02.0000] <mookid>
ajax is slightly difference

[05:20:03.0000] <mookid>
different

[05:20:04.0000] <yecril71>
Nothing precludes the particular representations from having their separate URIs.

[05:20:05.0000] <mookid>
and there's that^

[05:21:00.0000] <mookid>
support HTTP - whcih is both methods

[05:21:01.0000] <yecril71>
However, the only use case for that I can think of is internal editorial review.

[05:21:02.0000] <mookid>
don't take that descision away from develoeprs

[05:21:03.0000] <mookid>
that's arrogant and it's wrong

[05:22:00.0000] <yecril71>
I still think the choice of format is on the viewer.

[05:22:01.0000] <Philip`>
The whole point of standards is to take decisions away from developers

[05:22:02.0000] <mookid>
er

[05:22:03.0000] <mookid>
the whole point?

[05:23:00.0000] <Philip`>
I might be exaggerating :-p

[05:23:01.0000] <yecril71>
It also helps the authors to produce reliable content.

[05:23:02.0000] <yecril71>
And it can help end users in obtaining technical support from their providers.

[05:24:00.0000] <mookid>
I love how hyper text's markup language has complete contempt for it's transfer protocol

[05:24:01.0000] <Philip`>
The whole point of standards is to take decisions away from developers and authors

[05:24:02.0000] <mookid>
the attitude of "well yeah they put that in but we dont need it"

[05:24:03.0000] <mookid>
what exactly did the HTTP guys produce an rfc for if it wasnt to tell you guys how to use their protocol properly

[05:25:00.0000] <mookid>
?

[05:25:01.0000] <Philip`>
(as evidenced by standards being written as a set of restrictions and requirements on developers and authors)

[05:25:02.0000] <mookid>
like an rfc?

[05:25:03.0000] <yecril71>
HTTP and HTML should be kept separate.

[05:26:00.0000] <yecril71>
They meet at the implementor�s desk, and both are binding.

[05:26:01.0000] <yecril71>
HTML is for HTTP, but not vice-versa, since HTTP can trasfer other resources equally well.

[05:27:00.0000] <yecril71>
What did I say?  It should be HTTP is for HTML.

[05:28:00.0000] <yecril71>
Since HTTP is not limited to HTML, there is no need for HTML to mirror HTTP in full.

[05:29:00.0000] <yecril71>
I hope this makes sense.

[05:35:00.0000] <Dashiva>
Is Smylers in here?

[05:37:00.0000] <Philip`>
/me frowns

[05:39:00.0000] <mookid>
I'm being asked to explain the advantages.. I'm giving them: It makes use of HTTP conneg and appropriate use of URIs..

[05:39:01.0000] <mookid>
when I do that

[05:39:02.0000] <mookid>
I get the response

[05:39:03.0000] <mookid>
"real world examples please"

[05:40:00.0000] <mookid>
wel.. there are very few real world examples ebcause the browser support isnt there

[05:40:01.0000] <mookid>
that doesnt exactly encourage use does it

[05:40:02.0000] <mookid>
it encourages the complete opposite

[05:40:03.0000] <mookid>
(conneg in URIs)

[05:41:00.0000] <mookid>
it's like dealing with politicians or something.. constantly arguing circles so that it ends up being too late to do anything about it

[05:43:00.0000] <mookid>
the issue here is that the benefits are hard to understand because you have to ignore your preconceptions about what a URI is and how conneg should work

[05:43:01.0000] <mookid>
it's viewed as 'difficult' because its a change to what people are used to thinking

[05:44:00.0000] <mookid>
there are obvious benefits of using protocol level conneg with HTTP.. if there weren't - they wouldnt have spent the time including it in the HTTP spec

[05:45:00.0000] <mookid>
the fact that it is not used is a criticism of HTML and browsers

[05:45:01.0000] <Philip`>
It doesn't have to be examples that already exist in the real world and use the feature that doesn't exist - it's more about concrete examples where a realistic developer would encounter a problem and the best way to solve the problem would be the proposed solution, rather than abstract ideas like "supporting HTTP is inherently good"

[05:45:02.0000] <mookid>
what like the examples I gave

[05:45:03.0000] <mookid>
of telling people theres a report at a given URI

[05:45:04.0000] <mookid>
and they can use several different UA's ?

[05:46:00.0000] <hallvors>
..my counter-claim is that those URLs are more usable with extensions..

[05:46:01.0000] <mookid>
more usable?

[05:46:02.0000] <Philip`>
mookid: Yes, like that

[05:47:00.0000] <mookid>
read the latest email I wrote - it's completely confusing

[05:47:01.0000] <mookid>
if I update report.pdf.. does that update report.html ?

[05:47:02.0000] <Philip`>
mookid: (but in that example, people aren't convinced that it's easier than providing a different URI for each UA, so they want more convincing examples)

[05:47:03.0000] <ehird>
hallvors: i disagree

[05:48:00.0000] <mookid>
right.. they're two seperate methods

[05:48:01.0000] <mookid>
that both 'work'

[05:48:02.0000] <mookid>
having content negotiation in the URI has massive potential for misleading users, actually

[05:51:00.0000] <mookid>
at th end of the day.. the benefits are that you support HTTP transactions in browsers better - yuo provide more options for developers to use URIs appropriately.. whilst at the same time protecting backwards compatability

[05:52:00.0000] <mookid>
if the Accept attribute is optional, it makes no difference to developers who insist on using URIs for conneg

[05:52:01.0000] <mookid>
and gives the ability to those who would like to.. do I have to do market research or can we all be grown up about it?

[05:53:00.0000] <Philip`>
mookid: But one method (separate URI per target UA) already works, and the other would require dozens of browser developers to spend time writing and testing and security-analysing and documenting the feature and would result in some users being unable to use some sites because they haven't upgraded their UA yet, so it has a non-negligible cost

[05:54:00.0000] <mookid>
security analysing?

[05:54:01.0000] <mookid>
at the moment the browser sends Accept */*

[05:55:00.0000] <Philip`>
Someone might write <a href="whatever" accept="&#x0a;&#x0a;oh no it's an unexpected request body"> and that'd be kind of bad, so you'd have to make sure nobody can do that

[05:55:01.0000] <mookid>
Actualyl there are clever things you can do at the server side to look at the client - if that was a problem

[05:55:02.0000] <mookid>
so you wouldn't necessarily have to update the UA

[05:56:00.0000] <mookid>
why would that be bad?

[05:56:01.0000] <mookid>
explain to me why that's bad

[05:57:00.0000] <mookid>
it's stupid.. but I can't see how "allowing" that is inherently a bad thign

[05:57:01.0000] <Philip`>
It would be bad for the same reasons that XMLHttpRequest doesn't let you set arbitrary HTTP headers, and blacklists things like "Host" and "Referer"

[05:57:02.0000] <mookid>
what reason is that?

[05:57:03.0000] <hsivonen>
/me agrees with hallvors about wanting to address particular representations

[05:58:00.0000] <mookid>
that's because that's what you're used to

[05:58:01.0000] <mookid>
it's just fear of change

[05:58:02.0000] <hallvors>
mookid: look up HTTP request splitting

[05:58:03.0000] <mookid>
why?

[05:59:00.0000] <hallvors>
it's a cache-poisoning technique that causes potential security problems

[05:59:01.0000] <mookid>
and..?

[05:59:02.0000] <Philip`>
mookid: Some sites trust Referer to check that users came from the right place, to prevent dangerous cross-site requests, and that works now because Referer is only set by UAs and an attacker can't set it to an arbitrary value

[05:59:03.0000] <hallvors>
that's why adding methods to set http headers must be done very carefully

[05:59:04.0000] <mookid>
hahahahaha

[06:00:00.0000] <mookid>
that's awesome

[06:00:01.0000] <mookid>
does security by obscurity mean anything to you?

[06:00:02.0000] <hsivonen>
mookid: in general, "if there weren't - they wouldn't have spent the time of including it in the HTTP spec" is a bogus argument

[06:00:03.0000] <mookid>
no it's not.. they designed the protocol they would knwo the most appropriate way of using it

[06:00:04.0000] <mookid>
en dof

[06:01:00.0000] <mookid>
the fact that HTML has FAILED miserably to provide adequate mechanisms for making full use of their protocol

[06:01:01.0000] <mookid>
and so it is common place for peolpe to hack conneg into URIs

[06:01:02.0000] <Philip`>
mookid: Sure, but the Referer thing can actually achieve its goal of making it impossible for a user using a standard web browser to be tricked into making cross-site requests

[06:02:00.0000] <mookid>
what has that got to do with anything?

[06:02:01.0000] <hallvors>
if they didn't consider "URL usability" when authoring the specs they may have drawn the wrong conclusions..

[06:02:02.0000] <mookid>
well they did..

[06:02:03.0000] <mookid>
URLs (funnily enough) are used to locate resoruces

[06:02:04.0000] <Philip`>
It's a bit of a rubbish way of doing it, but it works in practice, so people rely on it, so browsers can't now start violating the assumptions people made, so they have to be very careful about how a web page can control the HTTP requests they send

[06:02:05.0000] <mookid>
resources.. and nott representations

[06:03:00.0000] <hsivonen>
mookid: in general, it's bogus to assume that the WG that defined HTTP 1.1 (or HTML 4) was axiomatically enlightened but that the WHATWG isn't equally axiomatically enlightened

[06:03:01.0000] <hallvors>
well, I disagree that the result is usable :) because I want to identify representations..

[06:03:02.0000] <mookid>
well from what I've read so far I'm onyl seeing circular disstractive arguments

[06:03:03.0000] <mookid>
so I'd say that wouldn't be a bad guess

[06:03:04.0000] <hsivonen>
to be clear, I'm not suggesting that the WHATWG is axiomatically enlightened. I'm suggesting that the HTTP 1.1 spec can contain bad stuff.

[06:04:00.0000] <Philip`>
Clearly the WHATWG should write HTTP5 which defines Uniform Representation Identifiers, and then we'd be okay

[06:04:01.0000] <hsivonen>
Philip`: nope, Locators!

[06:04:02.0000] <mookid>
you havent even given HTTP conneg a decent shot

[06:04:03.0000] <mookid>
so that is ridiculous logic

[06:05:00.0000] <Philip`>
hsivonen: Whatever, they're basically the same thing :-p

[06:05:01.0000] <mookid>
Uniform Resource and/or Representation Indentifiers

[06:05:02.0000] <mookid>
no.. no they arent.

[06:05:03.0000] <hsivonen>
Hixie: we should call the things that are not quite URIs in the HTML5 spec "Uniform Representation Locators"

[06:06:00.0000] <Dashiva>
Can't we endlessly debate what makes a Resource instead?

[06:06:01.0000] <hsivonen>
http://lists.w3.org/Archives/Public/www-tag/2008Mar/0026.html

[06:06:02.0000] <ehird>
well, a representation has to be a representation of SOMETHING

[06:06:03.0000] <ehird>
and resource is the generic term for something that can be represented, I guess

[06:07:00.0000] <Philip`>
hsivonen: Why does the acronym need to have an expansion? Just define it as an atomic term whose meaning is defined by some spec and not by a three-word description

[06:08:00.0000] <jcranmer>
like ISO

[06:08:01.0000] <mookid>
ehird: exactly.

[06:08:02.0000] <ehird>
hmm

[06:08:03.0000] <ehird>
you can represent a representation

[06:08:04.0000] <ehird>
os representations of resources are resources

[06:08:05.0000] <ehird>
xD

[06:08:06.0000] <ehird>
bizarre, but consistent

[06:09:00.0000] <mookid>
sure, there's nothing to say you cant do that in the http spec either

[06:09:01.0000] <ehird>
everything is a resource, a representation is of a resource.

[06:09:02.0000] <ehird>
representations are just a subset of resources

[06:09:03.0000] <mookid>
hmm

[06:09:04.0000] <mookid>
not really

[06:09:05.0000] <jcranmer>
it doesn't matter what the spec says

[06:09:06.0000] <Philip`>
If I can represent *any* representation, and representations represent resources, then all representations are resources, so we might as well call them all the same thing :-)

[06:10:00.0000] <jcranmer>
it matters what who implements the spec do

[06:10:01.0000] <ehird>
mookid: why not?

[06:10:02.0000] <ehird>
Philip`: well, yes, but

[06:10:03.0000] <ehird>
there are resources that are not representations

[06:10:04.0000] <hsivonen>
at the end of the day, representations are what cross the wire, so resources don't matter as much

[06:10:05.0000] <mookid>
jcranmer: well no it doesnt from my perspective - I know what the best practice is and why it would work - I've spent a while studying it

[06:10:06.0000] <ehird>
hsivonen: exactly

[06:10:07.0000] <ehird>
you can't send me, a person over the wire

[06:10:08.0000] <mookid>
that's how you see it at the moment..

[06:10:09.0000] <ehird>
but you can send a representation of me over the wire

[06:10:10.0000] <ehird>
and yet that representation is also a resource

[06:10:11.0000] <mookid>
you're all looking at URIs in tehw ay they're used at the moment

[06:10:12.0000] <ehird>
just one that happens to not be in meatspace

[06:11:00.0000] <mookid>
can we just step back from that a second

[06:11:01.0000] <jcranmer>
mookid: example

[06:11:02.0000] <ehird>
mookid: sure, if you give us some reasons why

[06:11:03.0000] <jcranmer>
HTML is SGML, by definition

[06:11:04.0000] <jcranmer>
ever try using the advanced SGML features in HTML?

[06:11:05.0000] <mookid>
why do you ask?

[06:11:06.0000] <jcranmer>
in practice, then, HTML uses a subset of SGML

[06:11:07.0000] <Lachy>
jcranmer, HTML5 is not SGML by definition

[06:12:00.0000] <jcranmer>
Lachy: lemme clarify... HTML 4.01

[06:12:01.0000] <jcranmer>
(hence why HTML5 is not SGML AIUI)

[06:12:02.0000] <mookid>
I still dont understand why it's so bad to include an optional attribute that will significantly help RESTful developers

[06:12:03.0000] <mookid>
is it laziness?

[06:12:04.0000] <mookid>
what?

[06:12:05.0000] <ehird>
mookid: what exactly is it

[06:12:06.0000] <ehird>
i'm uninformed

[06:13:00.0000] <Philip`>
mookid: It's cost vs benefit, both of which are non-zero

[06:13:01.0000] <mookid>
is it really that costly to implement oen attribute?

[06:13:02.0000] <mookid>
I think you might have problems with yoru processes.

[06:13:03.0000] <mookid>
-_-

[06:13:04.0000] <jcranmer>
mookid: you don't understand, I think

[06:13:05.0000] <ehird>
mookid: if every semi-useful attrib was added

[06:13:06.0000] <hsivonen>
mookid: it makes the address a pair instead of a single value, which sucks, because existing systems deal with one value

[06:13:07.0000] <ehird>
we'd have a monstrosity

[06:13:08.0000] <Dashiva>
Opportunity cost, maintenance costs, attack surface...

[06:13:09.0000] <ehird>
now someone tell me what it actually is ;)

[06:14:00.0000] <mookid>
hsivonen: it doesn't do anything of the sort

[06:14:01.0000] <jcranmer>
the mere virtue of adding something is a csot

[06:14:02.0000] <mookid>
HTTP already is a pri.. it's headers and body

[06:14:03.0000] <mookid>
pair^

[06:14:04.0000] <ehird>
whats that got tot do with uris.

[06:14:05.0000] <jcranmer>
then there's the costs of the actual implementation of the attribute, and what it does

[06:14:06.0000] <mookid>
everything

[06:15:00.0000] <ehird>
i see.

[06:15:01.0000] <mookid>
well the attribute isn't exactly complicated is it

[06:15:02.0000] <mookid>
and there's no greater attack surface

[06:15:03.0000] <mookid>
it's restricting the header which is allready catcha ll

[06:15:04.0000] <mookid>
so..

[06:15:05.0000] <mookid>
chill out on that one Neo

[06:15:06.0000] <ehird>
someone wanna tell me what it is yet? :q

[06:15:07.0000] <ehird>
or at least a link

[06:15:08.0000] <mookid>
a URI?

[06:16:00.0000] <jcranmer>
mookid: no HTTP uri mechanism, that I know of, allows you to specify headers

[06:16:01.0000] <Philip`>
ehird: What the proposed feature we're discussing is? I think it's <a href="report" accept="application/pdf">

[06:16:02.0000] <ehird>
Philip`: ewwwwwwwww

[06:16:03.0000] <mookid>
HTTP uri mechanism?!

[06:16:04.0000] <ehird>
how is that represented in the browser's URI box?

[06:16:05.0000] <mookid>
wth is that?

[06:16:06.0000] <ehird>
if you copy the URI there

[06:16:07.0000] <ehird>
and paste it in a new window

[06:16:08.0000] <ehird>
might you get a different representation?

[06:16:09.0000] <ehird>
if so... awful, awful, just plain awful

[06:16:10.0000] <jcranmer>
/me heads off to class

[06:16:11.0000] <hsivonen>
ehird: details! :-)

[06:17:00.0000] <mookid>
if you put a URI into a browser 'uri box' then it will fetch HTML.. that's the desired behaviour in that use case

[06:17:01.0000] <Philip`>
ehird: Yes, but it'll be a representation appropriate to the UA you paste it into, and it'll be the same resource so you'll never mind the difference

[06:17:02.0000] <ehird>
Philip`: Except that the page evidently thought that it was important I get THIS representation

[06:17:03.0000] <ehird>
So... not irrelevant, evidently

[06:17:04.0000] <mookid>
why is it awful to get the most appropriate representation of the exact same data?

[06:18:00.0000] <Philip`>
Depends on what you mean by "most appropriate" - when someone tells me to look at something specific, the most appropriate thing for me to see is the same thing they're seeing, and not a UA-dependent variant in a different format

[06:18:01.0000] <mookid>
oh damn... my user agent got the information in the best format for it.. shock horror

[06:19:00.0000] <mookid>
well then use the same UA..

[06:19:01.0000] <mookid>
einstein..

[06:19:02.0000] <ehird>
mookid: snide jabs help nobody

[06:19:03.0000] <mookid>
well cmon.. that's just ridiculous

[06:19:04.0000] <Philip`>
Then they'd have to tell me what UA they were using, so they're now using a (uri, ua) pair (for which no nice uniform syntax is defined) as the reference

[06:20:00.0000] <mookid>
what?

[06:20:01.0000] <mookid>
you would just send them the URI and say 'i'm looking at the pdf'

[06:20:02.0000] <mookid>
why does it matter that much what the format of the data is?

[06:20:03.0000] <ehird>
mookid: I thought you were advocating ADDING the conneg attribute?

[06:20:04.0000] <ehird>
Bit of a turn around there .. ?

[06:21:00.0000] <mookid>
I am - conneg in this isntance is seperated from the URI

[06:21:01.0000] <mookid>
so, normally your UA woud have fixed defaults

[06:21:02.0000] <mookid>
like a PDF document

[06:21:03.0000] <mookid>
reader

[06:21:04.0000] <mookid>
or iTunes (mp3)

[06:21:05.0000] <mookid>
btu browsers Get lots of different stuff and pass it to the OS

[06:21:06.0000] <mookid>
so you need a way of making the browser Accept header dynamic

[06:21:07.0000] <ehird>
/me shrugs. My thoughts: Overly complex, not a nice user experience when copying&pasting the resulting URI, not worth the effort required to implement even for the gains it might bring.

[06:22:00.0000] <mookid>
well it's a better user experience because they can use the UA that suits them

[06:22:01.0000] <Philip`>
What about an iTunes-like music player that had an embedded web browser?

[06:22:02.0000] <ehird>
Philip`: cough - songbird - cough?

[06:22:03.0000] <ehird>
;)

[06:22:04.0000] <mookid>
well it depends where in teh GUI you put the URI..

[06:22:05.0000] <mookid>
doesnt it?

[06:23:00.0000] <mookid>
you keep saying stuff that you clearly think is valid when it's not.... :/

[06:23:01.0000] <Philip`>
You put it in the text box at the top of the window, which is where you put all your URIs and search queries and other textual input

[06:24:00.0000] <mookid>
so the brwoser would open up the html page which would (heopfully) provide hyper text links to all the other representations.. if HTML5 works properly that will be the same URI but with a differet Accept headerconfiguration  indicated

[06:24:01.0000] <mookid>
YES you can do that with URIs

[06:24:02.0000] <mookid>
I know.

[06:25:00.0000] <mookid>
please don't tell me that AGAIN

[06:25:01.0000] <Philip`>
You can do that with UR... oh, okay

[06:25:02.0000] <ehird>
you can do that with UR*shot*

[06:25:03.0000] <ehird>
damnit Philip` 

[06:25:04.0000] <ehird>
:(

[06:25:05.0000] <Philip`>
/me is probably not being very helpful

[06:26:00.0000] <mookid>
essentially what you're saying is that in WHATWG's opinion.. the HTTP Accept header should be deprecated

[06:26:01.0000] <ehird>
what I'm saying is that you don't "get" Accept

[06:27:00.0000] <ehird>
accept is for the user/browser to dictate what formats they want.

[06:27:01.0000] <ehird>
not the page.

[06:27:02.0000] <mookid>
er

[06:27:03.0000] <ehird>
if the page must dictate the format, use file extensions

[06:27:04.0000] <ehird>
in the URI

[06:27:05.0000] <mookid>
I'm not even going to bother.. :/

[06:27:06.0000] <ehird>
then you won't convince me.

[06:27:07.0000] <ehird>
nice way to have an argument, but a bit peculiar

[06:27:08.0000] <mookid>
I KNOW!!! :D

[06:28:00.0000] <ehird>
sooo

[06:28:01.0000] <ehird>
why are you talking if you don't want to convince

[06:28:02.0000] <mookid>
well you're arguing in circles and you arent making sense

[06:28:03.0000] <mookid>
I'm saying

[06:28:04.0000] <mookid>
ok

[06:28:05.0000] <mookid>
bear with me

[06:28:06.0000] <hsivonen>
mookid: I suggest going back to what BenMillard said: giving a rationale in terms of use cases instead of appealing to the authority of the definers of HTTP.

[06:29:00.0000] <mookid>
developers need a way to construct HTML that can allow the user.. a way of modifying the Accept header in an HTML interface

[06:29:01.0000] <ehird>
no

[06:29:02.0000] <ehird>
i disagre

[06:29:03.0000] <ehird>
e

[06:29:04.0000] <mookid>
that isnt possible at the moment.. so URIs are used instead as a hack

[06:30:00.0000] <mookid>
it's a hack though

[06:30:01.0000] <ehird>
i disagree with your first point

[06:30:02.0000] <mookid>
it's just what people are used to

[06:30:03.0000] <ehird>
therefore you cannot convince me without convincing me of your first point.

[06:30:04.0000] <hsivonen>
mookid: It's a hack only from the point of view of a particular dogma that you haven't justified

[06:30:05.0000] <BenMillard>
ehird, file extensions in href (or src and so forth) seem like an solution to me, too. I consider this a Solved Problem, although if mookid follows hsivonen's advice maybe that'll unearth a widespread user need we haven't thought of.

[06:30:06.0000] <mookid>
dogma?

[06:30:07.0000] <hsivonen>
mookid: yes

[06:30:08.0000] <mookid>
lol

[06:31:00.0000] <ehird>
i hate it when people say "lol" when they run themselves into a corner

[06:31:01.0000] <ehird>
and don't justify their position...

[06:31:02.0000] <mookid>
what

[06:31:03.0000] <mookid>
it's not dogma

[06:31:04.0000] <mookid>
he's being a douchebag

[06:31:05.0000] <ehird>
gee, that's very kind

[06:31:06.0000] <mookid>
so is describing my poitn of view as dogma.

[06:32:00.0000] <ehird>
that's not a personal attack, that's a criticism of your views.

[06:32:01.0000] <mookid>
that's not a criticism

[06:32:02.0000] <jcranmer>
mookid: look at it like this

[06:32:03.0000] <jcranmer>
under your proposal

[06:32:04.0000] <hsivonen>
mookid: you can make it non-dogma by explaining in terms of use cases and addressing use cases and problems and solutions why putting ".pdf" in the URI is inappropriate

[06:32:05.0000] <jcranmer>
the accept is static

[06:32:06.0000] <mookid>
I DID EXPLAIN THE USE CASE

[06:33:00.0000] <jcranmer>
the page itself defines what it wants

[06:33:01.0000] <mookid>
"here theres somethign interesting here: example.com/report - have a look"

[06:33:02.0000] <mookid>
user can use whatever UA they want

[06:33:03.0000] <mookid>
that's a use case.

[06:33:04.0000] <jcranmer>
mookid: if the page says, "I want this to be HTML"

[06:33:05.0000] <jcranmer>
why not just tack on the .html at the end?

[06:33:06.0000] <hsivonen>
mookid: s without specifying accept, they get whatever

[06:33:07.0000] <ehird>
mookid: and?

[06:33:08.0000] <hsivonen>
s/s /so /

[06:34:00.0000] <mookid>
you can't do that using URIs for conneg

[06:34:01.0000] <mookid>
how would you say

[06:34:02.0000] <hsivonen>
mookid: if you want to send someone a link to a pdf, why is putting ".pdf" in the URI not an acceptable way to address the use case?

[06:34:03.0000] <ehird>
hsivonen ++

[06:34:04.0000] <ehird>
it's not a hack at all

[06:34:05.0000] <ehird>
if the type really is important, specify .pdf

[06:34:06.0000] <ehird>
BUT

[06:34:07.0000] <ehird>
Accept is for the user agent

[06:34:08.0000] <mookid>
"hey look at this URI and use whatever UA you want to look at it" if conneg is done in the URI ?

[06:34:09.0000] <ehird>
s

[06:34:10.0000] <ehird>
and the users

[06:34:11.0000] <ehird>
NOT the content authors

[06:34:12.0000] <mookid>
what?

[06:35:00.0000] <ehird>
the users, and their user agents, specify what types THEY personally want

[06:35:01.0000] <ehird>
it's not for the content authors to decide

[06:35:02.0000] <ehird>
for that, they should use .pdf and similar

[06:35:03.0000] <hsivonen>
mookid: should I try Firefox, iTunes, Word, Excel, OOo, etc. to see what you serve to each?

[06:35:04.0000] <mookid>
it's the SAME resource

[06:35:05.0000] <mookid>
it's the SAME data

[06:35:06.0000] <mookid>
it's just a DIFFERENT REPRESENTATION

[06:35:07.0000] <ehird>
I agree with that much.

[06:35:08.0000] <jcranmer>
at some point

[06:36:00.0000] <jcranmer>
you realize that idealism doesn't match up with reality

[06:36:01.0000] <mookid>
oh right

[06:36:02.0000] <ehird>
But I say that pragmatically, file extensions are the way to solve this

[06:36:03.0000] <ehird>
and the Accept changing damages repeatability

[06:36:04.0000] <mookid>
maybe that's becaue of belligerant spec writerS?

[06:36:05.0000] <ehird>
(copy&paste the uri and get the same thing)

[06:36:06.0000] <jcranmer>
no

[06:36:07.0000] <mookid>
possibly?

[06:36:08.0000] <hsivonen>
mookid: if it's same enough that the representation doesn't matter, why do you care about the accept value?

[06:36:09.0000] <mookid>
no?

[06:36:10.0000] <mookid>
thoguht not.

[06:36:11.0000] <ehird>
and is not good use of Accept

[06:36:12.0000] <ehird>
it's for users/user agents

[06:36:13.0000] <jcranmer>
mookid: no, implementors

[06:36:14.0000] <mookid>
omg..

[06:36:15.0000] <BenMillard>
/me realises what the most productive thing to do at this point is.

[06:37:00.0000] <jcranmer>
specs mean nothing if they're not implemented

[06:37:01.0000] <ehird>
mookid: If you weren't being condescending and insulting, I'd talk more.

[06:37:02.0000] <jcranmer>
ask the ACAP author

[06:37:03.0000] <mookid>
hsivonen: because the accept value is the only way you can serve multiple content types from one URI

[06:37:04.0000] <ehird>
And?

[06:37:05.0000] <mookid>
that's important..

[06:37:06.0000] <mookid>
that's how HTTP was designed

[06:37:07.0000] <ehird>
Is it now.

[06:37:08.0000] <mookid>
yes it is now.

[06:37:09.0000] <jcranmer>
mookid: are you are an HTTP developer?

[06:37:10.0000] <ehird>
Well, it'd be nice if you justified that.

[06:37:11.0000] <hsivonen>
mookid: but you shouldn't serve them from one URI unless you are indifferent towards which representation a user gets

[06:37:12.0000] <ehird>
exactly

[06:37:13.0000] <mookid>
no I'm an architect

[06:37:14.0000] <ehird>
if it matters, separate the URIS

[06:38:00.0000] <mookid>
I *actually use* this stuff

[06:38:01.0000] <jcranmer>
do you know the contents of the HTTP development meetings?

[06:38:02.0000] <danbri>
are the list archives borken? http://lists.whatwg.org/pipermail/help-whatwg.org/

[06:38:03.0000] <mookid>
to do new interesting stuff.

[06:38:04.0000] <mookid>
and progress this stuff forwards..

[06:38:05.0000] <mookid>
rather than continuing to do the same stuff that doesnt really work very well

[06:38:06.0000] <danbri>
ah, i'm mixing up lists

[06:38:07.0000] <ehird>
danbri: wfm

[06:38:08.0000] <jcranmer>
you seem to be claiming to channel the HTTP spec writers

[06:39:00.0000] <ehird>
spec development via spiritual energy

[06:39:01.0000] <mookid>
no I'm not.. I just look at the Accept haeder and thing.. hmm WHAT MIGHT BE THE REASON FOR THAT BEIGN THERE

[06:39:02.0000] <mookid>
OH I DONT KNOW

[06:39:03.0000] <mookid>
for fun.

[06:39:04.0000] <ehird>
mookid: for users

[06:39:05.0000] <ehird>
and

[06:39:06.0000] <ehird>
user agents

[06:39:07.0000] <mookid>
for suers?

[06:39:08.0000] <jcranmer>
for UAs

[06:39:09.0000] <ehird>
to specify what types they prefer to get data in.

[06:39:10.0000] <ehird>
NOT for content authors

[06:39:11.0000] <mookid>
YES

[06:39:12.0000] <mookid>
AND MOST UAs

[06:39:13.0000] <danbri>
(i read meaning into http://www.whatwg.org/mailing-list url and assumed was only one list)

[06:39:14.0000] <ehird>
USE ALL CAPS

[06:39:15.0000] <mookid>
have fixed Accept headers

[06:39:16.0000] <ehird>
that's their problem

[06:39:17.0000] <mookid>
brwosers

[06:39:18.0000] <ehird>
next

[06:39:19.0000] <mookid>
dont

[06:39:20.0000] <mookid>
they're Hypermedia brwosers

[06:40:00.0000] <mookid>
hypermedia provides a window for the OS

[06:40:01.0000] <ehird>
yawn, this is the part where I realise you're not actaully listening to us and stop talking to you

[06:40:02.0000] <jcranmer>
yep

[06:40:03.0000] <mookid>
well yuo're tlaking crap that's why

[06:40:04.0000] <mookid>
no offence

[06:40:05.0000] <jcranmer>
[ citation needed ]

[06:40:06.0000] <ehird>
mookid: no, i'm talking a representation of crap

[06:40:07.0000] <ehird>
obviously.

[06:40:08.0000] <mookid>
hillarious.

[06:41:00.0000] <ehird>
almost as hilarious as calling people who disagree with you douchebags!

[06:41:01.0000] <mookid>
you clearly don't knwo what yuo're tlaking abuot just pipe down

[06:41:02.0000] <ehird>
anyway, i thought i said i'd stop talking to you

[06:41:03.0000] <mookid>
good you do that

[06:41:04.0000] <jcranmer>
ehird: ever visit any Usenet groups recently?

[06:41:05.0000] <ehird>
jcranmer: No -- and there's a good reason for that

[06:41:06.0000] <ehird>
:)

[06:41:07.0000] <Philip`>
It's much easier to say you're going to stop talking to someone, than to actually stop talking to them

[06:42:00.0000] <ehird>
Philip`: yep

[06:42:01.0000] <hsivonen>
http://www.joelonsoftware.com/articles/fog0000000018.html

[06:42:02.0000] <mookid>
jcranmer: you stopped that train of thought I noticed

[06:42:03.0000] <mookid>
omg URI conneg!

[06:42:04.0000] <jcranmer>
ehird: shame... he reminds me of JSH or Twisted on sci.math and comp.lang.java.programmer in recent months

[06:42:05.0000] <jcranmer>
hmm, not so much JSH though

[06:42:06.0000] <ehird>
hsivonen: i was gonna link to that

[06:42:07.0000] <Philip`>
/me thinks joelonsoftware.com's URLs indicate great optimism as to how many articles will be published

[06:42:08.0000] <ehird>
when he said architet 

[06:43:00.0000] <ehird>
*architect

[06:43:01.0000] <ehird>
Philip`: lol :)

[06:43:02.0000] <ehird>
Philip`: interestingly, the URIs actually go downwards

[06:43:03.0000] <mookid>
"there's no such thing as superior beings" - actually there is that's how genetics work deal with it

[06:43:04.0000] <ehird>
http://www.joelonsoftware.com/Archive.html

[06:44:00.0000] <ehird>
and then up again

[06:44:01.0000] <ehird>
go figure

[06:44:02.0000] <mookid>
jcranmer: we didnt finisht aht off

[06:45:00.0000] <mookid>
so.. most UAs have specific Accept requirements that dont change

[06:45:01.0000] <mookid>
Browsers are a special case that do hypermedia but ALSO pass other content types to the OS

[06:45:02.0000] <mookid>
HTML needs to take account for that if people are to start using URIs properly

[06:45:03.0000] <ehird>
*crickets*

[06:45:04.0000] <mookid>
the reasont hey dont do that now

[06:46:00.0000] <mookid>
is becaue its impossible

[06:46:01.0000] <mookid>
not because it's "hard"

[06:46:02.0000] <ehird>
i considered something similar when i first starting using content negotiation

[06:46:03.0000] <mookid>
if you look at Rails.. Jax-RS..

[06:46:04.0000] <ehird>
except I realised why it was silly a few minutes after.

[06:46:05.0000] <mookid>
yeah silly..

[06:46:06.0000] <mookid>
that's why JAX-RS is built around it

[06:46:07.0000] <ehird>
yeah- silly

[06:46:08.0000] <mookid>
I suppose

[06:46:09.0000] <mookid>
do you know what JAX-RS is?

[06:47:00.0000] <ehird>
/me sighs

[06:47:01.0000] <ehird>
appeal to authorityyyyyyy

[06:47:02.0000] <mookid>
look it up then

[06:47:03.0000] <mookid>
I am not affiliated with them

[06:47:04.0000] <hsivonen>
Repetitive Strain XML binding for Java?

[06:47:05.0000] <mookid>
no it's got nothing to do with XML actually

[06:48:00.0000] <mookid>
it's content-type neutral

[06:48:01.0000] <mookid>
but the syntax is designed around HTTP methods.. and as a subset.. requested representation

[06:49:00.0000] <mookid>
I wonder why they did that..

[06:49:01.0000] <mookid>
a URI.. split up by methods..

[06:49:02.0000] <Philip`>
/me thinks it's fun how http://www.intertwingly.net/code/mombo/htaccess process the Accept header via a collection of UA-specific regexps, so it's basically an obscured form of UA sniffing

[06:49:03.0000] <mookid>
and content-type..

[06:49:04.0000] <ehird>
Philip`: cute

[06:49:05.0000] <hsivonen>
mookid: because WS-* went downhill, REST is in vogue and for each buzzword, Java has to hava a framework?

[06:50:00.0000] <mookid>
REST is just codeword for not using HTTP retardedly

[06:51:00.0000] <hsivonen>
mookid: right, so why do you need a framework beyond an HTTP framework?

[06:51:01.0000] <mookid>
RAD

[06:51:02.0000] <mookid>
duuuuh...

[06:51:03.0000] <mookid>
you can make a REST interface with folders and index.php scripts if you want

[06:51:04.0000] <mookid>
yuo've been able to do that for years

[06:51:05.0000] <ehird>
totally RAD

[06:51:06.0000] <mookid>
idiot.

[06:51:07.0000] <ehird>
hey, there go the personal insults again

[06:51:08.0000] <ehird>
i love this

[06:51:09.0000] <hsivonen>
/me subclasses HttpServlet, but whatever

[06:52:00.0000] <Philip`>
If I upload a few static pages and serve them via Apache, is that REST?

[06:52:01.0000] <mookid>
yes if you do it right

[06:52:02.0000] <mookid>
of course it is

[06:52:03.0000] <mookid>
REST isnt very complicated

[06:52:04.0000] <mookid>
people use big words to make it sound all clever

[06:53:00.0000] <mookid>
it's just "looking at HTTP and designing around it"

[06:53:01.0000] <mookid>
partically speaking at least

[06:53:02.0000] <mookid>
they get all beardy about it

[06:53:03.0000] <mookid>
because you can be RESTful with another protocol or whatveer

[06:53:04.0000] <ehird>
mookid: I believe we all know this

[06:54:00.0000] <mookid>
I dont

[06:54:01.0000] <hsivonen>
mookid: so why do you assume we are clueless but the HTTP folks and JAX-RS folks did everything for a good reason?

[06:54:02.0000] <mookid>
REsources + representations

[06:54:03.0000] <mookid>
caching

[06:54:04.0000] <mookid>
HTTP 

[06:55:00.0000] <ehird>
REST = REsources?

[06:55:01.0000] <mookid>
conneg

[06:55:02.0000] <jcranmer>
hsivonen: because we disagree with him

[06:55:03.0000] <ehird>
lol!

[06:55:04.0000] <ehird>
Representational State Transfer

[06:55:05.0000] <mookid>
jcranmer: you've stopped our conversation before

[06:55:06.0000] <mookid>
why?

[06:55:07.0000] <ehird>
mookid: because you're ignoring us, treating us as idiots, insulting us, being condescending, ...

[06:56:00.0000] <mookid>
because I explained what a browser is supposed to do

[06:56:01.0000] <jcranmer>
I'm also in middle of class

[06:56:02.0000] <mookid>
and the function of hypermedia

[06:56:03.0000] <ehird>
jcranmer: :-P

[06:56:04.0000] <mookid>
ooh right ok my bad enjoy

[06:56:05.0000] <jcranmer>
mookid: to be precise, you explaing what you THINK a browser is supposed to do

[06:56:06.0000] <mookid>
well you tell me the alternative definition then

[06:56:07.0000] <ehird>
... but not WHY

[06:57:00.0000] <mookid>
The benefits of protocol conneg are reasonably apparent I thin I've done a reasonable job putting them forwar

[06:57:01.0000] <mookid>
when I do that

[06:57:02.0000] <mookid>
I get

[06:57:03.0000] <mookid>
"real world examples please?"

[06:57:04.0000] <ehird>
well, you haven't

[06:57:05.0000] <ehird>
unfortunately

[06:57:06.0000] <mookid>
just be quiet you no one cares what you think.. :/

[06:58:00.0000] <jcranmer>
invert speaker and listener of last sentence

[06:58:01.0000] <Dashiva>
mookid: It seems you just dismiss anyone who doesn't agree with you

[06:58:02.0000] <ehird>
and this is why peopel don't talk to you

[06:58:03.0000] <mookid>
no I dont I just dismiss people who dont do me the respect of reading and comprehending what I'm communicating to them

[06:58:04.0000] <Philip`>
ehird: People here are doing a very bad job if they're intending to not talk to him

[06:59:00.0000] <ehird>
indeed...

[06:59:01.0000] <mookid>
indded?

[06:59:02.0000] <mookid>
just shush please

[06:59:03.0000] <jcranmer>
true

[06:59:04.0000] <ehird>
/me shushes please

[06:59:05.0000] <mookid>
yeah I completely respect people for having thei opinion I just dont feel ilke some people are really opening up

[07:00:00.0000] <mookid>
in how they consier what I'm saying

[07:00:01.0000] <Dashiva>
Have you ever considered that you might be wrong?

[07:00:02.0000] <mookid>
because most of the respnses I get are already dealt with in other stuf I've said

[07:00:03.0000] <ehird>
Dashiva: oh, you joker!

[07:00:04.0000] <mookid>
well I've been looking at this for a while now

[07:00:05.0000] <mookid>
so..

[07:00:06.0000] <mookid>
I havent heard antyhign from anyone

[07:00:07.0000] <mookid>
and actually

[07:00:08.0000] <mookid>
most of the peolpe who sound like they know what they're talkign abuot

[07:01:00.0000] <mookid>
arent suggesting i'm wrong

[07:01:01.0000] <jcranmer>
look up Michelson-Moore

[07:01:02.0000] <mookid>
they're suggesting I need to come up with a use case and real world examples

[07:01:03.0000] <jcranmer>
you can work on something for a while and still be wrong

[07:01:04.0000] <mookid>
yeah true

[07:01:05.0000] <mookid>
I agree.

[07:01:06.0000] <mookid>
you can get so caught up in a certain way of doing thigns

[07:01:07.0000] <Dashiva>
jcranmer: Morley, you mean?

[07:01:08.0000] <mookid>
that the alternative seems totally unviable

[07:01:09.0000] <jcranmer>
Dashiva: maybe?

[07:01:10.0000] <Philip`>
jcranmer: <troll>Or look up religion</troll>

[07:01:11.0000] <jcranmer>
I don't know off the top of my head

[07:02:00.0000] <jcranmer>
Philip`: what if they're ALL right?

[07:02:01.0000] <hsivonen>
http://en.wikipedia.org/wiki/Michelson-Morley_experiment

[07:03:00.0000] <jcranmer>
g'evening, MikeSmith 

[07:03:01.0000] <mookid>
The strangest thing abuot this is tha tI'm not even the one claiming to be right.. I'm just saying we cant tell

[07:03:02.0000] <mookid>
until you provision it

[07:03:03.0000] <mookid>
there's no feasible way of testing HTTP conneg

[07:03:04.0000] <MikeSmith>
jcranmer: hei

[07:03:05.0000] <mookid>
because developers wont move

[07:03:06.0000] <mookid>
because they cant..

[07:04:00.0000] <hsivonen>
mookid: you don't need to try everything. You can see problems with some things before implementation.

[07:04:01.0000] <mookid>
if your persepective is wrong sure.

[07:05:00.0000] <mookid>
The caching issue is an interesting one which no one seems to have adressed

[07:06:00.0000] <Philip`>
We can't afford to try every idea, so we have to imperfectly predict whether they'll be worthwhile based on what we already know

[07:06:01.0000] <Philip`>
particularly since it's impossible to ever remove a feature once it's been added, even if it turns out to be a terrible idea

[07:07:00.0000] <mookid>
why is it terrible to provide an optional mechanism for controlling Accept headers?

[07:09:00.0000] <Dashiva>
How about an optional mechanism to control font size?

[07:10:00.0000] <mookid>
css?

[07:10:01.0000] <Dashiva>
No, in the resources provided by the server

[07:10:02.0000] <mookid>
css?

[07:10:03.0000] <Dashiva>
CSS in a powerpoint file? A pdf?

[07:10:04.0000] <mookid>
er what?

[07:11:00.0000] <mookid>
we talkign abuot HTML here or not?

[07:11:01.0000] <Philip`>
/me doesn't understand Dashiva

[07:11:02.0000] <Dashiva>
<a href="file.pdf" accept-font-size="14pt">

[07:11:03.0000] <Dashiva>
So it doesn't give me documents where I have to zoom in

[07:11:04.0000] <jcranmer>
:-)

[07:12:00.0000] <jcranmer>
/me winks

[07:12:01.0000] <mookid>
HTTP conneg is a bti mroe fundamental than font size

[07:12:02.0000] <jcranmer>
but why not?

[07:12:03.0000] <Dashiva>
Surely adding just this one attribute is easy

[07:12:04.0000] <Dashiva>
And we don't know if it'll work before we try

[07:12:05.0000] <jcranmer>
it's not like it's terribly costly to implement

[07:12:06.0000] <mookid>
if you think that's the same thing that explains alot

[07:12:07.0000] <mookid>
:)

[07:12:08.0000] <hsivonen>
mookid: why? Font size is relevant to all text. Conneg isn't.

[07:13:00.0000] <mookid>
Conneg is relevant to every request

[07:13:01.0000] <mookid>
conneg is part of the browser request

[07:13:02.0000] <jcranmer>
it's also pointing out something of your method

[07:14:00.0000] <mookid>
not really i't sjust more diversionary rubbish

[07:14:01.0000] <mookid>
to be expected

[07:14:02.0000] <jcranmer>
you rejected it offhand without giving it thought

[07:14:03.0000] <jcranmer>
the priniciples are the same

[07:15:00.0000] <jcranmer>
we essentially gave the same "reasons" that you did

[07:16:00.0000] <mookid>
we?

[07:16:01.0000] <jcranmer>
Dashiva and I

[07:16:02.0000] <mookid>
I really hope you arent involved in this in any significant sense

[07:16:03.0000] <mookid>
you've added nothing so far.

[07:16:04.0000] <jcranmer>
you've convinced us of nothing

[07:17:00.0000] <Dashiva>
Have you talked to XHTML2 yet?

[07:17:01.0000] <mookid>
not yet

[07:17:02.0000] <mookid>
why?

[07:18:00.0000] <Dashiva>
Just wondering

[07:18:01.0000] <Philip`>
mookid: That's just the standard "go talk to some group we don't like because maybe your idea is the kind of crazy thing they'll like" insult

[07:19:00.0000] <Dashiva>
Busted!

[07:19:01.0000] <Dashiva>
/me shakes fist

[07:19:02.0000] <Philip`>
rather than a constructive suggestion

[07:19:03.0000] <mookid>
god forbid.

[07:19:04.0000] <Philip`>
/me takes all Dashiva's guns and $100

[07:19:05.0000] <jcranmer>
Philip`: you didn't have to say that out loud ;-)

[07:19:06.0000] <Dashiva>
But it's also a place where you're more likely to find agreement

[07:20:00.0000] <Dashiva>
Since they're more interested in architectural purity than HTML5 is

[07:22:00.0000] <ehird>
mookid is still talking?

[07:22:01.0000] <ehird>
wow :D

[07:22:02.0000] <Philip`>
It's also a place where (according to the groupmind here) the harmful effects of any crazy ideas will be prevented because nobody will implement XHTML2

[07:23:00.0000] <Dashiva>
We don't get to be a hivemind?

[07:23:01.0000] <Philip`>
so it's a safe place to send people to

[07:23:02.0000] <jcranmer>
Philip`: so XHTML2 is the new ACAP?

[07:23:03.0000] <Philip`>
jcranmer: An Agreement on the Conservation of Albatrosses and Petrels? I don't think it is

[07:24:00.0000] <Dashiva>
Say, is html4all still around?

[07:24:01.0000] <jcranmer>
+++ Binet-Simon test: 1st modern intel test

[07:24:02.0000] <jcranmer>
++++ abstract verbal reason skills

[07:24:03.0000] <jcranmer>
++++ metnal age: metal ability typical of child

[07:24:04.0000] <jcranmer>
+++ ->

[07:24:05.0000] <jcranmer>
dammit

[07:25:00.0000] <jcranmer>
http://www.rfc-editor.org/rfc/rfc2244.txt

[07:25:01.0000] <jcranmer>
(forgot to Ctrl-C the URI)

[07:26:00.0000] <Philip`>
Dashiva: Their public list isn't dead, but I guess all the juicy stuff is happening behind closed doors

[07:26:01.0000] <Philip`>
(or maybe it isn't happening at all, but that would be boring so I prefer my world-view)

[07:27:00.0000] <Philip`>
jcranmer: What is its analogy with XHTML2?

[07:27:01.0000] <jcranmer>
Philip`: dead, highly dead

[07:27:02.0000] <jcranmer>
only one server implementation

[07:28:00.0000] <Dashiva>
Mr. Last week is failing

[07:28:01.0000] <Dashiva>
His suggestion for the new doctype to be <HIXIE> isn't compatible with standards mode

[07:29:00.0000] <Philip`>
jcranmer: There are plenty of RFCs without even that many implementations

[07:30:00.0000] <hsivonen>
Dashiva: and now there'd even be IRC material available

[07:30:01.0000] <ehird>
Dashiva: <!doctype hixie>

[07:30:02.0000] <ehird>
<hixie>...</hixie>

[07:30:03.0000] <Dashiva>
Oh right

[07:31:00.0000] <Dashiva>
We'll be famous for this

[07:31:01.0000] <Dashiva>
*infamous

[07:31:02.0000] <ehird>
just name every element after someone in here.

[07:31:03.0000] <jcranmer>
Philip`: I've always used ACAP as my example of a dead protocol

[07:31:04.0000] <Philip`>
Dashiva: That's why we just need everyone to upgrade to the Hixiefox and Hixiekit and Hixplorer and Hixpera browsers, and they can render it in standards mode

[07:32:00.0000] <Dashiva>
Philip`: Surely it would then be hixie mode

[07:32:01.0000] <jcranmer>
and quirks mode is hixie pixie mode!

[07:33:00.0000] <ehird>
<!doctype hixie><hixie><philip>hello world</philip><dashiva>An example Hixie5 page</dashiva><danbri>This is a paragraph</danbri></hixie>

[07:33:01.0000] <Philip`>
Actually, Hixplorer sounds stupid, it should be HixIE

[07:33:02.0000] <Dashiva>
FireFoxie

[07:33:03.0000] <ehird>
Philip`: HixIEplorer

[07:34:00.0000] <Dashiva>
What was HTML again, Hixie something Markup Language

[07:34:01.0000] <danbri>
that should be <danbri xmlns='http://danbri.org/' /> :)

[07:34:02.0000] <ehird>
Dashiva: It's HIXIE

[07:34:03.0000] <ehird>
Stands for Hixie Ixie Xie Ie E

[07:34:04.0000] <ehird>
Because it's an "echo" of the content. Totally.

[07:35:00.0000] <ehird>
danbri: http://danbri.org/danbri

[07:35:01.0000] <Dashiva>
ehird: Yeah, but we had an expansion of HTML last year

[07:35:02.0000] <ehird>
:p

[07:35:03.0000] <ehird>
ah, wait

[07:35:04.0000] <ehird>
hmm, no

[07:38:00.0000] <Philip`>
jcranmer: You could use HTML 3.0 as a more appropriate example, perhaps

[07:38:01.0000] <jcranmer>
HTML 1.0

[07:39:00.0000] <Philip`>
/me likes how HTML 3.0's <math> element defines a new syntax that would parse to the same DOM tree as the usual SGMLish tags

[07:40:00.0000] <Philip`>
e.g. <math>x^{1+2}</math> seems to be the same as <math>x<sup><box>1+2</box></sup></math>

[07:40:01.0000] <ehird>
wow, never heard of that

[07:41:00.0000] <Philip`>
since that's a nice extension of the idea that the character-level syntax and parsed representation are really independent concepts

[07:41:01.0000] <Lachy>
Dashiva, s/FireFoxie/FireHixie/

[07:42:00.0000] <Philip`>
Oh, actually, I think that should be <math>x^{1+2}^</math>

[07:43:00.0000] <Philip`>
Fire Hixie? Then we wouldn't have an editor

[07:43:01.0000] <danbri>
that is interesting, yeah

[07:49:00.0000] <Philip`>
Easy solution to the problem of thinking of precise antonyms for rel values: use some pattern along the lines of rel="rev-author"

[07:52:00.0000] <Lachy>
I'm not even convinced of the need for rel='author", and even less convinced of the need for the reverse relationship

[07:54:00.0000] <gsnedders>
/me awves

[07:54:01.0000] <gsnedders>
*&waves

[07:54:02.0000] <Lachy>
and considering that the concept of a reverse relationship is difficult for authors to grasp correctly, it would be a mistake to add it simply to cater for theoretical cases

[07:54:03.0000] <gsnedders>
**waves

[08:02:00.0000] <jcranmer>
wouldn't *&waves = waves?

[08:03:00.0000] <Philip`>
Not if you overload the unary operator*

[08:03:01.0000] <jcranmer>
let's talk C here

[08:03:02.0000] <Philip`>
How boring

[08:04:00.0000] <Dashiva>
I'm still waiting for D to become popular

[08:04:01.0000] <Dashiva>
It looks kinda neat

[08:04:02.0000] <jcranmer>
doesn't DTrace use D?

[08:04:03.0000] <jcranmer>
oh wait, it's a different D

[08:04:04.0000] <Philip`>
jcranmer: That's a completely different D

[08:04:05.0000] <Philip`>
and isn't even Turing-complete

[08:04:06.0000] <jcranmer>
D therefore shouldn't be popular

[08:04:07.0000] <jcranmer>
don't want to confuse people

[08:05:00.0000] <Philip`>
Dashiva: It looks kind of neat, but not neat enough to overcome the huge legacy advantage of C for systems programming, or to compete with proper languages for other types of programming

[08:06:00.0000] <jcranmer>
any new language should be compatible with the C ABI

[08:06:01.0000] <jcranmer>
for bonus point

[08:07:00.0000] <jcranmer>
support C++ ABI

[08:08:00.0000] <Philip`>
That's kind of hard when even C++ doesn't support the C++ ABI (if you mix compilers or compiler versions)

[08:08:01.0000] <Philip`>
(unless I'm sadly mistaken about this)

[08:09:00.0000] <Philip`>
Passing STL objects across DLL boundaries is fun

[08:10:00.0000] <jcranmer>
ooh, I got caught by something similar

[08:11:00.0000] <jcranmer>
an object I was working with (growable array) wouldn't work if the original array had 0 size, but with non-zero size, it worked

[08:12:00.0000] <jcranmer>
all because of hidden static variables

[08:12:01.0000] <jcranmer>
basically, &sEmptyHdr != &sEmptyHdr :-)

[08:28:00.0000] <yecril71>
Lynx uses LINK[rev="made"] for me to complain about a Web page.

[08:28:01.0000] <yecril71>
Quite handy: press "C", no need to look around.

[08:29:00.0000] <yecril71>
Is it possible to have type of a hyperlink influence the Accept header the browser sends?

[08:30:00.0000] <yecril71>
s/possible/reasonable/

[08:31:00.0000] <Philip`>
I imagine its handiness is significantly reduced by the problem that fewer than 1% of pages have rev=made

[08:32:00.0000] <Philip`>
and shouldn't that be using LINK[rel="author"] anyway?

[08:33:00.0000] <yecril71>
Perhaps it should; only Lynx does not support it.

[08:34:00.0000] <Philip`>
/me sees that nearly all <link rev=made> have href=mailto:...

[08:34:01.0000] <yecril71>
So, to make my page conveniently usable for Lynx users, I have to use rev.

[08:34:02.0000] <ehird>
the maker of the document is who you should contact about maintaining it

[08:34:03.0000] <ehird>
the author of a document is the author of the content in the document

[08:35:00.0000] <yecril71>
Now you are getting architectonic...

[08:35:01.0000] <yecril71>
I have decribed the reality, and the reality is brutal.

[08:36:00.0000] <Philip`>
/me sees about as many <link rev=made href=example⊙ec> (forgetting the "mailto:") as he sees <link rev=made href=http://...>

[08:37:00.0000] <Philip`>
whereas rel=author most commonly links to a page instead of an email address

[08:37:01.0000] <yecril71>
I guess that means it should really be rel="publisher"?

[08:38:00.0000] <Philip`>
so it seems the semantics are defined more by convention (i.e. people copying examples and tutorials that use rel=author and rev=made for pages and emails) than by what anyone intended the values to really mean

[08:39:00.0000] <yecril71>
rev=made is required as implemented.

[08:50:00.0000] <Lachy>
does anyone here understand RB's distinction of "browser behaviour" and "HTML parsing"?

[08:52:00.0000] <Dashiva>
does anyone here understand RB?

[08:53:00.0000] <Lachy>
rarely

[08:55:00.0000] <yecril71>
Is it reasonable to expect that specifying the type attribute of a hyperlink

[08:55:01.0000] <yecril71>
will also influence the Accept header sent by the browser?

[08:55:02.0000] <Dashiva>
Give an example?

[08:57:00.0000] <Philip`>
"making HTML depend on HTTP-specific features sounds like A Bad Idea" - like <form method> and <a ping> and <eventsource> and so on?

[09:00:00.0000] <yecril71>
<A HREF="report" TYPE="application/pdf" >?

[09:01:00.0000] <Dashiva>
I'm getting this intense feeling of deja vu here

[09:01:01.0000] <yecril71>
BTW, what does METHOD=POST mean for non-HTTP?

[09:02:00.0000] <yecril71>
D�ja vu of what?

[09:03:00.0000] <ehird>
yecril71: aaah!

[09:03:01.0000] <ehird>
kill!

[09:03:02.0000] <yecril71>
(except that METHOD=POST is a good way to avoid opening a HTA in IE)

[09:03:03.0000] <ehird>
begon, clone of mookid!

[09:04:00.0000] <yecril71>
Please, I only want to understand what is already supported.

[09:04:01.0000] <ehird>
i was joking

[09:04:02.0000] <ehird>
:)

[09:05:00.0000] <ehird>
but, no

[09:05:01.0000] <ehird>
that won't.

[09:05:02.0000] <Dashiva>
(or was he? DUN DUN DUN)

[09:05:03.0000] <ehird>
Dashiva: i am actually travelling to where yecril71 (aka MOOKID) is right now so that i may cut up his intestines

[09:05:04.0000] <ehird>
... only kidding! haha!

[09:05:05.0000] <yecril71>
Joking is a good thing when you are also willing to help.

[09:05:06.0000] <Dashiva>
yecril71: The user agent can use the value as a hint if it wants to

[09:06:00.0000] <Dashiva>
But as far as I know, none of the major browsers do

[09:06:01.0000] <Philip`>
yecril71: http://www.whatwg.org/specs/web-apps/current-work/multipage/forms.html#form-submission-algorithm step 14 defines what it means in HTML5 for any protocol

[09:06:02.0000] <ehird>
yecril71: i answered

[09:06:03.0000] <ehird>
<ehird> but, no

[09:06:04.0000] <ehird>
<ehird> that won't.

[09:06:05.0000] <ehird>
re: <yecril71> <A HREF="report" TYPE="application/pdf" >?

[09:07:00.0000] <yecril71>
I must have missed your answer

[09:10:00.0000] <yecril71>
MSHTA treats GET file: as mutation.

[09:11:00.0000] <yecril71>
I do not know about ftp.

[09:11:01.0000] <yecril71>
Why is file: out of scope?

[09:12:00.0000] <yecril71>
OTOH, it does not modify the URL for POST file:, which is good.

[09:15:00.0000] <Philip`>
Oops, when I said "any protocol" I was lying

[09:15:01.0000] <yecril71>
"Attributes don�t reflect" means that getAttribute returns the initial value specified.

[09:16:00.0000] <Philip`>
yecril71: file: is out of scope because it's not considered an area where interoperability is needed, and UAs can just do whatever they want

[09:17:00.0000] <Philip`>
(I'm not entirely sure how to argue for why interoperability isn't needed in that case, though)

[09:19:00.0000] <yecril71>
Do you thing that METHOD=POST is a wrong thing overall?

[09:19:01.0000] <yecril71>
s/thing/think/

[09:19:02.0000] <Philip`>
Uh... No?

[09:20:00.0000] <yecril71>
It works with HTTP only, and you are for separation of HTML and HTTP.

[09:20:01.0000] <Philip`>
I'm not for separation of HTML and HTTP :-)

[09:21:00.0000] <yecril71>
But here HTML depends on HTTP-specific features,

[09:21:01.0000] <yecril71>
which is a bad thing, Philip` said.

[09:21:02.0000] <yecril71>
I am trying to decipher that text.

[09:23:00.0000] <Philip`>
Do you mean when I said

[09:23:01.0000] <Philip`>
17:00 < Philip`> "making HTML depend on HTTP-specific features sounds like A Bad Idea" - like <form method> and <a ping> and <eventsource> and so on?

[09:23:02.0000] <Philip`>
?

[09:23:03.0000] <ehird>
yecril71: he was quoting

[09:23:04.0000] <ehird>
and using sarcasm

[09:23:05.0000] <ehird>
to rebut it :P

[09:23:06.0000] <Philip`>
That was quoting jcranmer

[09:24:00.0000] <Philip`>
/me attempts to increase the ratio of technical discussion on public-html, by pointing out a couple of terribly exciting typos

[09:24:01.0000] <jcranmer>
method is something that could be easily specifiable to not be protocol-dependent (in my eyes at least)

[09:25:00.0000] <jcranmer>
I'm not a fan of <a ping> to be honest

[09:25:01.0000] <jcranmer>
and I've not read up on <eventsource>

[09:28:00.0000] <yecril71>
How would you POST to file:?

[09:28:01.0000] <ehird>
replace the file, obviously :-P

[09:28:02.0000] <ehird>
<form method="POST" action="/etc/passwd">

[09:28:03.0000] <yecril71>
That is not POST, that is PUT

[09:28:04.0000] <ehird>
well, true.

[09:29:00.0000] <ehird>
but it'd amount to the same thing for file:, probably.

[09:29:01.0000] <yecril71>
But it would be very different from what GET does,

[09:29:02.0000] <yecril71>
whereas it is not that different for http:.

[09:30:00.0000] <jcranmer>
get would be a URI composition

[09:30:01.0000] <yecril71>
How do you POST a file to file:?

[09:31:00.0000] <yecril71>
The exciting thing is POST, not GET.

[09:31:01.0000] <yecril71>
GET is already covered, assuming file: is similar to ftp:.

[09:32:00.0000] <jcranmer>
looks like Hixie disagrees with me on GET

[09:32:01.0000] <yecril71>
What does look like that?

[09:33:00.0000] <jcranmer>
a non-Flash, non-jemalloc-related FF crash?

[09:34:00.0000] <yecril71>
Am I getting every sixth message or what?

[09:57:00.0000] <yecril71>
All right, I got your point with GET.

[10:23:00.0000] <takkaria>
christ, it's obviously troll season at the moment

[10:25:00.0000] <yecril71>
I cannot see any at the moment.

[10:28:00.0000] <Philip`>
/me sees that occam has a keyword that is defined as performing an infinite loop

[10:29:00.0000] <Philip`>
which seems a peculiar but perfectly sensible thing to do

[10:33:00.0000] <takkaria>
I assume you can break out of it?

[10:43:00.0000] <tndH>
eh, 36 emails since yesterday

[10:43:01.0000] <tndH>
and i went to bed at 2 :|

[10:51:00.0000] <yecril71>
If a resource with alternative formats gets updated,

[10:51:01.0000] <yecril71>
all formats should be updated and not just one.

[10:52:00.0000] <yecril71>
Where "updated" can mean "removed until someone handles it".

[10:52:01.0000] <yecril71>
That is the publisher�s responsibility and it can be enforced by the server.

[10:54:00.0000] <yecril71>
However, the mechanisms of updating resources are often different from the mechanisms of viewing it.

[10:55:00.0000] <yecril71>
And that includes the URL, e.g. update ftp:, get http:.

[11:11:00.0000] <Philip`>
takkaria: No

[11:12:00.0000] <Philip`>
takkaria: but the language is designed to run stuff in parallel

[11:12:01.0000] <Philip`>
so all you're doing is sending one thread into an infinite loop

[11:12:02.0000] <Philip`>
which is observationally identical to terminating the thread

[11:12:03.0000] <Philip`>
(because it'll just stop outputting anything)

[11:13:00.0000] <Philip`>
(hence the keyword is "STOP")

[11:18:00.0000] <hallvors>
TAG fed the trolls..

[11:34:00.0000] <takkaria>
Philip`: heh

[11:36:00.0000] <takkaria>
yecril71: see, all formats should be updated, but not all will be, always.  and whilst a server *can* enforce this, it requires someone going out of their way to write code to do that, and as such will probably not be enforced most of the time

[11:54:00.0000] <yecril71>
Providing alternative representations already is going out of one�s way already.

[11:54:01.0000] <yecril71>
Once you decide to do that, you should make sure you have all the tools necessary to maintain this setup.

[12:08:00.0000] <webben>
Can anyone remember whether the problem with http://lists.w3.org/Archives/Public/public-html/2008Aug/0116.html was CSS generated text (content: "TODO") or just a border?

[12:09:00.0000] <webben>
wayback doesn't seem to have captured the relevant CSS

[12:28:00.0000] <Philip`>
In practice, if you have multiple formats for some resource, you'd quite possibly want to do the format conversion asynchronously as a batch operation rather than blocking the person who's doing the uploading while your really slow video transcoding / PDF OCRing / etc process goes on

[12:29:00.0000] <Philip`>
and so invalidating the cache at the moment of upload is not sufficient, since you'll need to invalidate all the other formats at some unknown time in the future

[12:31:00.0000] <Philip`>
webben: I'm pretty sure the CSS generated content was added ages ago, long before that email

[12:32:00.0000] <Philip`>
Maybe gsnedders could add non-CSS generated content support into Anolis :-)

[12:32:01.0000] <gsnedders>
Philip`: For what?

[12:33:00.0000] <gsnedders>
Philip`: I have feedback from Hixie about the ** stuff, and doing stuff to issues

[12:33:01.0000] <Philip`>
gsnedders: For the places where the spec currently uses CSS to add content:"Note" and content:"Warning" etc, since that's not just stylistic information and it means some UAs can't distinguish those notes/warnings

[12:34:00.0000] <Philip`>
*can't distinguish from plain text

[12:34:01.0000] <gsnedders>
Philip`: Yeah' I have an email from Hixie detailing what he wants there

[12:34:02.0000] <gsnedders>
s/'/,/

[12:34:03.0000] <Philip`>
gsnedders: Okay

[12:34:04.0000] <Philip`>
webben: Blame gsnedders for not fixing it yet :-)

[12:35:00.0000] <gsnedders>
Blame Apple for making a laptop that broke slowing me down :)

[12:36:00.0000] <Philip`>
That's your fault for buying shoddy hardware

[12:36:01.0000] <Philip`>
You should have got a Dell

[12:38:00.0000] <Philip`>
Also you should have got a RAIL setup, so you have a hot-spare laptop to fall back on if one breaks

[12:39:00.0000] <gsnedders>
Philip`: Buy me one. :P

[12:39:01.0000] <Philip`>
Don't try shifting the blame!

[12:41:00.0000] <Dashiva>
I wonder if the people complaining about DOM are aware that the charter includes DOM APIs

[12:42:00.0000] <gsnedders>
Our charter doesn't :P

[12:44:00.0000] <Philip`>
Be careful about arguing based on what the charter says, because other people might start doing the same when arguing against you on other issues :-)

[12:44:01.0000] <Dashiva>
Are we talking about the same charter?

[12:45:00.0000] <Philip`>
/me assumes gsnedders was making some point about WHATWG vs HTML WG

[12:45:01.0000] <gsnedders>
Dashiva: Philip` understands me better than you.

[12:46:00.0000] <Dashiva>
A charter that nobody lawyers about isn't interesting :)

[12:52:00.0000] <hsivonen_>
/me wishes the charter didn't have earmarks

[12:52:01.0000] <hsivonen_>
hmm. probably not the right political word

[12:55:00.0000] <Philip`>
Maybe "riders"?

[12:56:00.0000] <hsivonen_>
Philip`: that's the word I was looking for, yes

[12:56:01.0000] <hsivonen_>
thanks

[14:07:00.0000] <hallvors>
does anyone know documentation for differences in capabilities in IE between an XMLHttpRequest object created with ActiveXObject() and a native XMLHttpRequest() object?

[14:08:00.0000] <hallvors>
if that's hard to parse: it turns out that the object returned from "new XMLHttpRequest()" and "new ActiveXObject('MSXML2.XMLHTTP.3.0')" have different capabilities. I'd like to know if this is known/documented.

[14:10:00.0000] <gsnedders>
hallvors: IIRC they return exactly the same thing )the native object is just a wrapper)

[14:21:00.0000] <Hixie>
hm, burns is late with his trolls. he hasn't said anything since sunday.

[14:21:01.0000] <Hixie>
maybe he can only troll on weekends now.

[14:21:02.0000] <Hixie>
i wonder if he'll ever reply to http://lists.w3.org/Archives/Public/public-html/2008Nov/0161.html

[14:22:00.0000] <blooberry>
hallvors: using the ActiveXObject method requests a specific instantiation of a particular object. I believe there were previous iterations of xmlhttp? So, it makes sense that there might be some differences depending on which ActiveXObject you are requesting to create. The question I would have is: is there a version string for ActiveXObject that has the same behavior as new XMLHttpRequest? That would probably indicate which ActiveXO

[14:22:01.0000] <blooberry>
using as its default.

[14:38:00.0000] <Lachy>
Hixie, apparently, according to the comments in the forum, a new version of Requiem 1.8.2 has been released.

[14:39:00.0000] <Lachy>
I'm still waiting for my copy of freenet to update the download page for me before I can get it

[14:41:00.0000] <Philip`>
/me suggests using the web to download it instead

[14:44:00.0000] <Lachy>
Philip`, I would, but I can't find it elsewhere on the web yet

[14:49:00.0000] <Philip`>
Is there anything to stop some random person releasing a file on Freenet and claiming it's a new version of Requiem?

[14:50:00.0000] <takkaria>
hashes?

[14:51:00.0000] <Lachy>
they could, but only Brahms is able to update the official Requiem page with the links to the files

[14:52:00.0000] <Lachy>
so if you found a link to the file from elsewhere, you could be fooled into downloading the wrong file, but that's the same with anything.

[14:52:01.0000] <hallvors>
gsnedders, blooberry: I have a brand new test case where IE only supports xhr.open('SEARCH', ...) if the XHR object is created with the ActiveX call. Weird, and I have no idea why.

[15:01:00.0000] <blooberry>
hallvors: that is odd. Did you see if there are any other ActiveX strings that would work for XHR?

[15:03:00.0000] <hallvors>
since the next TC I wrote actually froze IE (IE8 beta2, that is) I haven't gotten much further yet, no :-p

[15:03:01.0000] <Hixie>
Lachy: i'm still on 1.7.x

[15:03:02.0000] <Hixie>
Lachy: but good to know :-)

[15:03:03.0000] <Hixie>
Lachy: any idea what it adds?

[15:04:00.0000] <blooberry>
hallvors: 8-}

[15:12:00.0000] <hallvors>
fun. tried all the identifiers from http://en.wikipedia.org/wiki/XMLHttpRequest 

[15:12:01.0000] <hallvors>
they all work. only "new XMLHttpRequest()" makes it fail

[15:16:00.0000] <Lachy>
Hixie, I don't know yet. It probably adds windows support

[15:17:00.0000] <Hixie>
ah ok

[15:17:01.0000] <Lachy>
yeah, it does. My freenet node has updated now

[15:17:02.0000] <Hixie>
/me finds his mac is crashing on loginwindow on startup now

[15:17:03.0000] <Hixie>
wtf

[15:17:04.0000] <blooberry>
hallvors: I would have expected ActiveXObject to be the more restrictive specification method as MS has had the most probs with that tech over the years. 

[15:22:00.0000] <hallvors>
If they have started limiting what HTTP verbs you're allowed to use it's pretty interesting. Especially since Outlook Web Access uses lots of funky verbs including SEARCH. The OWA team might not like IE8... :p

[15:23:00.0000] <takkaria>
they'll be in IE7mode, I bet


2008-11-19
[16:37:00.0000] <hallvors>
where did TimBL speak about "change in philosophy" from "improving the web" to "letting it fester while describing it" as one of his concerns about HTML5? I'm writing a blog post and trying to find the source of that quote.. #-]

[16:39:00.0000] <Philip`>
hallvors: http://www.w3.org/2001/tag/2008/09/23-minutes ?

[16:39:01.0000] <Philip`>
(It doesn't seem clear whether that's his own opinion, or him reporting other people's apparent opinions)

[16:42:00.0000] <hallvors>
thanks!

[16:42:01.0000] <hallvors>
2001 in the URL?! didn't realise it was such an ancient quote

[16:42:02.0000] <Philip`>
2008 is in the URL too

[16:42:03.0000] <ehird>
TAG started 2001

[16:42:04.0000] <ehird>
talk was 2008

[16:42:05.0000] <hallvors>
right. cool URIs must confuse

[16:43:00.0000] <hallvors>
:p

[16:43:01.0000] <ehird>
indeed.

[16:43:02.0000] <ehird>
/1995/news for stuff happening in 2130 makes soo much sense

[16:43:03.0000] <ehird>
very durable :p

[16:43:04.0000] <Hixie>
are you mocking the web's architecture?

[16:44:00.0000] <Hixie>
careful, you might get in trouble!

[16:44:01.0000] <ehird>
/me gets mauled.

[16:44:02.0000] <ehird>
By web owls.

[16:45:00.0000] <ehird>
They're just like regular owls, except they sting you with cool URIs and... um... resource talons.

[17:04:00.0000] <hallvors>
URLs are just the UI of HTTP. As all UIs they are a pretty bad compromise between what a human can get used to and what a computer can understand.

[17:28:00.0000] <jwalden>
"modularized"?

[17:28:01.0000] <jwalden>
sigh

[17:30:00.0000] <jwalden>
oh, it gets worse, they mention E4X

[17:31:00.0000] <jwalden>
apparently not seriously, tho, which is a relief

[17:34:00.0000] <Philip`>
Wasn't it just hsivonen_ who mentioned E4X?

[17:34:01.0000] <jwalden>
possibly, I don't know who any of these people are except the one obvious one and DanC

[18:45:00.0000] <MikeSmith>
http://www.w3.org/QA/2008/06/shipbuilding.html#c169011

[18:46:00.0000] <MikeSmith>
Dmitry Turin: "Western world does not interested in implementation of progressive ideas, because Indian slaves have been executing job for Western world free of charge"... "Nobody allow me to kill his investments, even it's investment into frankly nonsense (you guessed, what term i implied)"... "P.S. >to produce standards that actually get implemented"

[18:49:00.0000] <MikeSmith>
I think I like the old mostly just quixotic and relatively conspiracy-theory-free Dmitry Turin better than this one

[18:49:01.0000] <MikeSmith>
but you do have to give him credit in that even when he thinks conspiracy theory, he thinks big and goes all the way

[18:50:00.0000] <MikeSmith>
e.g, the entire Western world vs. his ideas

[23:29:00.0000] <Hixie>
aw, all the threads died out

[23:51:00.0000] <Hixie>
i guess it makes sense that roy would stop responding after i showed an interest in listening to his feedback

[23:51:01.0000] <Hixie>
since that would break his world view of us being people who ignore his feedback

[00:17:00.0000] <Hixie>
where's Philip`'s static copy of the whatwg issues list at again?

[00:43:00.0000] <Philip`>
Hixie: http://canvex.lazyilluminati.com/misc/cgi/issues.cgi/

[00:43:01.0000] <Hixie>
thanks

[02:06:00.0000] <hsivonen_>
it feels weird that rickg's dated comments in nsIParser and friends are now over ten years old

[02:07:00.0000] <yecril71>
If a script exhausts browser resources, an exception should be thrown.

[02:08:00.0000] <yecril71>
(unless the failing action is known to be discardable)

[02:11:00.0000] <yecril71>
We can define that the HTML token "red" really means "blue",

[02:11:01.0000] <yecril71>
but I am afraid this is going to cause many misunderstandings.

[02:14:00.0000] <yecril71>
rel="rev-made" is not very good.

[02:14:01.0000] <hsivonen_>
yecril71: consider what danbri said about changing author to creator in DC. it wasn't about changing the concept but about effective bikeshedding the English word

[02:14:02.0000] <yecril71>
rel="made-me" sounds much better.

[02:14:03.0000] <yecril71>
or rel="made-by".

[02:14:04.0000] <Philip`>
Where is rev=made defined?

[02:15:00.0000] <yecril71>
Lynx documentation.

[02:15:01.0000] <Philip`>
Is there something more like a spec where it's defined?

[02:16:00.0000] <yecril71>
The only problem with removing rev is that Lynx uses it.

[02:16:01.0000] <hsivonen_>
for all practical purposes, rev=made is an ancient Lynx-only feature

[02:17:00.0000] <hsivonen_>
yecril71: we've made IE-only UI-only stuff non-conforming, too.

[02:18:00.0000] <yecril71>
I thought it was sort a political decision, IE being the equivalent of Dr. No?

[02:19:00.0000] <yecril71>
Whereas Lynx, as it is used by disabled people, should be treated with some condescending?

[02:20:00.0000] <Philip`>
More disabled people use IE than any other browser, from what I've heard

[02:23:00.0000] <yecril71>
There are various disabilities, perhaps IE is good for some of them?

[02:24:00.0000] <yecril71>
Using numbers in this context is rather dangerous.

[02:25:00.0000] <yecril71>
Most of the people are not disabled.

[02:27:00.0000] <mookid>
Hi fans

[02:27:01.0000] <yecril71>
Hi chief

[02:33:00.0000] <mookid>
well the email barrage dried up which is a shame!

[02:35:00.0000] <yecril71>
Time to catch up with things

[02:38:00.0000] <yecril71>
a rel="in-reply-to" is nonsense.

[02:39:00.0000] <yecril71>
It can be used as LINK[rel="in-reply-to"] only, to describe this document.

[02:42:00.0000] <yecril71>
And that would mean THAT document is in reply to THIS document.

[02:44:00.0000] <yecril71>
In this sense, rel="made" and rev="made" have the same meaning.

[02:45:00.0000] <yecril71>
rel="made" means THAT made THIS

[02:45:01.0000] <yecril71>
and so does rev="made".

[02:46:00.0000] <Hixie>
i think you misunderstand how rev="" was meant to work

[02:46:01.0000] <Hixie>
but it's academic

[02:46:02.0000] <Hixie>
since rev="" is gone now

[02:49:00.0000] <zcorpan>
hsivonen_: here's a feature request that i tried to implement myself but gave up before having a proof of concept impl...

[02:49:01.0000] <zcorpan>
hsivonen_: have the messages inline with the source

[02:50:00.0000] <yecril71>
I think we can all agree that LINK[rel="next"] means THAT is next to THIS.

[02:51:00.0000] <yecril71>
So LINK[rel="made"] means THAT made THIS, doesn�t it?

[02:51:01.0000] <zcorpan>
hsivonen_: so that you get e.g.

[02:51:02.0000] <zcorpan>
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd"> ↩

[02:51:03.0000] <zcorpan>
1. Error: Almost standards mode doctype. Expected <!DOCTYPE html>.

[02:51:04.0000] <zcorpan>
<html>↩

[02:51:05.0000] <zcorpan>
<head>↩

[02:51:06.0000] <zcorpan>
<meta http-equiv="Content-Type" content="text/html; charset=windows-1252">↩

[02:51:07.0000] <zcorpan>
2. Error: The internal character encoding declaration must be the first child of the head element.

[02:51:08.0000] <zcorpan>
etc

[02:52:00.0000] <zcorpan>
if there are multiple errors on the same line you have multiple messages for that line

[02:54:00.0000] <zcorpan>
i tried implementing it by walking unsortedOl and doing a regexp on the location para to get the line and then inserting the message into the line-minus-one-th child of the source ol

[02:54:01.0000] <yecril71>
I was just trying to figure out what LINK[rel="banana"] would mean.

[02:54:02.0000] <yecril71>
THAT is MY banana?

[02:55:00.0000] <hsivonen_>
zcorpan: I think that could be a UI improvement, yes.

[02:55:01.0000] <hsivonen_>
IIRC, PageValet had something along those lines a few years ago.

[02:57:00.0000] <hsivonen_>
zcorpan: btw, in case you're wondering about the slow reaction to your recent bug reports, it's because I'm trying to reach a particular calendar goal with HTML5 parsing in Gecko. I'll get back to Validator.nu bugs in due course.

[02:57:01.0000] <zcorpan>
hsivonen_: no worries

[02:58:00.0000] <Hixie>
yecril71: it means whatever we define it to mean, these keywords are opaque strings

[02:58:01.0000] <Hixie>
ok bed time

[02:58:02.0000] <Hixie>
nn

[02:58:03.0000] <zcorpan>
i'm used to Hixie's rate of response i.e. somewhere between immediately and a few years :)

[03:00:00.0000] <yecril71>
But we cannot define rel="banana" to mean something unrelated to bananas.

[03:00:01.0000] <hsivonen_>
yecril71: yes, we can.

[03:00:02.0000] <yecril71>
But it would be confusing to the authors.

[03:00:03.0000] <hsivonen_>
yecril71: yes

[03:01:00.0000] <yecril71>
I think it is not an accident that the HEAD tag was not renamed QWERTY.

[03:01:01.0000] <yecril71>
Although QWERTY would be much easier to type.

[03:02:00.0000] <hsivonen_>
yecril71: there's value in names that make sense, yes.

[03:04:00.0000] <zcorpan>
/me thinks head is easier to type than qwerty

[03:05:00.0000] <hsivonen_>
Hixie: do I understand correctly that it's possible for document.close() to cause previously document.written data to be discarded?

[03:06:00.0000] <hsivonen_>
/me tries it

[03:08:00.0000] <zcorpan>
hsivonen_: do you want me to file a bug about the feature request?

[03:08:01.0000] <hsivonen_>
zcorpan: preferably, yes

[03:08:02.0000] <yecril71>
To me, document.close() means that any subsequent document.write() rewrites the document.

[03:09:00.0000] <hsivonen_>
yecril71: yes, but as far as I can tell, document.close() closes the stream, but the spec says it puts an explicit EOF at insertion point, which seems wrong.

[03:09:01.0000] <yecril71>
The previous content is discarded only at document.write().

[03:09:02.0000] <hsivonen_>
consider

[03:09:03.0000] <hsivonen_>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cscript%3Edocument.write(%22%3Cscript%3Edocument.write(%27a%27)%3Bdocument.close()%3B%3C\%2Fscript%3EEND%22)%3B%3C%2Fscript%3E

[03:11:00.0000] <yecril71>
I am unable to evaluate that in Internet Explorer.

[03:11:01.0000] <hsivonen_>
I'm starting Windows now, but Firefox, Opera and Safari all agree

[03:11:02.0000] <yecril71>
Takes 94% CPU.

[03:12:00.0000] <Hixie>
hsivonen_: oh it should probably insert it at the end of the stream

[03:12:01.0000] <hsivonen_>
Hixie: right. I'll send email.

[03:12:02.0000] <zcorpan>
hsivonen_: filed. fyi i got the internal error message again (worked fine the last few times)

[03:12:03.0000] <Hixie>
hsivonen_: didn't think of testing nested document.close() in document.write(), but i'm sure the web does it somewhere :-)

[03:12:04.0000] <hsivonen_>
zcorpan: thanks. (I still have no clue about the error.)

[03:13:00.0000] <Hixie>
/me hacked the xcode demo npapi plugin today to output the parameters passed to the plugin

[03:13:01.0000] <Hixie>
mac only, sadly

[03:13:02.0000] <Hixie>
but should help spec out what safari and firefox do for <embed> a bit better

[03:13:03.0000] <zcorpan>
Hixie: that would be useful

[03:14:00.0000] <Hixie>
for some reason it crashes when i use <object> in firefox

[03:14:01.0000] <Hixie>
(not in safari)

[03:14:02.0000] <Hixie>
no idea how to debug that

[03:14:03.0000] <hsivonen>
the IE8 XSS filter is really annoying with the live dom viewer

[03:14:04.0000] <Hixie>
and i dunno how to make opera accept a plugin

[03:15:00.0000] <Hixie>
it doesn't seem to use the plugins in the Internet Plugins directory

[03:16:00.0000] <hsivonen>
yay. it indeed freezes IE8 beta2

[03:17:00.0000] <Hixie>
how odd

[03:17:01.0000] <Hixie>
i wonder why

[03:36:00.0000] <hsivonen>
I guess I disagree with Dr. Hoffmann about CURIEs

[03:49:00.0000] <hallvors>
Hixie: add something relevant to plugin path in opera6.ini

[03:51:00.0000] <zcorpan>
Hixie: it's supposed to look in /Library/Internet Plug-Ins

[03:51:01.0000] <hallvors>
well.. that's on Windows, I don't really know on Mac. Sorry.

[04:12:00.0000] <Philip`>
hsivonen: Hixie could modify the Live DOM Viewer to add a header to disable IE8's XSS filter

[04:16:00.0000] <hallvors>
yes, I suggested that on IRC a while ago. the header is  X-XSS-Protection: 0

[04:18:00.0000] <hsivonen>
http://twitter.com/drawohara/statuses/1012168434

[04:19:00.0000] <hsivonen>
is "the shite" a praise while "shite" isn't?

[04:19:01.0000] <BenMillard>
hsivonen, it can be

[04:20:00.0000] <Philip`>
hallvors: http://www.google.com/search?q=site:krijnhoetmer.nl+x-xss-protection suggests it was only me that suggested it on IRC around here :-)

[04:20:01.0000] <hsivonen>
BenMillard: ok. tweeted spec feedback could be less ambiguous :-)

[04:23:00.0000] <takkaria>
I think that twitter is missing a "the", but yeah, seems like praise

[04:24:00.0000] <hsivonen>
takkaria: I thought it was missing "are"

[04:27:00.0000] <Philip`>
http://www.urbandictionary.com/define.php?term=the+shit is seemingly the intended meaning, not http://www.urbandictionary.com/define.php?term=the+shite

[04:29:00.0000] <yecril71>
I believe that keywords like rel="author" should be left in English.

[04:29:01.0000] <yecril71>
They are not meant to be directly visible to the human reader.

[04:30:00.0000] <BenMillard>
Philip`, looks more like vandalism to me.

[04:31:00.0000] <BenMillard>
(especially considering the extremely small number of votes)

[04:34:00.0000] <Philip`>
Hmph, and I thought Urban Dictionary was a trustworthy source :-(

[04:34:01.0000] <takkaria>
hsivonen: uh,yeah,you're right

[04:36:00.0000] <Dashiva>
Mr. Last Week seems to be on the ball for once

[04:37:00.0000] <Philip`>
Hmm, the OED doesn't seem to have the positive definition at all - it just has the dirty ones, and notes "Not now in decent use."

[04:37:01.0000] <hsivonen>
/me notes that the W3C is funded by its corporate masters

[04:37:02.0000] <Philip`>
Dashiva: Like one of those people at the circus who does clever balancing tricks but often falls off?

[04:38:00.0000] <hsivonen>
http://www.w3.org/2008/07/ogws-cfp

[04:38:01.0000] <Dashiva>
Philip`: Yes, and much entertainment is had by all

[04:41:00.0000] <hallvors>
Philip`: http://krijnhoetmer.nl/irc-logs/whatwg/20081008#l-608 :)

[04:42:00.0000] <Philip`>
hallvors: Oh, right :-)

[04:42:01.0000] <Philip`>
Maybe I should have read the logs I was pointing at

[05:02:00.0000] <zcorpan>
hsivonen: i tried my bookmarklet on http://html5.validator.nu/?doc=http%3A%2F%2Fgoogle.se&showsource=yes and it looks really nice... except for "<body bgcolor=#ffffff text=#000000 link=#0000cc vlink=#551a8b alink=#ff0000 onload="sf();if(document.images)new Image().src='/images/nav_logo3.png'" topmargin=3 marginheight=3>"

[05:02:01.0000] <zcorpan>
but i'm not sure that's worth optimizing

[05:04:00.0000] <hsivonen>
zcorpan: I don't follow. Which bookmarklet?

[05:04:01.0000] <zcorpan>
hsivonen: sorry, in the bug. http://bugzilla.validator.nu/show_bug.cgi?id=325

[05:08:00.0000] <hsivonen>
zcorpan: cool! I think it need more styling to distinguish messages and source quotes.

[05:08:01.0000] <zcorpan>
hsivonen: yeah

[05:09:00.0000] <zcorpan>
hsivonen: perhaps insert the <li> instead

[05:12:00.0000] <hsivonen>
I wonder if were feasible to make the source text take 50% of available width and make the messages float into the other 50% of available width and draw a line from the error highlight to the error message CSS box

[05:14:00.0000] <zcorpan>
interesting idea... not sure it would be better though :)

[05:15:00.0000] <hsivonen>
Hixie: what do you suggest for block annotations inserted into inline content?

[05:15:01.0000] <hsivonen>
in terms of content model conformance

[05:15:02.0000] <hsivonen>
I think there's a use case here :-)

[05:17:00.0000] <yecril71>
If you want to indicate a representation, you can use A[type="application/pdf"].

[05:17:01.0000] <yecril71>
There is no need to introduce A[accept].

[05:20:00.0000] <yecril71>
Developers can rig their browsers with all extensions they consider valuable.

[05:20:01.0000] <mookid>
erm

[05:20:02.0000] <mookid>
no.

[05:20:03.0000] <mookid>
maybe if your web app serves a couple of ted in ashed companies

[05:20:04.0000] <mookid>
ted in a shed*

[05:21:00.0000] <mookid>
designing enterprise level web applications you need standards

[05:21:01.0000] <mookid>
not to mention development frameworks

[05:22:00.0000] <yecril71>
The standard, as it is, reflects what developers are unwilling to do.

[05:22:01.0000] <mookid>
type doesn't perform that function btw

[05:22:02.0000] <yecril71>
Why not?

[05:22:03.0000] <mookid>
..?

[05:22:04.0000] <zcorpan>
hsivonen: you just have to split the elements (<b>s and <code>s) in the parent chain until the current element is <li> and then insert a new <ol> there

[05:23:00.0000] <hsivonen>
zcorpan: ok

[05:23:01.0000] <mookid>
yecril71: did you read that email I just sent out?

[05:23:02.0000] <yecril71>
Why isn�t the type attribute good for the purpose?

[05:23:03.0000] <yecril71>
I am just reading it.

[05:25:00.0000] <yecril71>
"Getting involved" is not that easy.

[05:26:00.0000] <yecril71>
I would consider myself happy if I could get involved.

[05:26:01.0000] <yecril71>
And you surely cannot get involved in everything.

[05:26:02.0000] <yecril71>
Not because you do not have enough time, because of the trade secrets.

[05:28:00.0000] <mookid>
well no but the only reason I can find not to do this is "it's too much work"

[05:28:01.0000] <yecril71>
HTML is not hypermedia, HMML is :-)

[05:28:02.0000] <mookid>
whatever :P

[05:29:00.0000] <mookid>
u get the point =)

[05:29:01.0000] <mookid>
a PDF is not the same thing as an HTML document.

[05:29:02.0000] <yecril71>
The problem is that it does not help that you are willing to help.

[05:30:00.0000] <mookid>
well I don't mind not being able to do the work provided someone else does it

[05:30:01.0000] <hsivonen>
mookid: PDF, ODF, OOXML and .doc can have hyperlinks to HTTP resources. Do you classify them as hypermedia formats?

[05:30:02.0000] <mookid>
I still ahvent' been given a good reason why this would not be beneficial

[05:31:00.0000] <mookid>
they will be eventually they aren't used for that currently

[05:31:01.0000] <mookid>
hsivonen: ^

[05:31:02.0000] <yecril71>
It seems nobody is going to work on it.

[05:31:03.0000] <mookid>
oh great.

[05:31:04.0000] <mookid>
well at least we have video tags!

[05:32:00.0000] <hsivonen>
mookid: you have been given reasons why what you are suggesting is bad. Presumably you are classifying the reasons as not "good" reasons.

[05:32:01.0000] <mookid>
no I havent at all

[05:32:02.0000] <mookid>
what are those reasons?

[05:32:03.0000] <hsivonen>
see the logs from yesterday

[05:32:04.0000] <mookid>
I read them

[05:33:00.0000] <jcranmer>
mookid: can you name one drawback of your feature?

[05:33:01.0000] <mookid>
if you look at the conversations I'm having on the mailling list.. I'm constantly repeating myself and referring back to stuff I've already said - just explained differently because peopel aren't comprehending what I'm saying

[05:33:02.0000] <yecril71>
Nobody is going to free Tibet, or Tuva.

[05:34:00.0000] <yecril71>
Or cease cutting the rainforest.

[05:34:01.0000] <yecril71>
There are more dire things to be worried about.

[05:34:02.0000] <mookid>
my skulls tougher than alot of brick walsl

[05:34:03.0000] <jcranmer>
I'll take that as a `no'

[05:34:04.0000] <mookid>
actually there arent - better interoperability of internet applications is paramount to the human race

[05:35:00.0000] <mookid>
-_-

[05:35:01.0000] <hsivonen>
mookid: have you considered the possibility that conneg might be a hindrance to interop?

[05:35:02.0000] <jcranmer>
why do people think that new features come without costs?

[05:35:03.0000] <mookid>
ITS NOT A NEW FEATURE

[05:35:04.0000] <mookid>
it's part of the protocol

[05:35:05.0000] <mookid>
jesus.

[05:36:00.0000] <Dashiva>
And browsers support it

[05:36:01.0000] <hsivonen>
mookid: you are suggesting that something be added. ergo, new feature.

[05:36:02.0000] <mookid>
no they don;t.. javacsript virtual machines do

[05:36:03.0000] <jcranmer>
mookid: you're adding a feature to HTML

[05:36:04.0000] <mookid>
I'm "adding" something that should already be there

[05:36:05.0000] <jcranmer>
you're still adding

[05:36:06.0000] <mookid>
oh great

[05:36:07.0000] <mookid>
ok well I guess when the slaves all came along asking for freedom

[05:36:08.0000] <mookid>
we should've been like

[05:36:09.0000] <mookid>
you know what

[05:36:10.0000] <mookid>
I'm busy right now

[05:37:00.0000] <mookid>
BYE

[05:37:01.0000] <yecril71>
We shall overcome, some day.

[05:37:02.0000] <mookid>
Why would I want to free you anyway?? I make lots of money from cheap labour.

[05:38:00.0000] <yecril71>
That�s my boy.

[05:38:01.0000] <mookid>
that's a "rational" defense against freeing slaves

[05:38:02.0000] <mookid>
is it not?

[05:38:03.0000] <yecril71>
Of course it is.

[05:38:04.0000] <mookid>
perspective.

[05:38:05.0000] <yecril71>
You are on the ridge of being plonked.

[05:38:06.0000] <mookid>
translate..!

[05:39:00.0000] <yecril71>
killfiled.

[05:39:01.0000] <Dashiva>
HTTP headers do not have human rights

[05:39:02.0000] <Dashiva>
Oddly enough

[05:39:03.0000] <mookid>
really?

[05:39:04.0000] <hsivonen>
mookid: are you implying that using the Accept header is a moral or human right issue?

[05:39:05.0000] <mookid>
no it was an analogy

[05:39:06.0000] <yecril71>
HTTP headers have headerly rights.

[05:40:00.0000] <mookid>
you know.. the kind of thing you use when you're battling ignorance -_-

[05:40:01.0000] <Dashiva>
Analogies are like statistics, 87% of them are made up on the spot.

[05:40:02.0000] <mookid>
yeah it was mostly a joke anyway relax

[05:41:00.0000] <mookid>
so.. about the problems.. what were they again?

[05:41:01.0000] <Philip`>
Dashiva: I won't trust your statistics unless you make up a standard deviation too

[05:41:02.0000] <hsivonen>
mookid: http://krijnhoetmer.nl/irc-logs/whatwg/20081118#l-822

[05:42:00.0000] <mookid>
oh dear oh dear

[05:42:01.0000] <mookid>
you see a URI - and your immediate reaction is that it needs to go through a browser

[05:42:02.0000] <yecril71>
My mediaeval English is not good enough

[05:42:03.0000] <Dashiva>
Philip`: Maybe that's what I want you to think

[05:43:00.0000] <yecril71>
but the weapon is called ep�e in French.

[05:43:01.0000] <mookid>
There is no pair there at all - if I want to get a resource.. I put the URI into the appropriate UA

[05:43:02.0000] <mookid>
there's no pairing there at all

[05:44:00.0000] <yecril71>
OK, it is rapier in English.

[05:45:00.0000] <mookid>
this is my point - you have your subjective opinion on what a URI should do

[05:45:01.0000] <hsivonen>
mookid: and you don't?

[05:45:02.0000] <mookid>
and your projecting it onto developers by restricting the capabilites of browsers

[05:45:03.0000] <mookid>
I don't know which is 'better' - I just know I want to make good use of HTTP

[05:46:00.0000] <mookid>
I can't do that in a brwoser unless you support it

[05:46:01.0000] <mookid>
that's why I suggested it as optional

[05:46:02.0000] <Dashiva>
You want to make the internet worse for everyone else

[05:46:03.0000] <mookid>
fuck off

[05:46:04.0000] <yecril71>
The video tag is not an evidence of hypermedialness.

[05:46:05.0000] <mookid>
excuse my language

[05:46:06.0000] <yecril71>
It is evidence for compositeness.

[05:47:00.0000] <yecril71>
Are you suffering from Turette�s?

[05:47:01.0000] <mookid>
every time this discussion gets interesting it goes quiet..

[05:47:02.0000] <yecril71>
PDF can embed video in exactly the same way.

[05:48:00.0000] <Philip`>
/me sees that Martin McEvoy chooses to interpret the statistics to say that only 0.09% of pages use rev=stylesheet, instead of to say that of the pages that use rev and don't use rev=made (which is redundant with rel=author), 57% use it for rev=stylesheet

[05:48:01.0000] <mookid>
hsivonen: so...

[05:49:00.0000] <mookid>
I don't know exactly what a URI should do.. but I'm sure that restricting client capabilities to use HTTP will lead it down one way when there is another available

[05:49:01.0000] <mookid>
which is why I dont consider "well that's the way its done" an acceptable response

[05:50:00.0000] <mookid>
and neither should you.

[05:51:00.0000] <Dashiva>
Have you produced an answer for "Why should the page author control what my user agent wants to accept on my behalf?" yet?

[05:51:01.0000] <yecril71>
The client is free to use HTTP as it sees fit, within the framework of the HTTP specification.

[05:52:00.0000] <mookid>
Dashiva: what? how is that any different from specifying the content-type in the URI ?

[05:52:01.0000] <Dashiva>
mookid: Because that's just a URI, my browser can still Accept: whatever it wants

[05:52:02.0000] <mookid>
how is saying <a href="foo.pdf">>link</a> not doing that?

[05:52:03.0000] <mookid>
you don't understand Accept then

[05:52:04.0000] <mookid>
Accept is an indicator

[05:53:00.0000] <mookid>
it's not a fixed contract

[05:53:01.0000] <mookid>
the server can return whatever content-type it wants

[05:53:02.0000] <yecril71>
If you specify the content type in the URL, it is there.

[05:53:03.0000] <yecril71>
Otherwise, it is not there.

[05:53:04.0000] <yecril71>
That is the difference.

[05:53:05.0000] <mookid>
er, ok

[05:53:06.0000] <mookid>
:D

[05:53:07.0000] <Dashiva>
mookid: Accept is a header the user agent controls. The URI is something the server controls. They should stick to their own. It's simple, isn't it?

[05:54:00.0000] <mookid>
no.. you misunderstand the purpose of accept

[05:54:01.0000] <mookid>
Accept does nothing at the moment in browsers

[05:54:02.0000] <mookid>
it's just a catchall essentially

[05:54:03.0000] <mookid>
it may aswell not be there

[05:54:04.0000] <yecril71>
Why is it a problem for HTML?

[05:54:05.0000] <Dashiva>
Okay, that shows how much you know

[05:55:00.0000] <mookid>
because HTML is how the browser traverses HTTP

[05:55:01.0000] <yecril71>
The browser does not traverse HTTP at all.

[05:55:02.0000] <mookid>
you know what I mean - dont be pedantic

[05:56:00.0000] <mookid>
traverses URIs over HTTP

[05:56:01.0000] <yecril71>
My guess is that the browser uses HTTP to get data.

[05:56:02.0000] <mookid>
it's the way that brwosers do HTTP

[05:56:03.0000] <yecril71>
It does not traverse URIs.

[05:56:04.0000] <mookid>
yes it does - what is a browser for if it's not for movign around URIs ?

[05:57:00.0000] <yecril71>
The browser is for surfing the Web.

[05:57:01.0000] <yecril71>
(or browsing, if you prefer).

[05:57:02.0000] <mookid>
lay off the crack

[05:57:03.0000] <Dashiva>
mookid: How about you get a packet sniffer and look at what Accept headers browsers actually send

[05:58:00.0000] <yecril71>
(Actually, Microsoft�s documentation nowhere says what Internet Explorer is for).

[05:58:01.0000] <yecril71>
(It only says how cute it is.)

[05:58:02.0000] <mookid>
they send a long string of preferences.. and then a */* at the end..

[05:58:03.0000] <mookid>
so it really makes no differences

[05:58:04.0000] <mookid>
particularly when, by all of your own admission, it doesnt make a difference because each URI should only serve one content-type

[05:59:00.0000] <yecril71>
Why is this a problem for HTML?

[05:59:01.0000] <Dashiva>
Sure, but it lets hardliners like you do it still

[05:59:02.0000] <Dashiva>
So everyone can be happy

[05:59:03.0000] <mookid>
what?

[05:59:04.0000] <mookid>
I'm not 'hard lining' I'm doing the opposite

[06:00:00.0000] <Dashiva>
You're going on and on about a feature we already have

[06:00:01.0000] <mookid>
the future we already have?

[06:00:02.0000] <mookid>
pwahahaha

[06:00:03.0000] <Dashiva>
For no reason but theoretical purity

[06:00:04.0000] <mookid>
theoretical purity?

[06:00:05.0000] <mookid>
your a joke.

[06:00:06.0000] <mookid>
if that's literally the only criticism you have

[06:00:07.0000] <mookid>
you're seriously scraping the bottom of the barrel

[06:02:00.0000] <jcranmer>
Philip`: anyone can use the same stastics to support their case

[06:02:01.0000] <yecril71>
HTTP is a standard protocol for transmitting HTML documents.

[06:02:02.0000] <hallvors>
/me mangles Opera's support for window.setTimeout() if and only if it's called on nytimes.com from the NYTD.require() method.

[06:02:03.0000] <mookid>
you're trolling yecril71  :)

[06:02:04.0000] <yecril71>
Not the other way round.

[06:02:05.0000] <mookid>
HTTP has nothing to do with HTML

[06:02:06.0000] <mookid>
what about

[06:02:07.0000] <yecril71>
I am replying to your letter.

[06:02:08.0000] <mookid>
RSS

[06:03:00.0000] <mookid>
Atom

[06:03:01.0000] <mookid>
xml

[06:03:02.0000] <mookid>
pdf

[06:03:03.0000] <yecril71>
They are document formats, IIRC.

[06:03:04.0000] <hsivonen>
hallvors: how do you undo a patch like that when the site fixes itself?

[06:03:05.0000] <Dashiva>
browserjs, probably

[06:03:06.0000] <hsivonen>
hallvors: or how do site owners test their fix?

[06:03:07.0000] <mookid>
hsivonen: why did you not continue that discussion before?

[06:04:00.0000] <hsivonen>
mookid: I have work to do

[06:04:01.0000] <mookid>
oooh

[06:04:02.0000] <mookid>
i see

[06:04:03.0000] <jcranmer>
Philip`: yet my reaction was the same as yours upon seeing those states

[06:04:04.0000] <jcranmer>
s/te/t/

[06:04:05.0000] <jcranmer>
"rev=stylesheet is by far the most used rev link"

[06:06:00.0000] <yecril71>
HTTP is not limited to HTML, but if you want to expose HTTP, you rather choose HTTP to do the job.

[06:06:01.0000] <yecril71>
If you want to expose HTML, you choose HTTP.

[06:06:02.0000] <Philip`>
jcranmer: No it's not - rev=made is the most used rev link

[06:06:03.0000] <krijnh>
I send all my pages to my customers as email attachments

[06:07:00.0000] <hallvors>
hsivonen: undoing it is as simple as releasing a new browser.js update without that code. Getting site owners to test with opera:config#Browser%20JavaScript set to 0 is somewhat trickier, I admit.. :-o

[06:07:01.0000] <hsivonen>
hallvors: ok

[06:08:00.0000] <jcranmer>
Philip`: must have missed that, then

[06:08:01.0000] <yecril71>
I presume the customers publish krijnh�s pages themselves?

[06:09:00.0000] <krijnh>
No, they email me, "Hi, I want info about your company" and then I send them back an HTML page with that info

[06:09:01.0000] <krijnh>
Fuck HTTP

[06:10:00.0000] <yecril71>
That is rather peculiar, I would say.

[06:11:00.0000] <jcranmer>
yecril71: needless to say, people send HTML email a fair amount

[06:11:01.0000] <yecril71>
Sending is not exposing.

[06:12:00.0000] <yecril71>
All right, in today�s world it is, but it is another story :-(

[06:13:00.0000] <yecril71>
HTML does not support content negotiation at all, in URIs or not in URIs.

[06:14:00.0000] <yecril71>
Except maybe for A[type].

[06:14:01.0000] <yecril71>
User agents can, and perhaps should, support content negotiation.

[06:14:02.0000] <yecril71>
But that has nothing to do with HTML.

[06:15:00.0000] <Philip`>
krijnh: That sounds like Richard Stallman's approach

[06:16:00.0000] <yecril71>
HTML, as a markup language, does not make use of URIs, it only contains them here and there.

[06:16:01.0000] <yecril71>
Indeed, being a language, it does not make use of anything.

[06:18:00.0000] <yecril71>
Actually, SVG tiger is untimately smaller than PNG tiger.

[06:20:00.0000] <yecril71>
ultimately

[06:21:00.0000] <yecril71>
You can have multiple e-mail addresses in rev="made

[06:21:01.0000] <Philip`>
Not if you resize the PNG to be really small

[06:22:00.0000] <yecril71>
SVG is ultimately smaller.

[06:22:01.0000] <Philip`>
yecril71: How do you put multiple email addresses in it?

[06:22:02.0000] <ehird>
but is it AWESOMELY smaller?!

[06:22:03.0000] <Philip`>
yecril71: I bet I can make a 1x1-pixel PNG of the tiger that's smaller than the SVG version

[06:23:00.0000] <yecril71>
mailto:a1⊙ec;a2⊙ec

[06:23:01.0000] <yecril71>
But it is not an equivalent representation any more.

[06:27:00.0000] <Philip`>
Quite a few people just seem to put their names in the href attribute

[06:30:00.0000] <Philip`>
yecril71: If a PNG is ever equivalent to an SVG, where is the threshold at which a resized n*n version of the PNG is no longer equivalent to the SVG?

[06:31:00.0000] <yecril71>
There are various measures.

[06:31:01.0000] <yecril71>
Quantitative: the spectrum of the histogram.

[06:31:02.0000] <yecril71>
Qualitative, if B/W:

[06:32:00.0000] <yecril71>
Graph isomorphism.

[06:32:01.0000] <yecril71>
The threshold depends on how complicated the image is.

[06:33:00.0000] <mookid>
if they represent the same image it's the same resource -_-

[06:33:01.0000] <yecril71>
Fractals cannot be represented as PNG at all.

[06:34:00.0000] <Philip`>
"Lachy[...] knows more about HTML5 than almost anyone else which is why he's the right guy to write the authoring guide." - hmm, that sounds to me like a reason why he shouldn't write it, because he won't understand HTML5 from the perspective of a typical ignorant developer :-)

[06:35:00.0000] <Philip`>
mookid: If I have a PNG image and a thumbnail version of it, those are two representations of the same resource?

[06:36:00.0000] <Philip`>
If so, does HTTP provide any way to request the desired representation from the same URI?

[06:36:01.0000] <Lachy>
Philip`, where's that quote from?

[06:36:02.0000] <Lachy>
/me was an ignorant developer once, so I'm more than qualified :-)

[06:37:00.0000] <Philip`>
Lachy: http://www.w3.org/mid/49241573.8080003⊙don

[06:37:01.0000] <Lachy>
ok

[06:37:02.0000] <yecril71>
The content type can contain image resolution as an option

[06:39:00.0000] <Philip`>
Ah, so I can do "Accept: image/png;size=little;q=1"? I suppose that sounds vaguely reasonable

[06:40:00.0000] <yecril71>
I cannot give you the details, check it up with the registry.

[06:40:01.0000] <yecril71>
The resolution should be explicit IIRC.

[06:43:00.0000] <mookid>
Philip`: a thumbnail is a seperate resource right

[06:44:00.0000] <mookid>
example.com/dog/thumbnail

[06:45:00.0000] <Philip`>
mookid: Hmm, I don't quite see what makes it a separate resource, since it seems to be representing the same image

[06:45:01.0000] <mookid>
it's a seperate image - it's a thumbnail of the real image

[06:46:00.0000] <Philip`>
But a JPEG and a PNG are different representations of the same image (if and only if they decode to the same number of pixels)?

[06:47:00.0000] <Philip`>
/me isn't sure where the distinction lies

[06:47:01.0000] <mookid>
why would they be a different size?

[06:47:02.0000] <mookid>
well the distinction lies in what the entity is..

[06:47:03.0000] <jcranmer>
but JPEG is lossy and PNG is lossless

[06:47:04.0000] <mookid>
a thumbnail isn't the image.. that's why it's a thumbnail of the image..

[06:48:00.0000] <mookid>
yeah - they're different representation of the same image

[06:48:01.0000] <Philip`>
mookid: One might be a low-resolution lossy JPEG, and the other a high-resolution lossless PNG

[06:49:00.0000] <mookid>
what's your point?

[06:49:01.0000] <Philip`>
(so people might decide which to download based on the speed of their internet connection, because otherwise they're basically the same image)

[06:49:02.0000] <mookid>
..?

[06:51:00.0000] <mookid>
Sorry, I dont understand the point you're trying to make dude

[06:51:01.0000] <Philip`>
mookid: My point is to try to understand why if I have a low-resolution lossy JPEG (which could act as a thumbnail) and a high-resolution lossless PNG, you seem to be saying they are different resources; but if I have equally-sized JPEG and PNGs, they're just different representations of the same resource

[06:52:00.0000] <mookid>
That's a design descision obviously

[06:52:01.0000] <mookid>
there's nothign stopping you from resizing the image

[06:52:02.0000] <Philip`>
I'm not attempting to prove anything, I just want to get a better understanding of what a resource is, since it seems a bit vague at the moment

[06:52:03.0000] <mookid>
and specifying Accept jpeg

[06:52:04.0000] <mookid>
for your thumbnails..

[06:52:05.0000] <mookid>
that's application design though right?!

[06:53:00.0000] <mookid>
well resource is supposed to be vague - that's the poitn I'm trying to get across - you need to be able to accuotn for everyone's interpretation

[06:54:00.0000] <mookid>
some people think resources are tied to a singl representation for each URI

[06:54:01.0000] <mookid>
which is fine.. but some people don't and that is the reason Accept exists

[06:54:02.0000] <Philip`>
/me doesn't like things being vague :-(

[06:54:03.0000] <mookid>
vague is the wrong word..

[06:55:00.0000] <mookid>
abstract is better

[06:55:01.0000] <Philip`>
I wouldn't mind if people had different specific technical interpretations of what a "resource" is, as long as they could each define it, and then we could give them all different names to make it clearer and then it wouldn't be confusing

[06:56:00.0000] <mookid>
it is defined in fielding's disseratation - there's plenty ofpapers written abuot it

[06:58:00.0000] <yecril71>
Web authors are not going to read that.

[06:59:00.0000] <yecril71>
You have to come up with something short to recount and easy to grasp.

[06:59:01.0000] <mookid>
do yuo have to say things?

[06:59:02.0000] <mookid>
why dont you not say things..

[06:59:03.0000] <mookid>
try that out

[06:59:04.0000] <Philip`>
"a resource R is a temporally varying membership function M_R(t), which for time t maps to a set of entities, or values, which are equivalent." clears everything up ;-)

[07:00:00.0000] <mookid>
:P

[07:00:01.0000] <mookid>
that's explained better in th enext paragrpah by way of explaining the benefits

[07:01:00.0000] <Philip`>
/me wonders why M_R is defined and then seemingly never referred to again

[07:02:00.0000] <Philip`>
Maybe my concern is that it seems inadequate to use media type as the way to select a representation, since the media type is only meant to be identifying the data format of a representation and you might have multiple distinct representations of the same resource with the same data format

[07:03:00.0000] <mookid>
how would that work?

[07:03:01.0000] <mookid>
wtf was that? o.O

[07:04:00.0000] <mookid>
how would that work?

[07:05:00.0000] <Philip`>
I could represent my hypertext document as text/html with uppercase tag names, or as text/html with lowercase tag names, and they're distinct representations (since a representation is a sequence of bytes) with the same data format

[07:05:01.0000] <mookid>
page/uppercase + page/lowercase

[07:05:02.0000] <mookid>
seperate formats

[07:05:03.0000] <mookid>
formats

[07:05:04.0000] <mookid>
lol

[07:05:05.0000] <mookid>
seperate resoruces

[07:05:06.0000] <mookid>
not formats

[07:06:00.0000] <yecril71>
page/uppercase is not registered

[07:06:01.0000] <Philip`>
But they're the same resource, just with different representations, and it's not nice if I have to pretend they're different resources to hack around HTTP's assumption that URI+media-type is adequate identification

[07:07:00.0000] <Philip`>
yecril71: I think page/uppercase is intended as a URI, not a media type

[07:07:01.0000] <yecril71>
you would have to register an option for text/html

[07:08:00.0000] <Philip`>
That sounds like an awful lot of work, and also impossible

[07:08:01.0000] <yecril71>
but what if you want every other letter to be capitalized?

[07:08:02.0000] <Philip`>
because the IETF is not going to listen to me saying "please register text/html;uppercase=yeah"

[07:08:03.0000] <yecril71>
are x-options allowed?

[07:08:04.0000] <yecril71>
It depends whether they deem it is important.

[07:09:00.0000] <yecril71>
In this particular case, it probably is not.

[07:09:01.0000] <yecril71>
And I would rather say "flavor=uppercase".

[07:11:00.0000] <Philip`>
The media type is the wrong place for that information anyway, since it's not part of the data format - the data format is still just HTML

[07:12:00.0000] <yecril71>
I can imagine it can be essential for somebody to have it like that.

[07:13:00.0000] <yecril71>
As a more real-world example, consider "linewidth=80" versus "indented=pretty".

[07:13:01.0000] <Philip`>
The aforementioned dissertation doesn't seem to say anything about how you only content-negotiate based on the data format, so I guess if you're using HTTP you should do "X-Accept-Uppercase-Tags: please" or something

[07:13:02.0000] <mookid>
are they the same resrouce though - it would have a different include for the css

[07:14:00.0000] <Philip`>
in which case simply adding "Accept" to HTML wouldn't be sufficient to do proper content negotiation, since that's restricted to negotiating the data format :-)

[07:15:00.0000] <yecril71>
I think all those things belong to the data format as options.

[07:15:01.0000] <mookid>
right - it's GUI stuff

[07:15:02.0000] <mookid>
you can do that with css and javascript

[07:15:03.0000] <mookid>
that's design stuff again

[07:17:00.0000] <yecril71>
Do what with css and javascript?

[07:17:01.0000] <yecril71>
I can do everything with JavaScript because it is Turing-complete.

[07:17:02.0000] <yecril71>
And I am the king of gorillas.

[07:18:00.0000] <mookid>
this whole uppercase of the same page stuf that's apparently created a massive flaw in RESTful architecture

[07:18:01.0000] <mookid>
-_-

[07:19:00.0000] <mookid>
either way - some data is different between the two resources whether it's the css include or the ASCII values

[07:20:00.0000] <mookid>
how you approach that is a design descision - I shouldn't be the person to say one way is better than the other

[07:20:01.0000] <mookid>
the same way as I shouldn't be the person to say that using URIs is better one way or the other

[07:21:00.0000] <mookid>
but neither should anyone else - we just need to support the standarsd (like HTTP) and let developers mak those descisions

[07:22:00.0000] <mookid>
that's the most pragmatic approach

[07:22:01.0000] <mookid>
Philip`: did some of the latest email make more sense - I only ask because you seem more intereted in my opinion now :p

[07:24:00.0000] <yecril71>
I am sorry to say most of id did not make sense.

[07:24:01.0000] <yecril71>
most of it

[07:25:00.0000] <mookid>
feel free to respond on tehmailing list

[07:25:01.0000] <mookid>
:)

[07:25:02.0000] <yecril71>
The problem is, I am sort of banned.

[07:25:03.0000] <mookid>
I can understand why

[07:26:00.0000] <yecril71>
Good for you.

[07:26:01.0000] <mookid>
yes it is.

[07:27:00.0000] <Philip`>
mookid: The ASCII values are an artifact of the representation, not of the underlying conceptual resource - if I write <html> or <HTML> then it gets parsed by exactly the same parsing algorithm into exactly the same DOM, and the only difference is the bytes sent over the wire

[07:28:00.0000] <Philip`>
(I'm not uppercasing the textual content of the document, only the tag names)

[07:29:00.0000] <Philip`>
(which is a slightly pointless thing to do, but I'm sure there exist more plausible real-world cases for something like this :-p )

[07:29:01.0000] <yecril71>
Like linewidth=80 for example.

[07:30:00.0000] <Philip`>
mookid: I don't remember the latest email saying much that hadn't been said before; I'm just procrastinating by thinking out loud about some of this stuff :-)

[07:32:00.0000] <Philip`>
Also it's comforting to know that a PhD dissertation doesn't have to actually make sense

[07:34:00.0000] <jmb>
suits me :)

[07:34:01.0000] <yecril71>
Why is it comforting?

[07:34:02.0000] <jcranmer>
who is it that wrote that nonsensical PhD

[07:34:03.0000] <jcranmer>
?

[07:35:00.0000] <Philip`>
jcranmer: Do you mean Roy Fielding's dissertation?

[07:35:01.0000] <yecril71>
It is rather depressing to me.

[07:35:02.0000] <Philip`>
jcranmer: If so, Roy Fielding wrote it

[07:36:00.0000] <Philip`>
yecril71: Because I need to write one, and so I'm happier if I think the bar is lowered :-)

[07:36:01.0000] <yecril71>
It is depressing because I know I am not shameless enough to follow that path :-(

[07:37:00.0000] <Philip`>
/me has been working for a year and still doesn't quite know what his dissertation is going to be on, but never mind minor problems like that

[07:38:00.0000] <mookid>
well I'd be interested to hear these real world cases - and also how that relates to content type negotiation

[07:39:00.0000] <jcranmer>
ah

[07:40:00.0000] <jcranmer>
Bogdanov

[07:40:01.0000] <jmb>
Philip`: if it helps, it took me somewhere between 18 & 21 months to work out exactly what I was doing

[07:41:00.0000] <Philip`>
/me notes that he is being mostly unfair in thinking that Roy Fielding's dissertation doesn't make sense, since he hasn't read it in much detail, but it does seem mostly like vague descriptions and post-hoc rationalisations, which isn't great, though it's kind of unavoidable because you can't set up two wildly popular world-wide network systems and evaluate the differences so you just have to guess

[07:42:00.0000] <yecril71>
/me has been working for half a year to learn that the unjustified assumptions his master�s thesis was based on actually can be defended

[07:44:00.0000] <jgraham>
/me moans about people discussing PhDs here

[07:46:00.0000] <Philip`>
mookid: I think the not-quite-real-world cases still demonstrate that there is seemingly a disconnect between the abstract REST theory (where a resource has multiple representations and clients can negotiate which one they get) and practical suggestions for adding <a href accept="media-type"> (which is much more limited)

[07:47:00.0000] <Philip`>
so a good solution to the content negotiation issue should do much more than simply Accept

[07:48:00.0000] <Philip`>
and would have to allow arbitrary X-... HTTP headers, because that's the only mechanism that lets you properly implement content negotiation

[07:50:00.0000] <Philip`>
Anyway, real PhD dissertations should just be full of maths rather than woolly words

[07:52:00.0000] <jgraham>
/me suspects that depends somewhat on your thesis topic

[07:52:01.0000] <Philip`>
/me is actually trying to avoid the maths as much as possible, because it's scary

[07:52:02.0000] <mookid>
Philip`: The accept header would be one component of that

[07:53:00.0000] <mookid>
why would it not make sense to at least provide the mechanism for that even if it is 'incomplete'

[07:54:00.0000] <mookid>
beyond that you are getting into additions to the HTTP spec

[07:54:01.0000] <Philip`>
How come nobody has commented yet on the <link rev="D. Bailey Management" ...> and <link rev="YouTube" ...> and <link rev="text/css" ...> and so on? Those are far more fun than simply mixing rev and rel

[07:55:00.0000] <Philip`>
mookid: HTTP already explicitly allows extension header fields for server-driven content negotiation

[07:56:00.0000] <Philip`>
so it wouldn't need additions

[07:56:01.0000] <mookid>
right - but that's for use in special cases

[07:56:02.0000] <mookid>
Accept is a standard

[07:56:03.0000] <mookid>
specific standard

[07:56:04.0000] <mookid>
I suppose you could go as far as to provide a header attribute instead

[07:57:00.0000] <Philip`>
There are four other specific standard non-extension header fields that it mentions for negotiation

[07:57:01.0000] <mookid>
content-encoding?

[07:57:02.0000] <mookid>
user-agent ?

[07:58:00.0000] <Philip`>
Accept-{Charset,Encoding,Language}, User-Agent

[07:58:01.0000] <mookid>
user-agent I'd say is pretty irrelevant for obvious reasons

[07:59:00.0000] <mookid>
but sure these could also be accounted for

[07:59:01.0000] <Philip`>
If the problem is that you want to expose control over HTTP content negotiation to HTML pages, then Accept by itself is not a solution to that problem

[07:59:02.0000] <mookid>
well it's the most useful

[07:59:03.0000] <mookid>
by far

[08:00:00.0000] <mookid>
multiple content-types is the most common variation on resource representation

[08:00:01.0000] <mookid>
I take your point though, maybe it makes sense to ammend my proposal to include all of these

[08:01:00.0000] <mookid>
or even just a header attribute

[08:01:01.0000] <mookid>
this is more like it, though  thanks for your input :)

[08:02:00.0000] <Philip`>
/me is just trying to make the proposed feature more and more complex, so it's easier to kill it later for being too bloated ;-)

[08:02:01.0000] <mookid>
haha I did suspect ;P

[08:02:02.0000] <mookid>
is there no way we can work at this to make it feasible?

[08:03:00.0000] <mookid>
who can I work through the nitty-gritty with ?

[08:04:00.0000] <mookid>
I don't think it has to be that bloated necessarily, HTTP isn't a really bloated protocol

[08:04:01.0000] <mookid>
it's certianly no SOAP.

[08:08:00.0000] <hsivonen>
http://lists.w3.org/Archives/Public/www-html/2006Feb/0069.html

[08:08:01.0000] <hsivonen>
last paragraph in particular

[08:09:00.0000] <hsivonen>
and http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2007-March/009732.html

[08:10:00.0000] <jgraham>
mookid: Are you still looking for a way to specify http accept on a link? It seems to me that that, apart from any theoretical issues has significant compatibility issues since it would cause new browsers to return different documents to old browsers

[08:11:00.0000] <Philip`>
mookid: Given that you don't seem too concerned about supporting e.g. Accept-Charset, I assume you agree that "it is a feature of HTTP, therefore it must be exposed via HTML" is not a convincing argument, because sometimes nobody cares enough about that feature of HTTP to be worth the effort

[08:11:01.0000] <mookid>
but you're assuming nobody cares becasue noone has the ability to

[08:11:02.0000] <mookid>
I addressed this

[08:12:00.0000] <mookid>
that's an unwinnable argument purely on the basis that it's not practical *at the moment*

[08:12:01.0000] <mookid>
that's the reason I've brought this up

[08:12:02.0000] <mookid>
to change it so it is feasible..

[08:12:03.0000] <Philip`>
mookid: But you seem to care about Accept and not about Accept-Charset, even though you can't control either via HTML

[08:12:04.0000] <mookid>
in a way which doesn't effect existing implementations

[08:12:05.0000] <mookid>
right and I've said I'm happy to loko at including that aswell

[08:12:06.0000] <mookid>
content-type is by far the most common cocnern

[08:13:00.0000] <jgraham>
mookid: Is this directed at me? I don't understand how one could implement the feature so that it didn't break existing implementations

[08:13:01.0000] <mookid>
if that's going to make the difference then I am more than happy to include them in my proposal

[08:13:02.0000] <hsivonen>
mookid: here's the thing is short: If you care about which representation you are linking to, the choice of representation isn't indifferent to you and you need to address a particular representation. We already have an addressing mechanism: URL.

[08:13:03.0000] <mookid>
jgraham: becaue it's an option attribute

[08:13:04.0000] <jgraham>
fwiw I also agree with hsivonen 

[08:13:05.0000] <hsivonen>
mookid: so whenever you care about linking to a particular representation, you really have two resources

[08:13:06.0000] <mookid>
THATS ONE APPROACH

[08:14:00.0000] <mookid>
I know

[08:14:01.0000] <mookid>
you can use

[08:14:02.0000] <mookid>
URLs

[08:14:03.0000] <mookid>
to do that

[08:14:04.0000] <Philip`>
mookid: So when looking at something like Accept-Charset, we should consider the possible costs (of language complexity and implementation effort and everything else) and the possible benefits (like what it would enable people to do) in order to decide whether it's worthwhile?

[08:14:05.0000] <hsivonen>
mookid: if you had only one resource, you by definition wouldn't have to care about linking to a particular representation

[08:14:06.0000] <mookid>
what?

[08:14:07.0000] <mookid>
what is this 'link' you talk abuot

[08:14:08.0000] <mookid>
it's a request

[08:14:09.0000] <hsivonen>
(aside: it may follow that with this definition, the set of resources with multiple representations is the empty set. If so, so be it.)

[08:14:10.0000] <mookid>
and a request is more than just GET (URI)

[08:15:00.0000] <mookid>
thre are headers

[08:15:01.0000] <mookid>
for a reason

[08:15:02.0000] <jgraham>
mookid: ? So how does that change the fact that if I do <a href="foo" accept="application/pdf"> a new browser will return a PDF and an old one will not?

[08:15:03.0000] <mookid>
in my application

[08:15:04.0000] <mookid>
that's a design descision

[08:15:05.0000] <mookid>
old applications will continue to work fine

[08:15:06.0000] <mookid>
with new browsers

[08:15:07.0000] <hsivonen>
mookid: link is the <a> element with your proposed attribute

[08:15:08.0000] <mookid>
jgraham: your response to that is..?

[08:16:00.0000] <jgraham>
You mean you are happy if different users get different representations of the resource?

[08:16:01.0000] <mookid>
yes - they're just representations

[08:16:02.0000] <jgraham>
Are your users also happy with that?

[08:16:03.0000] <jgraham>
I guess not

[08:16:04.0000] <mookid>
of course I am

[08:16:05.0000] <mookid>
it's a diffrent way of using URIs

[08:16:06.0000] <mookid>
which you clearly dont *think* is the 'right' way

[08:16:07.0000] <hsivonen>
mookid: well then you don't need to specify the accept header in your <a> element!

[08:17:00.0000] <mookid>
it's not 'right' it's just a way

[08:17:01.0000] <mookid>
hsivonen: yeah you do

[08:17:02.0000] <mookid>
of course you do

[08:17:03.0000] <jgraham>
I would be extremely confused as an end user if clicking the same link consistently returned an HTML document in one browser and a PDF in another

[08:17:04.0000] <hsivonen>
mookid: you just said you're happy even without it!

[08:17:05.0000] <mookid>
It's an application design issue..

[08:17:06.0000] <jgraham>
Indeed I would assume that the browser that returned the PDF was broken

[08:18:00.0000] <mookid>
I'm happy that a user would have to either know the UA that was most appropriate

[08:18:01.0000] <mookid>
and/or the html version of that resource could link their up to date browser

[08:18:02.0000] <mookid>
to all the representations

[08:18:03.0000] <mookid>
aswell as provide it's own representation

[08:18:04.0000] <mookid>
that's an application design descision

[08:18:05.0000] <mookid>
microsoft office talks HTTP now

[08:19:00.0000] <Philip`>
jgraham: I think you'd actually blame Adobe for having a broken Acrobat plugin in that case

[08:19:01.0000] <Philip`>
because they're the easiest target for any PDF-related complaints

[08:19:02.0000] <jgraham>
I am not a design expert but having a single UI element do different things based on some unspecified external variable is, I expect, never a good idea

[08:19:03.0000] <hsivonen>
/me starts to think "that's an application design descision" is some kind of architect cop-out

[08:20:00.0000] <mookid>
it's not a single UI element..

[08:20:01.0000] <jgraham>
It's a single link

[08:20:02.0000] <mookid>
how many users use multiple browsers?

[08:20:03.0000] <jgraham>
Almost all, hopefully

[08:21:00.0000] <mookid>
If HTML5 did implement that - up to date browsers would understand the markup so that's not a fair point

[08:21:01.0000] <mookid>
what?

[08:21:02.0000] <jgraham>
Browsers change with time

[08:21:03.0000] <jgraham>
Few people still use NS4

[08:21:04.0000] <mookid>
yeah.. that's why I'd want it in the HTTP5 spec

[08:21:05.0000] <Philip`>
It's more of a concern when multiple users, each with a different single browser, communicate and expect web sites to work the same for everyone

[08:21:06.0000] <mookid>
well up to date browsers taht were HTTP5 compliant would understand the marup for the accept header

[08:21:07.0000] <mookid>
so I'd be willing to accept that older browsers wouldn't work and insist that my users upgrade

[08:22:00.0000] <mookid>
that's my descision as a developer

[08:22:01.0000] <Philip`>
(Do you mean HTML5, not HTTP5?)

[08:22:02.0000] <mookid>
HTLM

[08:22:03.0000] <mookid>
yeah

[08:22:04.0000] <mookid>
my bad ;)

[08:22:05.0000] <mookid>
what's wrong with that approach - it's just a different way of using URIs I dont see what the big deal is

[08:22:06.0000] <mookid>
and it's not like it stops you from continuing to use URIs for conneg

[08:23:00.0000] <jgraham>
Yeah so a user used to one behaviour in an old browser would be surprised if their new browser did something different with a familiar application. This would possibly be enough to convince them that the new browser was broken and that they should not upgrade

[08:23:01.0000] <mookid>
what

[08:23:02.0000] <hsivonen>
mookid: over here we don't just make stuff up for choice and vive la difference

[08:23:03.0000] <mookid>
if a link that says 'pdf version of reoprt' returns a pdf?

[08:23:04.0000] <mookid>
if you don't indicate that a certain link does something

[08:23:05.0000] <mookid>
that's a GUI insufficiecy

[08:24:00.0000] <mookid>
hsivonen:  I'm not making anything up - HTTP conneg exists. deal with it.

[08:24:01.0000] <mookid>
hsivonen: you're opinionated and you're shrouding it in psuedo-logic

[08:24:02.0000] <mookid>
just admit you want conneg in URIs

[08:24:03.0000] <hsivonen>
mookid: I'm dealing with the consequences right now

[08:24:04.0000] <jgraham>
I really don't understand this. You seem to be imaganing that there is actually a single canonical version of the resource e.g. a PDF

[08:25:00.0000] <jgraham>
If that is the case, just link to is

[08:25:01.0000] <jgraham>
s/is/it/

[08:25:02.0000] <mookid>
what?

[08:25:03.0000] <mookid>
that doesnt make sense

[08:25:04.0000] <mookid>
I'm suggesting there's lots of representations

[08:25:05.0000] <mookid>
of a resource

[08:25:06.0000] <mookid>
one URI has many content types to return

[08:26:00.0000] <mookid>
and we need a way of letting browsers negotiate that URI on the fly

[08:26:01.0000] <mookid>
a PDF reader, rightly so, already has it's Accept headers fixed

[08:26:02.0000] <jgraham>
So how can you write "pdf version of report" if it might return HTML or plain text or Word?

[08:26:03.0000] <mookid>
because that's all it accepts

[08:26:04.0000] <mookid>
it's just data

[08:26:05.0000] <hsivonen>
http://thedailywtf.com/Articles/The_Complicator_0x27_s_Gloves.aspx

[08:26:06.0000] <mookid>
fick.

[08:26:07.0000] <mookid>
dick*

[08:27:00.0000] <mookid>
passive aggresive is what defensive peopel with no rationale do

[08:27:01.0000] <jgraham>
Heh

[08:27:02.0000] <mookid>
grow a pair hsivonen 

[08:27:03.0000] <jgraham>
/me has met hsivonen

[08:28:00.0000] <jgraham>
That is not a good description :)

[08:28:01.0000] <Philip`>
/me thinks the 0x27 in the URL is a bit weird

[08:28:02.0000] <mookid>
jgraham: you look at web applications differenet to me

[08:28:03.0000] <mookid>
I see a pdf as the same thing as an html document - just a different format

[08:29:00.0000] <jgraham>
mookid: I think mos users see them as different, even if they have the same prose content

[08:29:01.0000] <mookid>
so if I PUT the pdf resource back to the URI it's the same as if I PUT the html or the word doc format back

[08:29:02.0000] <mookid>
it changes the resource

[08:29:03.0000] <mookid>
users dont care

[08:29:04.0000] <mookid>
they do what they're told

[08:29:05.0000] <mookid>
particularly with html

[08:29:06.0000] <mookid>
they dont look at the markup

[08:29:07.0000] <mookid>
or the status bar

[08:29:08.0000] <Philip`>
Evidence suggests people don't actually do what they're told :-)

[08:29:09.0000] <mookid>
no - they like to do it their way

[08:30:00.0000] <takkaria>
well, no.  pdfs load adobe acrobat into the browser and make it slow and draggy, and give it an extra set of menu bars.  html doesn't.  most people understand the stuff you do in web browsers != PDF

[08:30:01.0000] <mookid>
if you couple content-type to a URI you restrict them

[08:30:02.0000] <jgraham>
mookid: I don't claim that they care about the markup or about the status bar. They care about the results of their actions though

[08:30:03.0000] <yecril71>
PUT is not enough to make a resource fork.

[08:30:04.0000] <mookid>
how will they notice any differnet if the conneg is in the URI or whether it's HTTP then?

[08:30:05.0000] <yecril71>
Some server magic is needed.

[08:31:00.0000] <mookid>
like JAX-RS ? -_-

[08:31:01.0000] <yecril71>
Like vi .htaccess

[08:31:02.0000] <mookid>
it's not magic that's how modern web apps are being designed

[08:31:03.0000] <jgraham>
I think the core concept of content negoiation is user-hostile

[08:31:04.0000] <mookid>
htacces..

[08:31:05.0000] <yecril71>
whatever

[08:31:06.0000] <mookid>
is that a joke?

[08:31:07.0000] <jgraham>
Or, at least there are very few occasions where I think it is reasonable

[08:31:08.0000] <yecril71>
How do you configure Apache to do content negotiation?

[08:32:00.0000] <mookid>
hahaha

[08:32:01.0000] <mookid>
that says it all

[08:32:02.0000] <jgraham>
e.g. serving XHTML vs HTML is often OK

[08:32:03.0000] <jgraham>
Serving PDF or HTML is not

[08:32:04.0000] <mookid>
do any of you actually develop enterprise applications?

[08:32:05.0000] <mookid>
-_-

[08:32:06.0000] <jgraham>
because users interact with those representations very differently

[08:32:07.0000] <takkaria>
/me concludes the Language spec thread on public-html is now entirely bikeshedding

[08:33:00.0000] <jgraham>
mookid: Fortunatley not :)

[08:33:01.0000] <hsivonen>
/me used to develop enterprise apps but doesn't anymore

[08:33:02.0000] <mookid>
maybe that's why you dont see this from my perspective

[08:33:03.0000] <Philip`>
yecril71: It's not a thing for static content that's all handled by Apache; it's for complex dynamic systems written in an Enterprise language like Visual Basic

[08:33:04.0000] <mookid>
you're seeing html and PDF as seperate

[08:33:05.0000] <smedero>
/me likes his bike shed painted purple

[08:33:06.0000] <jgraham>
/me noticed that after three days away public-html had almost entirely meta discussion and whatwg almost entirely technical discussion

[08:33:07.0000] <takkaria>
how has enterprise programming got anything to do with how users see html vs pdf?

[08:33:08.0000] <mookid>
I see HTML and PDF documents at one URI beign generated from the same data store < IMPORTANT

[08:34:00.0000] <mookid>
and updated to the same store aswell

[08:34:01.0000] <mookid>
generated.. on the fly..

[08:34:02.0000] <yecril71>
Really?  So even Apache does not support declarative content negotiation?

[08:34:03.0000] <jgraham>
mookid: Right but do users care about which representation they get? 

[08:35:00.0000] <mookid>
well if they do they'll use the UA that's appropriate

[08:35:01.0000] <yecril71>
Then we should probably wait until it does.

[08:35:02.0000] <mookid>
I can write a content negotating app in JAX-RS right now

[08:35:03.0000] <mookid>
I just can't get a brwoser to use it properly

[08:35:04.0000] <jgraham>
mookid: So you expect me to load my PDF viewer to get a PDF version of a resource?

[08:36:00.0000] <yecril71>
If you cannot do it without writing server-side code, you cannot require it.

[08:36:01.0000] <mookid>
yecril71: you dont know what you're tlaking about

[08:36:02.0000] <Philip`>
yecril71: You can just use mod_rewrite if you want to redirect using .htaccess depending on the request headers

[08:36:03.0000] <mookid>
jgraham: yes, I dont see the big deal with that

[08:37:00.0000] <yecril71>
Right.  However, I cannot do it with PUT.

[08:37:01.0000] <yecril71>
I have to use another server-side technique to create the fork.

[08:37:02.0000] <jgraham>
mookid: Ah, I think we are getting closer to the fundamental disagreement here

[08:37:03.0000] <mookid>
yes!

[08:37:04.0000] <mookid>
It's design level

[08:37:05.0000] <Philip`>
yecril71: Not to static files via Apache, which is why this is more relevant for more dynamic systems, where you're updating databases and generating content from them

[08:38:00.0000] <mookid>
YES

[08:38:01.0000] <mookid>
thank you Philip` 

[08:38:02.0000] <yecril71>
But mookid said he can PUT.

[08:38:03.0000] <jgraham>
My PDF viewer doesn't even seem to open URIs

[08:38:04.0000] <yecril71>
Hie cannot.

[08:38:05.0000] <Philip`>
I don't think I have any icons on my computer to launch a PDF viewer :-(

[08:38:06.0000] <mookid>
well then update your PDF viewer then

[08:39:00.0000] <jgraham>
I believe I have the latest avaliable version

[08:39:01.0000] <mookid>
or browse to my html pages which will give you the HTML version and link yuo to the other representations using the brand new Accept attribute

[08:39:02.0000] <mookid>
the fact that the UA's dont support it yet doesnt mean they wont

[08:39:03.0000] <mookid>
it's all moving that way

[08:39:04.0000] <mookid>
I should know..

[08:39:05.0000] <mookid>
o.O

[08:40:00.0000] <mookid>
MS office does HTTP

[08:40:01.0000] <jgraham>
mookid: The point that I was trying to make earlier is that UAs not supporting it today may wel mean that they won't becauseany browser that adds support will change the behaviour of existing applications, so discouraging users from upgrading

[08:41:00.0000] <Philip`>
jgraham: You should get a better operating system / desktop environment - KDE automatically lets you load URLs into programs like KPDF :-)

[08:41:01.0000] <mookid>
you moved off 3.5 Philip`  ?

[08:41:02.0000] <Philip`>
mookid: No

[08:41:03.0000] <mookid>
me neither :>

[08:41:04.0000] <Philip`>
(Well, I have 4.1 installed but only use it when testing Konqueror)

[08:41:05.0000] <yecril71>
And a HTTP-compatible file system like HFS that supports forks

[08:42:00.0000] <yecril71>
HFS+ (HFS supports only two)

[08:42:01.0000] <Philip`>
yecril71: The filesystem is irrelevant - you write code that sits behind the URLs to do whatever magic it wants to do

[08:43:00.0000] <yecril71>
If I have to write code, I defect.

[08:43:01.0000] <yecril71>
I need a declarative interface for maintenance.

[08:44:00.0000] <ehird>
uh-oh mookid is still blabbing on about that conneg?

[08:44:01.0000] <ehird>
:P

[08:44:02.0000] <mookid>
jgraham: I dont  understand the point your making, explain please

[08:44:03.0000] <mookid>
:)

[08:44:04.0000] <mookid>
it's important stuff

[08:45:00.0000] <Philip`>
yecril71: If you don't want to write code, this feature is not for you :-)

[08:45:01.0000] <mookid>
jgraham: you see incompatability between the two methods that isn't there

[08:46:00.0000] <mookid>
UAs that are not browsers should be setting their Accept headers appropriately

[08:46:01.0000] <mookid>
so if you specify the conneg in the URI or you don't it shouldn't matter

[08:46:02.0000] <yecril71>
If it is not supported declaratively server-side, it is not mature enough.

[08:46:03.0000] <mookid>
yecril71: yes it is

[08:47:00.0000] <mookid>
becasue you dont' know how to do it doesnt mean it's not possible

[08:47:01.0000] <yecril71>
I never said it is not possible.

[08:47:02.0000] <mookid>
well then stop talking.

[08:47:03.0000] <mookid>
'not supported' = not possible

[08:47:04.0000] <yecril71>
I said it is immature.

[08:47:05.0000] <mookid>
immature

[08:47:06.0000] <mookid>
just stop.

[08:47:07.0000] <mookid>
stop please.

[08:48:00.0000] <mookid>
when in hole and all that

[08:48:01.0000] <mookid>
why did you say you got banned again?

[08:48:02.0000] <mookid>
=)

[08:48:03.0000] <yecril71>
I did not say I got banned again.

[08:49:00.0000] <mookid>
turn of phrase

[08:49:01.0000] <yecril71>
I just got banned, indefinitely.

[08:49:02.0000] <mookid>
I'm tempted to put you on ignore

[08:49:03.0000] <mookid>
I can't though I'm a liberal

[08:49:04.0000] <mookid>
damn my awesomeness

[08:49:05.0000] <yecril71>
Sweet temptations...

[08:49:06.0000] <takkaria>
putting people on ignore is something you do, not something you talk about. :)

[08:50:00.0000] <jgraham>
mookid: In browser A I click a link and get a HTML version of a resource with a link to a PDF version. In newer browser B that supports <a accept=> I get a PDF version. I depended on the HTML version. From my point of view browser B is broken. I stick with browser A.

[08:50:01.0000] <jgraham>
Hence browser B cannot gain marketshare without dropping support for <a accept>

[08:50:02.0000] <mookid>
right.. that's technical innovation

[08:50:03.0000] <mookid>
that happens all the time

[08:51:00.0000] <mookid>
so blu-ray should never have happened

[08:51:01.0000] <mookid>
because the discs dont work in my dvd player?

[08:51:02.0000] <yecril71>
Blue ray disks hold more data.

[08:52:00.0000] <yecril71>
That is an advantage.

[08:52:01.0000] <yecril71>
But there are innovations that gain acceptance, and there are others that do not.

[08:52:02.0000] <jgraham>
It's more like playing DVDs in your bluray players and discovering that they suddenly play the remake of psyco rather than the hitchcock original

[08:53:00.0000] <mookid>
not really those are two seperate resources

[08:53:01.0000] <mookid>
actually it's more like putting a blue ray disc in a dvd player and getting dvd quality

[08:53:02.0000] <mookid>
dissapointing but hey.. you're an idiot

[08:53:03.0000] <mookid>
:)

[08:54:00.0000] <yecril71>
I have always thought I am an idiot.

[08:54:01.0000] <jgraham>
They have the same script, same shots, just different versions of them

[08:54:02.0000] <yecril71>
So now there are two of us...

[08:54:03.0000] <yecril71>
:)

[08:56:00.0000] <mookid>
I wasnt actually calling him an idiot

[08:57:00.0000] <mookid>
I was calling the hypothetical person putting a blu ray disc in a dvd player an idiot

[09:00:00.0000] <mookid>
jgraham: yeah exactly - so you put the 'blu ray disc' into your cd player and it just plays the audio

[09:01:00.0000] <mookid>
with someone describing the scene aswell

[09:01:01.0000] <mookid>
so your cd player representation doesnt lose to much data

[09:02:00.0000] <takkaria>
no, but you still want to be able to differentiate between the cd/dvd/blu-ray versions

[09:02:01.0000] <mookid>
you want a seperate disc for each?

[09:02:02.0000] <mookid>
rather than one you can stick into all the players?

[09:03:00.0000] <mookid>
you're wierd.

[09:03:01.0000] <yecril71>
Meaning, having a blue-ray disk and a blue-ray player, you would rather prefer it to play CD audio?

[09:03:02.0000] <takkaria>
I might do, yeah

[09:03:03.0000] <yecril71>
For what reason?

[09:03:04.0000] <takkaria>
you know what directors are like, the DVD release will be closer to the cinema release than the blu-ray

[09:03:05.0000] <mookid>
yeah *MAYBE*

[09:04:00.0000] <mookid>
you're confusing yourself

[09:04:01.0000] <mookid>
-_-

[09:04:02.0000] <takkaria>
no, I'm just not taking this all as seriously as you are. :)

[09:04:03.0000] <mookid>
cmon admit - that analogy just killed you :P

[09:05:00.0000] <takkaria>
see, I think that your argument works fine given your assumptions

[09:05:01.0000] <takkaria>
I just don't think your assumptions are realistic

[09:05:02.0000] <mookid>
why - because that's not how it works at the moment?

[09:06:00.0000] <mookid>
no sh*t!

[09:07:00.0000] <takkaria>
no, because people sometimes want PDFs and sometimes want HTML, and being able to specify between them is quite useful

[09:07:01.0000] <mookid>
which you can do..

[09:07:02.0000] <mookid>
you just dont get it :/

[09:08:00.0000] <takkaria>
I'm sorry you think that

[09:08:01.0000] <mookid>
specifying between them is conneg

[09:08:02.0000] <mookid>
HTTP has conneg

[09:08:03.0000] <mookid>
that's what Accept is for

[09:08:04.0000] <mookid>
if you can't understand that

[09:08:05.0000] <mookid>
what m I supposed to do abuot that:?

[09:08:06.0000] <takkaria>
no, I get that :)  but to the user, that's all behind-the-scenes stuff

[09:09:00.0000] <takkaria>
if you want to paste someone a link to the PDF version of something, you go "copy link location" and paste it to them in an IM window

[09:09:01.0000] <takkaria>
or in an email, whatever

[09:09:02.0000] <takkaria>
and as soon as you do that, you lose the invisible metadata hidden in the HTML

[09:09:03.0000] <mookid>
I appreciate that.. then it's down to how the user wants to use the system

[09:10:00.0000] <mookid>
that's a design descision

[09:10:01.0000] <mookid>
you're not comfortable with that.. fair enough

[09:10:02.0000] <yecril71>
It would work if the negotiated content were an explicit REDIRECT

[09:10:03.0000] <mookid>
that's possible aswell

[09:10:04.0000] <mookid>
that's what the location header is for

[09:11:00.0000] <mookid>
if that was even a requirement

[09:11:01.0000] <mookid>
which it isnt given my methodology

[09:11:02.0000] <mookid>
look - it's not about being right

[09:11:03.0000] <mookid>
I really dont want to say 'my way is better than yours'

[09:11:04.0000] <mookid>
it's just different - and legitimately so.. HTTP has a conneg mechanism that I want to use

[09:11:05.0000] <mookid>
it's not for fun..

[09:11:06.0000] <yecril71>
What about the type attribute?

[09:12:00.0000] <yecril71>
Why is it not enough to say type="application/pdf">

[09:12:01.0000] <yecril71>
?

[09:14:00.0000] <mookid>
that attribute is standard across all referencing tags and specifically tied to the Accept header is it?

[09:15:00.0000] <yecril71>
I think the A tag is enough.

[09:15:01.0000] <mookid>
well you would because you dont know what you're talking abuot

[09:16:00.0000] <yecril71>
Your example contained an A.

[09:16:01.0000] <mookid>
'enough'

[09:16:02.0000] <mookid>
^compute. understand.

[09:17:00.0000] <yecril71>
Compute what?

[09:17:01.0000] <mookid>
look at what you wrote.. and then my response.

[09:17:02.0000] <mookid>
I was referring to your use of the word enough

[09:17:03.0000] <yecril71>
And what does it have to do with computing?

[09:18:00.0000] <yecril71>
Besides, in order to explain things, you have to talk to people who do not know what you are talking about.

[09:19:00.0000] <yecril71>
That is quite natural, and you should be prepared to explain things to them.

[09:19:01.0000] <mookid>
compentent people, yeah

[09:19:02.0000] <yecril71>
Making fun of them will not help you to make your case.

[09:19:03.0000] <mookid>
I gave you the benefit of the doubt about 24 hours ago

[09:19:04.0000] <mookid>
you only have yourself to blame

[09:20:00.0000] <mookid>
it wore thing.

[09:20:01.0000] <mookid>
thin^

[09:20:02.0000] <mookid>
there is no hope

[09:20:03.0000] <yecril71>
And everything ends in Armageddon.

[09:20:04.0000] <mookid>
decapitate yourself

[09:20:05.0000] <yecril71>
But what does it have to do with accept attribute wannabe?

[09:21:00.0000] <mookid>
ok he's on ignore now

[09:21:01.0000] <yecril71>
Encouraging people to commit suicide is a criminal offense.

[09:21:02.0000] <mookid>
takkaria: do you take the point that it's a design descision or not?

[09:22:00.0000] <yecril71>
At least in my country.

[09:26:00.0000] <mookid>
Philip`: can you wade in on the mailing list about this?

[09:27:00.0000] <mookid>
it seems more solid if we do it on there

[09:27:01.0000] <takkaria>
mookid: a design decision?  not really.  invisible metadata in cases like this is bad for users

[09:27:02.0000] <mookid>
in your opinion

[09:28:00.0000] <mookid>
that's ok I'm comfortable with that

[09:28:01.0000] <takkaria>
sure, it is a design decision to be hostile to users, but I don't generally think that's something HTML5 should allow in as much as it can stop it

[09:28:02.0000] <mookid>
if I get context menu when I right click a URI that lets me decide what UA to open it with

[09:28:03.0000] <mookid>
that would be super cool

[09:29:00.0000] <mookid>
people associate HTTP with browsers too much

[09:29:01.0000] <mookid>
takkaria: that's my problem with this - it's not your place to stop it

[09:29:02.0000] <mookid>
it's part of the HTTP protocol

[09:30:00.0000] <takkaria>
I think HTML5 in excellent place to stop things which are bad for users

[09:30:01.0000] <takkaria>
if not HTML5, then where?

[09:30:02.0000] <mookid>
within the scope of web interfaces

[09:30:03.0000] <mookid>
not 'how web developers use HTTP'

[09:31:00.0000] <mookid>
well nowhere - if anywhere in HTTP

[09:31:01.0000] <mookid>
since that's the protocol

[09:31:02.0000] <mookid>
we're tlakign about here

[09:31:03.0000] <mookid>
if you werent supposed to be able to do it with HTTP

[09:31:04.0000] <mookid>
HTTP wouldnt allow it

[09:31:05.0000] <mookid>
infact, they actively encourage it

[09:31:06.0000] <mookid>
by providing protocol level conneg

[09:32:00.0000] <takkaria>
just because HTTP provides something doesn't mean it's necessarily good for users

[09:32:01.0000] <mookid>
who are you to decide that?

[09:32:02.0000] <takkaria>
a user? :)

[09:32:03.0000] <mookid>
well that's probably the most acurate thing you've said all day

[09:33:00.0000] <mookid>
why don't you see HTML as somethign to liberate developers ratehr than restrict them

[09:33:01.0000] <takkaria>
actually, I talked to people about the impact of drug policy on crime earlier today, that was about as accurate as what I said about. :)

[09:33:02.0000] <mookid>
legalise maaaaaan

[09:33:03.0000] <takkaria>
why do you see HTML as something to do things users don't want? :)

[09:34:00.0000] <mookid>
english please

[09:34:01.0000] <takkaria>
well, you say that HTML is something to liberate developers

[09:34:02.0000] <takkaria>
I think it should be focusing more on what is helpful to users

[09:34:03.0000] <mookid>
lol

[09:34:04.0000] <mookid>
users dont use html

[09:34:05.0000] <mookid>
html is given to them by developers

[09:34:06.0000] <takkaria>
sure they do.  what else do I use when I browse the web?

[09:34:07.0000] <mookid>
it's just markup

[09:35:00.0000] <mookid>
you use a web browser

[09:35:01.0000] <takkaria>
if developers can't do things that aren't good for users, that's a good thing as far as I'm concerned, as a user

[09:35:02.0000] <mookid>
who are you to decide what developers should and shouldnt be doing?

[09:35:03.0000] <takkaria>
a user. :)

[09:35:04.0000] <mookid>
that's not a good answer

[09:35:05.0000] <mookid>
I know you think it is, it's not

[09:36:00.0000] <takkaria>
well, I have as much say in the matter as anyone else, really

[09:36:01.0000] <mookid>
it's pretty arrogant to assume you have a better grasp than other people and that they can't possibly work it out when you can

[09:36:02.0000] <mookid>
I just dont understand that perspectife

[09:36:03.0000] <mookid>
at all

[09:36:04.0000] <takkaria>
I don't think I said that

[09:36:05.0000] <mookid>
it's written between the ilines

[09:36:06.0000] <takkaria>
I just said that as a user, I'd rather developers didn't have the ability to do things that I don't want them to do

[09:37:00.0000] <takkaria>
like hiding invisible metadata that makes URL copy and pasting difficult

[09:37:01.0000] <mookid>
right.. and that's your rational for not allowing HTML to provision better control of HTTP conneg?

[09:37:02.0000] <takkaria>
yup

[09:37:03.0000] <mookid>
that's YOUR perspective

[09:37:04.0000] <takkaria>
and yours is yours. :)

[09:37:05.0000] <mookid>
how do you know that's right?

[09:37:06.0000] <mookid>
yes.. and I'm saying

[09:37:07.0000] <mookid>
I dont know definitively what's better

[09:38:00.0000] <mookid>
I just know that the most appropriate situation is where we can coexist

[09:38:01.0000] <mookid>
which an optional attribute provides

[09:38:02.0000] <takkaria>
I'd like to note that the design principles for the HTML WG (http://www.w3.org/html/wg/html5/principles/#priority-of-constituencies) mention: "n case of conflict, consider users over authors over implementors over specifiers over theoretical purity"

[09:38:03.0000] <takkaria>
so I'd guess that the WG's official position is pretty much in line with mine

[09:39:00.0000] <mookid>
so you create a conflict and then accuse me of rocking the boat and confusing users with change

[09:39:01.0000] <mookid>
that sounds like the lazy-boy manifesto right there

[09:39:02.0000] <takkaria>
I don't think I'm creating a conflict at all

[09:39:03.0000] <mookid>
congratulaitons you lot completely fooled me

[09:39:04.0000] <takkaria>
your proposal would make URL copy and pasting, a very common activity, harder

[09:39:05.0000] <mookid>
glad you're at the helm of HTML

[09:39:06.0000] <mookid>
no it wouldnt

[09:39:07.0000] <takkaria>
it would, because there'd be invisible metadata that wouldn't get copy-pasted

[09:40:00.0000] <mookid>
what is invisible?

[09:40:01.0000] <takkaria>
and the conflict is between you wanting to make it harder and users wanting it to work

[09:40:02.0000] <takkaria>
the accept="" attribute

[09:40:03.0000] <mookid>
the URI indicates where the resource is..

[09:40:04.0000] <mookid>
it's done it's job

[09:40:05.0000] <mookid>
the content negotiation is down to the UA yoru using and the context you are requesting from

[09:40:06.0000] <mookid>
the fact that you dont see it that way

[09:40:07.0000] <takkaria>
if someone copy and pastes the link from <a type="application/pdf" href="blah">, then they will lose the fact it was intended to PDF

[09:41:00.0000] <mookid>
I KNOW THAT

[09:41:01.0000] <mookid>
that's not complicated.

[09:41:02.0000] <takkaria>
aye.  and that is what the invisible metadata is

[09:41:03.0000] <takkaria>
and invisible metadata is harmful to users, especially in this case

[09:41:04.0000] <Philip`>
mookid: I'd rather not wade in since I don't think I have anything valuable to contribute and I'd just get my socks wet

[09:41:05.0000] <mookid>
a URI indicates a resource

[09:41:06.0000] <Philip`>
A URI indicates what the developer thinks is a resource, which may not be what the user thinks is a resource

[09:41:07.0000] <mookid>
yeah sure

[09:41:08.0000] <mookid>
bbl

[09:42:00.0000] <Philip`>
(so users might want a way to distinguish what they consider separate resources, even if the developer thought they were one resource and put them at the same URI)

[09:42:01.0000] <yecril71>
By the way, there are mouses that do not have the right button.

[09:43:00.0000] <yecril71>
Because they have only one button.

[09:44:00.0000] <yecril71>
OTOH, it is the same thing with METHOD=POST.

[09:45:00.0000] <yecril71>
Imagine METHOD=POST returning content that is inaccessible otherwise.

[09:45:01.0000] <yecril71>
And you cannot put the URL into the clipboard.

[09:53:00.0000] <Philip`>
POST is non-idempotent so there's a good reason to not let people arbitrarily copy around strings that let them accidentally repeat POST requests

[09:53:01.0000] <Philip`>
but that reason doesn't apply when you're simply GETting (content-negotiated) files

[10:01:00.0000] <yecril71>
The problem with hsivonen�s script is caused by the DOM viewer code.

[10:02:00.0000] <yecril71>
The content rendered separately does not cause Internet Explorer to hang.

[10:03:00.0000] <Philip`>
Oh, that's a useful observation - I think I saw a bug exactly like that at some point in the past...

[10:03:01.0000] <Philip`>
https://connect.microsoft.com/IE/feedback/ViewFeedback.aspx?FeedbackID=366200

[10:07:00.0000] <yecril71>
It is not a big problem given that inline frames are not strictly conforming.

[10:08:00.0000] <yecril71>
My experience shows it is safer to use window.open

[10:08:01.0000] <yecril71>
Which is nonconforming as well, but it works better at least.

[10:08:02.0000] <Philip`>
iframes are perfectly conforming in HTML5

[10:09:00.0000] <Philip`>
and any browsers shouldn't crash or freeze on any input, regardless of conformingness

[10:09:01.0000] <Philip`>
s/any/anyway/

[10:09:02.0000] <yecril71>
Of course.

[10:09:03.0000] <yecril71>
Hovewer, HTML5 is not implemented.


2008-11-20
[00:19:00.0000] <hsivonen>
hmm. Perhaps Roy hasn't looked at what some of the 2000 Apache committers have been up to lately...

[02:24:00.0000] <Hixie>
danbri: (i ask merely because different people have said similar things but in each case meant something different)

[03:02:00.0000] <danbri>
and dropped in html5?

[03:02:01.0000] <Hixie>
nope

[03:48:00.0000] <mookid>
I'm not saying you dont understand I'm saying you can't tell

[03:48:01.0000] <Hixie>
so how could i get the perspective to tell?

[03:48:02.0000] <mookid>
by allowing both

[03:48:03.0000] <mookid>
and seeing what happens

[03:48:04.0000] <mookid>
why do you even want to make that call

[03:48:05.0000] <Hixie>
we can't both allow and disallow content negotiation

[03:49:00.0000] <mookid>
if someone makes a web app with HTTP conneg

[03:49:01.0000] <mookid>
and it's crap for users..

[03:49:02.0000] <Hixie>
they're mutually exclusive options

[03:49:03.0000] <mookid>
gues what

[03:49:04.0000] <mookid>
NO ONE WILL USE IT

[03:49:05.0000] <mookid>
that's how capitalism works

[03:49:06.0000] <Hixie>
right, indeed, nobody uses it

[03:49:07.0000] <mookid>
NO

[03:49:08.0000] <Hixie>
we've seen that already

[03:49:09.0000] <mookid>
that's wrong

[03:49:10.0000] <Hixie>
well, a few people do

[03:49:11.0000] <gsnedders>
hsivonen: I was discussing it with him in private, but the current draft hal CC: www-tag

[03:49:12.0000] <mookid>
nobody uses it because it's not provisioed for by existing technmologies

[03:49:13.0000] <Hixie>
w3c does for some resources

[03:49:14.0000] <mookid>
that's  circular argument

[03:49:15.0000] <mookid>
and you know it

[03:49:16.0000] <Hixie>
though almost uniformly when they do, they screw things up

[03:49:17.0000] <mookid>
it's totally disingeous

[03:49:18.0000] <hsivonen>
"progression of a technologist" reminds me of TUITT (The Architect's Progress) :-)

[03:49:19.0000] <mookid>
no

[03:49:20.0000] <mookid>
go back

[03:49:21.0000] <hsivonen>
gsnedders: ah

[03:49:22.0000] <mookid>
just go back

[03:49:23.0000] <gsnedders>
*has

[03:49:24.0000] <mookid>
 < Hixie> right, indeed, nobody uses it

[03:49:25.0000] <mookid>
 < Hixie> right, indeed, nobody uses it

[03:50:00.0000] <mookid>
taht is nonsense

[03:50:01.0000] <mookid>
nobody has th oppotunity to use it

[03:50:02.0000] <mookid>
because you cant do it

[03:50:03.0000] <mookid>
there's no reason you can't provision for both HTTP and URI conneg

[03:50:04.0000] <Hixie>
how so? all the browsers support it, all the servers support it, why would people not use it if they wanted it?

[03:50:05.0000] <mookid>
they arent mututally exclusive

[03:50:06.0000] <mookid>
the browsers dont support it

[03:50:07.0000] <Hixie>
sure they do

[03:50:08.0000] <mookid>
no they dont

[03:50:09.0000] <Hixie>
they send Accept headers correctly

[03:51:00.0000] <mookid>
not dynamically there's no markup for it

[03:51:01.0000] <mookid>
that's my point.

[03:51:02.0000] <Hixie>
the markup idea doesn't do content _negotiation_, it does content _selection_, which is what URIs are _for_

[03:51:03.0000] <Hixie>
URIs are resource identifiers

[03:51:04.0000] <mookid>
URIs are USED THAT WAY AT THE MOMENT

[03:51:05.0000] <mookid>
BECAUSE OF THAT

[03:52:00.0000] <mookid>
you're looking at the relationship wrong

[03:52:01.0000] <Hixie>
right, that's the architecture of the web

[03:52:02.0000] <mookid>
the reason

[03:52:03.0000] <mookid>
URIs are used to do that

[03:52:04.0000] <mookid>
is because there's no other alternative 

[03:52:05.0000] <mookid>
it's NOT PROVISIONED FOR

[03:52:06.0000] <Hixie>
why do we need an alternative?

[03:52:07.0000] <mookid>
why not?

[03:52:08.0000] <mookid>
it's another pattern

[03:52:09.0000] <gsnedders>
Bloat?

[03:52:10.0000] <Hixie>
because we don't need it

[03:52:11.0000] <hsivonen>
alternatives are expensive

[03:52:12.0000] <Hixie>
and it's expensive to implement, spec, test, and debug

[03:52:13.0000] <mookid>
if you can provision for it simply and withotu breaking existing methodologies

[03:52:14.0000] <mookid>
why would you not do that

[03:52:15.0000] <mookid>
the answer is that

[03:52:16.0000] <mookid>
you dont agree with that way of doing thigns

[03:52:17.0000] <Hixie>
surely we should be focusing on making sure we fit the web's architecture, anyway

[03:52:18.0000] <gsnedders>
Unless we need something, it just bloats the spec and makes implementing it more expensive

[03:53:00.0000] <mookid>
well you should'nt have the authority to do that

[03:53:01.0000] <Hixie>
which is that uris are identifiers for resources

[03:53:02.0000] <mookid>
what is a resource?

[03:53:03.0000] <hsivonen>
mookid: should you have the authority?

[03:53:04.0000] <mookid>
I have my definition

[03:53:05.0000] <mookid>
you have you

[03:53:06.0000] <mookid>
yorus

[03:53:07.0000] <mookid>
I'm saying

[03:53:08.0000] <mookid>
lets both do it our ways

[03:53:09.0000] <mookid>
and see which one wins

[03:53:10.0000] <mookid>
your saying

[03:53:11.0000] <Hixie>
"you're"

[03:53:12.0000] <mookid>
dont be a dick

[03:54:00.0000] <Hixie>
every time you misuse "your" i get really confused

[03:54:01.0000] <mookid>
seriously?

[03:54:02.0000] <mookid>
wow.

[03:54:03.0000] <jgraham>
mookid: AFAICT in your "arcitecture" you would still need to have unique URIs for unique content types so that you could do something like a HTML page with links to specific versions of the resource, which you have said would be the fallback for browsers that didn't support <a accept>

[03:54:04.0000] <Hixie>
yeah i'm like "my saying? what saying? what?"

[03:54:05.0000] <jgraham>
architecture*

[03:54:06.0000] <mookid>
it's funny why you've stopped the conversation

[03:54:07.0000] <mookid>
is taht becaue you're in a hole

[03:55:00.0000] <mookid>
perhaps?

[03:55:01.0000] <Hixie>
i'm still confused about how this accept attribute would work when you want to copy and paste the url

[03:55:02.0000] <mookid>
ABORT ABORT

[03:55:03.0000] <mookid>
it doesnt work - it depends on the default accept headers of the client

[03:55:04.0000] <Hixie>
but more importantly, we've explained that new features are expensive, and that we don't see an advantage to this proposal, as it doesn't introduce anything

[03:55:05.0000] <Hixie>
and we've explained how it breaks the web's architecture

[03:55:06.0000] <mookid>
that's how I want my application to behave..

[03:55:07.0000] <mookid>
yhou dont

[03:55:08.0000] <mookid>
that's fine

[03:55:09.0000] <mookid>
why cant we both have our way?

[03:55:10.0000] <wilhelm>
Even if this suggestion makes it into the spec, you would still have to convince browser vendors to actually implement it.

[03:55:11.0000] <mookid>
why do you have to restrict me

[03:55:12.0000] <mookid>
on the basis that you think you're right

[03:56:00.0000] <mookid>
well at least I have one step done

[03:56:01.0000] <mookid>
then I can do the next

[03:56:02.0000] <Hixie>
are you going to pay the engineers to implement this, to test it, to spec it, to write the tutorials, to debug their apps when it breaks, etc?

[03:56:03.0000] <hsivonen>
mookid: should we have as many ways of doing things as there are people in the world?

[03:56:04.0000] <mookid>
hsivonen: no, clearly not shut up

[03:56:05.0000] <jgraham>
mookid: be civil

[03:56:06.0000] <mookid>
no he's being annoying

[03:57:00.0000] <Hixie>
so are you but we're still being polite :-)

[03:57:01.0000] <mookid>
passive agressive is agressive

[03:57:02.0000] <mookid>
I just dont act like agirl

[03:57:03.0000] <mookid>
:)

[03:57:04.0000] <Hixie>
his point is well made, though, if five people want five different things, we obviously don't want to add five features to the language

[03:57:05.0000] <mookid>
Hixie: I dont have to pay anyone - developers that believed in doing it that way would od it

[03:57:06.0000] <Hixie>
we have to pick the right nes

[03:57:07.0000] <mookid>
the ones that didnt

[03:57:08.0000] <mookid>
wouldnt

[03:57:09.0000] <mookid>
funnily enough

[03:57:10.0000] <mookid>
but btoh can co-exist

[03:58:00.0000] <mookid>
there's not imcopatability

[03:58:01.0000] <Hixie>
mookid: ok, well, i am deciding not to do it

[03:58:02.0000] <mookid>
WHY?

[03:58:03.0000] <Hixie>
mookid: because you won't pay me to add it to the spec :-)

[03:58:04.0000] <mookid>
erm

[03:58:05.0000] <Hixie>
mookid: it costs me time to do it

[03:58:06.0000] <mookid>
aaaahhhhhhh

[03:58:07.0000] <Hixie>
mookid: and you aren't willing to pay the cost

[03:58:08.0000] <mookid>
here we are

[03:58:09.0000] <mookid>
why didnt you just say that?

[03:58:10.0000] <Hixie>
i did

[03:58:11.0000] <mookid>
no you didnt

[03:58:12.0000] <Hixie>
i said adding teh feature was expensive

[03:58:13.0000] <mookid>
that's the first time you've mentioned money

[03:58:14.0000] <mookid>
how much does it cost?

[03:59:00.0000] <krijnh>
-_-

[03:59:01.0000] <mookid>
If I buy you a boko on web architecture will you do it?

[03:59:02.0000] <hsivonen>
do we have a link to an explanation of opportunity cost from the FAQ yet?

[03:59:03.0000] <mookid>
I nkow what an opporunity cost is

[03:59:04.0000] <mookid>
cost forgone by chosing the next best alternative

[04:00:00.0000] <Hixie>
well there's the 15 minutes for me to write the spec, the day or two spent fixing the spec over the next few months, the month or so of writing test cases, then the browser vendors each have to implement it so that's a dozen times a few days, then there's the tutorial writers who will have to spend a few hours or days explaining it

[04:00:01.0000] <mookid>
just for one optional attribute?

[04:00:02.0000] <Hixie>
then there's the cost of all the people who try to understand it, probably 5 minutes to an hour each, times maybe 100 million people

[04:00:03.0000] <wilhelm>
Implementation, testing, debugging, regression testing on every nightly build with all used configurations-- even the smallest features takes a lot of resources.

[04:00:04.0000] <mookid>
wow your processes are SERIOUSLY broken

[04:00:05.0000] <Hixie>
then there's the cost of regression testing that wilhelm mentions

[04:01:00.0000] <mookid>
I can't honestly believe that one optional attribute can make that much difference in the grand scheme of things

[04:01:01.0000] <Hixie>
then there's the cost of actual debugging for people using this

[04:01:02.0000] <Hixie>
let's assume an hour or so for a million people...

[04:01:03.0000] <mookid>
thos arent costs to this project that's not a reasonable argument

[04:02:00.0000] <krijnh>
(Don't forget the hours spent on this in IRC :)

[04:02:01.0000] <mookid>
those will be incurred by the 'insane' people who want to use HTTP conneg

[04:02:02.0000] <jgraham>
/me will be expecting his cheque in the post :)

[04:02:03.0000] <mookid>
I meant.. what does it cost to get this in the spec

[04:02:04.0000] <mookid>
not - what are the projected costs to industry

[04:02:05.0000] <mookid>
if it's not worth it

[04:03:00.0000] <mookid>
indstry wont pay for it

[04:03:01.0000] <Hixie>
so let's see... assuming $20/hour, that would be... about a billion dollars total, if my calculations are correct.

[04:03:02.0000] <mookid>
for one attribute?

[04:03:03.0000] <Hixie>
yup

[04:03:04.0000] <mookid>
you're a joke.

[04:03:05.0000] <Hixie>
that's why we only add the minimum

[04:03:06.0000] <Hixie>
well the cost is spread over the entire human race

[04:03:07.0000] <mookid>
ilke VIDEOTAGS

[04:03:08.0000] <Hixie>
over many decades

[04:03:09.0000] <mookid>
we already have video tags

[04:03:10.0000] <mookid>
we can already embed sounds

[04:03:11.0000] <Hixie>
consider how much money google paid to buy youtube

[04:03:12.0000] <mookid>
why not just fix the tags that exist?

[04:03:13.0000] <mookid>
rather than creat your own gay new markup

[04:03:14.0000] <Hixie>
and then you'll see that a billion dollars for a feature like that doesn't sound like much anymore

[04:04:00.0000] <mookid>
that no user actually cares about

[04:04:01.0000] <mookid>
who reads html

[04:04:02.0000] <mookid>
seriously

[04:04:03.0000] <Hixie>
because that feature makes money :-)

[04:04:04.0000] <mookid>
no ti doesnt

[04:04:05.0000] <mookid>
video on teh web makes money

[04:04:06.0000] <mookid>
you can do that with the tags that exist

[04:04:07.0000] <Hixie>
right

[04:04:08.0000] <mookid>
they need fixing

[04:04:09.0000] <Hixie>
not without paying adobe

[04:05:00.0000] <wilhelm>
Porting Flash to new platforms along with a browser is a pain.

[04:05:01.0000] <mookid>
so it would cost whatwg how much to add this attribute?

[04:05:02.0000] <wilhelm>
I'd prefer not to do that again. (c:

[04:05:03.0000] <Hixie>
it would cost hte whatwg nothing at all, the whatwg has no money. it would cost the human race about a billion dollars spread over a few years.

[04:05:04.0000] <mookid>
investment

[04:05:05.0000] <Hixie>
with what return?

[04:05:06.0000] <mookid>
who knows

[04:06:00.0000] <mookid>
they arent going to pay up if there's no ROI

[04:06:01.0000] <Hixie>
exactly

[04:06:02.0000] <mookid>
why dont you let them decide that

[04:06:03.0000] <mookid>
WHY DONT

[04:06:04.0000] <mookid>
YOU LET

[04:06:05.0000] <mookid>
THEM DECIDE

[04:06:06.0000] <Hixie>
because i am paid to make these judgement calls

[04:06:07.0000] <mookid>
how difficult is it to understand that is the most efficient way to make that descision

[04:06:08.0000] <Hixie>
that's my job

[04:06:09.0000] <mookid>
it's how our economies work

[04:06:10.0000] <Hixie>
well

[04:06:11.0000] <Hixie>
yes

[04:06:12.0000] <annevk2>
who is "them"?

[04:06:13.0000] <krijnh>
Wb annevk2 :)

[04:06:14.0000] <Hixie>
sometimes, the browser vendors implement things without a spec

[04:06:15.0000] <mookid>
do you know Smith's "invisible hand" ?

[04:06:16.0000] <annevk2>
thx krijnh

[04:06:17.0000] <hsivonen>
wilhelm: did Opera port Flash to Wii?

[04:07:00.0000] <Hixie>
and that's when i know i made a mistake in saying "no"

[04:07:01.0000] <mookid>
erm

[04:07:02.0000] <Hixie>
and sometimes, the browser vendors don't implement something that i put in the spec

[04:07:03.0000] <mookid>
lol

[04:07:04.0000] <Hixie>
and that's when i know i made a mistake in saying "yes"

[04:07:05.0000] <mookid>
by not putting it in the spec you create an articificial barrier to entry

[04:07:06.0000] <Hixie>
but in this particular case, i haven't heard a single browser vendor be positive about it

[04:07:07.0000] <wilhelm>
hsivonen: Yes.

[04:08:00.0000] <mookid>
it's something else they have to worry about 0 they probably dont understand it

[04:08:01.0000] <mookid>
so you've talked to browser vendors about this?

[04:08:02.0000] <mookid>
that's blatantly a lie

[04:09:00.0000] <Philip`>
Hixie: C's pointer syntax made more sense when I learned that the type declarations were meant to reflect how you would use variables of that type, e.g. "char **foo" means *foo is a char* and **foo is a char, and "char *foo[]" means foo[n] is a char* and *foo[n] is a char

[04:09:01.0000] <jgraham>
mookid: You've tlked to browser vendors

[04:09:02.0000] <wilhelm>
mookid: You are talking to browser vendors about this right now.

[04:09:03.0000] <Hixie>
mookid: pretty much everyone who has spoken in this conversation other than me and you is a browser vendor

[04:09:04.0000] <mookid>
ahh taht explains ALOT

[04:09:05.0000] <mookid>
:)

[04:09:06.0000] <Hixie>
actually i guess now that the G1 is out and Chrome is out even i'm technically a browser vendor

[04:10:00.0000] <hsivonen>
Philip`: putting * after the space has never made sense to me. I always think that pointerness is part of the type.

[04:10:01.0000] <Hixie>
but we'll gloss over that so i can continue to act as a vendor-neutral arbitor!

[04:10:02.0000] <mookid>
huhuhuhu

[04:10:03.0000] <tthorsen>
hsivonen: consider this "char* a, b" a is a pointer to a char, but b is just a char

[04:10:04.0000] <Hixie>
hsivonen: it only makes sense because of the binding of the star token in "int* a, b" (b's type is "in" not "int*")

[04:10:05.0000] <Hixie>
what tthorsen said

[04:11:00.0000] <Philip`>
hsivonen: It seems to make some sense either way - "char* x" means x is a char*, while "char *x" means *x is a char

[04:11:01.0000] <Hixie>
but that's a bug in C and i agree that the * should be before the space

[04:11:02.0000] <hsivonen>
tthorsen: I don't declare multiple variables on one line in C or C++.

[04:11:03.0000] <mookid>
no offence but how many 'big thinkers' are invovled in the nitty gritty of HTML5 -_-

[04:11:04.0000] <mookid>
brwoser vendors or not

[04:11:05.0000] <Hixie>
mookid: how does one determine if one is a big thinker?

[04:11:06.0000] <mookid>
I just really dont think this is the appropriate place to be making design descisions

[04:12:00.0000] <Hixie>
where is appropriate?

[04:12:01.0000] <mookid>
well..

[04:12:02.0000] <mookid>
DESIGN TIME?

[04:12:03.0000] <mookid>
maybe?

[04:12:04.0000] <mookid>
...

[04:12:05.0000] <Hixie>
in #whatwg it is design time all the time

[04:12:06.0000] <Hixie>
that's what we do

[04:12:07.0000] <mookid>
you dont do application design

[04:12:08.0000] <jgraham>
Apart from when we are discussing gsnedders poetry.

[04:13:00.0000] <mookid>
specifically web application design

[04:13:01.0000] <jgraham>
That is not spec design :)

[04:13:02.0000] <Hixie>
oh i thought you mean the design of the language

[04:13:03.0000] <Hixie>
i don't think it would make much sense to decide what tags html had when writing a web app :-)

[04:13:04.0000] <Hixie>
it's kinda late by then

[04:13:05.0000] <mookid>
It's notyruo place to start rejecting features on the basis that you need to 'protect' users from 'crazy' architects/developers

[04:13:06.0000] <mookid>
what?

[04:13:07.0000] <mookid>
Hixie: 

[04:14:00.0000] <mookid>
developers are entirely restricted by thtat

[04:14:01.0000] <Hixie>
should we just add all features that people request then?

[04:14:02.0000] <mookid>
no just the ones that don't break anything else and provide more alternatives to developers

[04:14:03.0000] <mookid>
users dont care about html

[04:14:04.0000] <mookid>
developers do

[04:15:00.0000] <Hixie>
your proposal breaks the web architecture

[04:15:01.0000] <mookid>
no it doesnt

[04:15:02.0000] <mookid>
you can still write URI based conneg with my proposal

[04:15:03.0000] <Hixie>
the web architecture is that a resource is identified by a uri

[04:15:04.0000] <mookid>
you can still do it your way

[04:15:05.0000] <Hixie>
you are proposing making them identify multiple resources

[04:15:06.0000] <mookid>
even if my proposal is implemetned

[04:15:07.0000] <mookid>
my proposal is to provide the OPTION

[04:15:08.0000] <Hixie>
the option to break the web architecture, right

[04:15:09.0000] <mookid>
that's compltely different

[04:15:10.0000] <mookid>
no

[04:15:11.0000] <mookid>
not to break it

[04:16:00.0000] <mookid>
the old design

[04:16:01.0000] <mookid>
and OLD applications

[04:16:02.0000] <mookid>
will still work

[04:16:03.0000] <mookid>
there's no change there..

[04:16:04.0000] <mookid>
it's only developers who make web apps that serve html with the attribute

[04:16:05.0000] <mookid>
that will have any change

[04:16:06.0000] <Hixie>
so we should add all features that people ask for if those features are optional?

[04:16:07.0000] <mookid>
provided they dont break anything else

[04:16:08.0000] <mookid>
and they have a valid use for them

[04:17:00.0000] <mookid>
which I'v eclrealy demonstrated

[04:17:01.0000] <Hixie>
wow, that would make for one hell of a big and confusing language

[04:17:02.0000] <mookid>
not really

[04:17:03.0000] <mookid>
if you use your brain

[04:17:04.0000] <Hixie>
dude i have literally THOUSANDS of e-mails of feature requests

[04:17:05.0000] <Hixie>
to go through

[04:17:06.0000] <mookid>
oh right ok then

[04:17:07.0000] <Hixie>
most of which are optional

[04:17:08.0000] <mookid>
maybve think abuot letting someone else help?

[04:17:09.0000] <Hixie>
it's not a matter of it taking time

[04:17:10.0000] <mookid>
or do you not feel like sharing the glory?

[04:17:11.0000] <Hixie>
it's a matter of the size of the language if we added them all!

[04:18:00.0000] <mookid>
or trust anyone to be as super intelligent as you obviously are

[04:18:01.0000] <mookid>
protector of the web

[04:18:02.0000] <mookid>
(architecturE)

[04:18:03.0000] <mookid>
it's a nonsense position to take

[04:18:04.0000] <Hixie>
(actually i've already requested volunteers to take over certain parts of the spec)

[04:18:05.0000] <Hixie>
(btu that's independent of my point here)

[04:18:06.0000] <Hixie>
but

[04:18:07.0000] <mookid>
I'll happilly work on that part and anything else you need

[04:18:08.0000] <Hixie>
sweet

[04:19:00.0000] <mookid>
it's a shame there's no web site for proposals though

[04:19:01.0000] <Philip`>
Specs are useless unless they're implemented, and implementors have limited resources, so even if we had infinite spec-editing capability we'd have to limit the size of the language

[04:19:02.0000] <Hixie>
http://lists.w3.org/Archives/Public/public-html/2008Oct/0127.html has the details of what needs doing

[04:20:00.0000] <Hixie>
right, the original point still stands, which is that we'd be irresponsible to just accept all (optional) proposals

[04:20:01.0000] <mookid>
right it's a trade off.. but it's pretty obvious that HTTP centric apps are emerging as a a new approach

[04:20:02.0000] <mookid>
otherwise why would you be implementing PUT/DELETE?

[04:21:00.0000] <Hixie>
ok if you agree that it is a tradeoff... what is it a tradeoff between?

[04:21:01.0000] <mookid>
between what it adds.. adding the functionality from HTTP conneg is a big addition with minimal cost and complete backwards compatability

[04:22:00.0000] <mookid>
the fact that you wouldn't use it

[04:22:01.0000] <mookid>
is neither here nor there

[04:22:02.0000] <mookid>
frankly

[04:22:03.0000] <mookid>
I mean

[04:22:04.0000] <mookid>
what realistically does a video tag add?

[04:22:05.0000] <Hixie>
well in the case of your proposal, it adds nothing, it's just another syntax for what you can do already

[04:22:06.0000] <mookid>
what?

[04:22:07.0000] <mookid>
no you CANT..

[04:22:08.0000] <Hixie>
the video element adds a way to do video without paying adobe

[04:23:00.0000] <mookid>
if you can please tell me how

[04:23:01.0000] <Hixie>
instead of src="foo" accept="text/html" you do src="foo.html"

[04:23:02.0000] <mookid>
that's not the same thing

[04:23:03.0000] <mookid>
that's totally different

[04:23:04.0000] <mookid>
that's not HTTP conneg

[04:23:05.0000] <Hixie>
it acts exactly the same to the user

[04:23:06.0000] <mookid>
not to DEVELOPERS

[04:23:07.0000] <mookid>
who are the people

[04:23:08.0000] <Hixie>
yes but HTTP conneg, as previously discussed, is a bad idea

[04:23:09.0000] <mookid>
USING HTML

[04:23:10.0000] <mookid>
users dont SUE html

[04:23:11.0000] <mookid>
USE

[04:24:00.0000] <mookid>
they dont CARE about html

[04:24:01.0000] <Hixie>
developers are developing for the users

[04:24:02.0000] <Hixie>
so if the users see no difference...

[04:24:03.0000] <mookid>
they couldn't care less if it was XHTML HTML whatever

[04:24:04.0000] <mookid>
er..

[04:24:05.0000] <mookid>
yeah..

[04:24:06.0000] <mookid>
but what if the developers see a huge difference?

[04:25:00.0000] <mookid>
what if it gives them a feasible way to use single URIs for many representations

[04:25:01.0000] <mookid>
that's not possible at the moment

[04:25:02.0000] <Hixie>
well in this paricular case, what difference do they see?

[04:25:03.0000] <mookid>
that's a big difference

[04:25:04.0000] <mookid>
well

[04:25:05.0000] <mookid>
not possible -> is now possible

[04:25:06.0000] <hesslow>
mookid: Why don't just use mod_rewrite and use the extension part as content-type?

[04:25:07.0000] <mookid>
that's more of a change than

[04:25:08.0000] <mookid>
crappy object and embed tags for video -> video tags for vide

[04:25:09.0000] <hsivonen>
mookid: you aren't considering things on high enough a level of abstraction

[04:26:00.0000] <Hixie>
what is possible now that wasn't before?

[04:26:01.0000] <hsivonen>
mookid: when you go high enough, the Accept header is just an implementation detal

[04:26:02.0000] <hsivonen>
detail

[04:26:03.0000] <mookid>
hesslow: I know you can do that.. that's what I do at the moemnt

[04:26:04.0000] <mookid>
everyone does it that way

[04:26:05.0000] <mookid>
because you cant do it any other way

[04:26:06.0000] <mookid>
because HTTP conneg isnt supported

[04:26:07.0000] <mookid>
you can make a browser GUI that makes use of HTTP conneg

[04:26:08.0000] <hesslow>
So as a developer is there any difference then?

[04:26:09.0000] <mookid>
so no one produces web apps that use http conneg

[04:27:00.0000] <mookid>
yes there's a huge difference

[04:27:01.0000] <mookid>
because you *can* use HTTP conneg

[04:27:02.0000] <Hixie>
as hsivonen says, the content negotiation pattern is supported, just not using the low-level implementation details of http vs uris

[04:27:03.0000] <mookid>
and so you *can* serve multiple content-types from one URI

[04:27:04.0000] <Hixie>
but that's an implementation detail

[04:27:05.0000] <Hixie>
it doesn't affect whether or not the pattern is possible

[04:27:06.0000] <mookid>
yes it does

[04:27:07.0000] <Hixie>
how?

[04:27:08.0000] <mookid>
it's completely different

[04:27:09.0000] <Hixie>
how?

[04:28:00.0000] <mookid>
well it's PROTOCOL level

[04:28:01.0000] <Hixie>
at the architectural level, how is it different?

[04:28:02.0000] <mookid>
masively

[04:28:03.0000] <Hixie>
how?

[04:28:04.0000] <mookid>
URI strcture?

[04:28:05.0000] <mookid>
cachine mechanisms

[04:28:06.0000] <mookid>
caching

[04:28:07.0000] <Hixie>
that's an implementation detail, not an architectural level detail

[04:28:08.0000] <Hixie>
i'm asking what difference does it make at the architecture level

[04:28:09.0000] <hsivonen>
caching has to happen on the representation level anyway

[04:29:00.0000] <mookid>
Vary

[04:29:01.0000] <Philip`>
If you encode the desired content-type in the URI, it's not possible for a general-purpose HTTP-processing device to understand what's the resource identifier and what's the content-type selector, because there's no standard for encoding that stuff in URIs

[04:29:02.0000] <mookid>
yes exactly

[04:29:03.0000] <mookid>
which means that if you were using HTTP conneg

[04:30:00.0000] <mookid>
and your brwoser knew it was only requesting text/html

[04:30:01.0000] <Hixie>
Philip`: such a standard or convention could be easily established

[04:30:02.0000] <hsivonen>
Philip`: is that a problem?

[04:30:03.0000] <mookid>
youc ould implement security mechanism

[04:30:04.0000] <mookid>
to reject that content

[04:30:05.0000] <mookid>
yes

[04:30:06.0000] <mookid>
read what I just wrote

[04:30:07.0000] <mookid>
is one example of the benefits

[04:30:08.0000] <mookid>
of UNIFORMITY

[04:30:09.0000] <Philip`>
whereas there is a standard for encoding the resource identifier in the URI and the content-type selector in Accept, so devices that don't know about your application-specific encoding mechanism can still separate the resource/representation identifiers

[04:30:10.0000] <hesslow>
Are there any good way to do fallback for your own protocols-handlers? I found the type-attribute on a-tags for content-type but there is no way to specify how the fallback should be done. What I would want is a fallback to a normal http-uri if the browser don't support the protocol

[04:31:00.0000] <jcranmer>
mookid: the sever need not match the Accept header

[04:31:01.0000] <mookid>
exactly..

[04:31:02.0000] <mookid>
so if it doesnt

[04:31:03.0000] <mookid>
there's a standard way of sayign

[04:31:04.0000] <mookid>
'hey look I wasnt expecting that'

[04:31:05.0000] <mookid>
if it's in the URI

[04:31:06.0000] <mookid>
that's not feasible

[04:31:07.0000] <mookid>
?type=html ;type=text/html .html

[04:32:00.0000] <Philip`>
Hixie: You couldn't establish it in a way that wouldn't conflict with other people not following that standard and perfectly legitimately happening to use the same magic URI query parameter or whatever the content-type thing is stored as

[04:32:01.0000] <mookid>
you couldn't ?

[04:33:00.0000] <mookid>
dunno if I'm reading that right

[04:33:01.0000] <Hixie>
Philip`: yeah, fair point

[04:34:00.0000] <Philip`>
hsivonen: I don't know if it's a practical problem or not, since it depends on whether general-purpose HTTP-processing devices can do interesting things with that information

[04:34:01.0000] <Hixie>
Philip`: but it's unclear that the convention needs to extend beyond the server

[04:34:02.0000] <Hixie>
Philip`: (i'm not sure i understand the use cases)

[04:34:03.0000] <mookid>
well the security mechnism would be nice..

[04:35:00.0000] <Philip`>
mookid: I mean you couldn't define a standard which says e.g. ?type=text/html in the URI is equivalent to Accept:text/html, because that would conflict with people who happen to use ?type=... for something else

[04:35:01.0000] <mookid>
yeah that's what I mean

[04:35:02.0000] <mookid>
it's not uniform

[04:35:03.0000] <mookid>
so you couldn't reaasonably protect against that

[04:35:04.0000] <mookid>
if it's inteh accept header that a certain request is expected to be only text/html or text/* or whateevr

[04:36:00.0000] <mookid>
and it's outside of that

[04:36:01.0000] <mookid>
you can throw errors or whatever

[04:36:02.0000] <mookid>
you cant do that if it's in the URI

[04:36:03.0000] <Philip`>
Why bother with sending Accept? Just send a normal request, and if you don't get the text/html you expected then throw it away

[04:37:00.0000] <mookid>
As a security feature - thre's probably something you could do to prevent CSRF/XSS

[04:37:01.0000] <mookid>
particularly if the HTML is all modular and stuf

[04:37:02.0000] <mookid>
modularity is great functionality but it's going to make CSRF and XSS go nuts

[04:37:03.0000] <Hixie>
CSRF wouldn't be helped here since the page containing the markup would be the source page

[04:38:00.0000] <Philip`>
mookid: That's sounding incredibly vague :-p

[04:38:01.0000] <Hixie>
and XSS as far as I can tell wouldn't happen with a link, and we've already got sandbox="" for <iframe>

[04:38:02.0000] <mookid>
:/

[04:38:03.0000] <mookid>
so you dont take the point that its easier to implement a security mechanism against HTTP conneg?

[04:39:00.0000] <Philip`>
If there is a specific instance of some vulnerability that could best be solved by this mechanism, then that'd be more useful than saying "there's probably something you do [with this feature]"

[04:39:01.0000] <mookid>
well yeah sure I can go away and come up with the case if you like

[04:39:02.0000] <mookid>
I mean

[04:39:03.0000] <mookid>
it's hard to deny that's the case

[04:40:00.0000] <mookid>
Either way - I think you an I agree that it is a significantly different approach to aplpication development

[04:40:01.0000] <Philip`>
s/you do/you could do/

[04:40:02.0000] <mookid>
particularly from the server standpoint

[04:40:03.0000] <mookid>
of data store

[04:40:04.0000] <mookid>
and representations

[04:41:00.0000] <mookid>
+ it's not imcompatible with the current methods

[04:41:01.0000] <Philip`>
I don't think it's significantly different - it's just a different syntax, and that syntax might make a few things work better and a few things work worse

[04:41:02.0000] <mookid>
the only worse is how do I know what this string URI means

[04:43:00.0000] <mookid>
in a business context there's only so many UAs to account for

[04:43:01.0000] <mookid>
and natural lanaguage goes along way aswell 

[04:43:02.0000] <mookid>
god forbid we should humanise human computer interaction

[04:44:00.0000] <mookid>
-_--

[04:44:01.0000] <mookid>
but hey thats all 'moon' stuff

[04:48:00.0000] <Philip`>
Why humanise it? I don't like humans

[04:50:00.0000] <Philip`>
Natural language doesn't seem great - "Visit whatwg.org in your web browser then click "HTML5" in the "Specs" box" is much less helpful to anyone than "Visit http://whatwg.org/html5", and the uniformity is a significant benefit of URIs

[04:51:00.0000] <jcranmer>
most computer problems emabate from between the chair and monitor

[04:51:01.0000] <jcranmer>
s/b/n/

[04:52:00.0000] <Philip`>
That's a very old-fashioned pre-mobile view of the world :-p

[04:53:00.0000] <jcranmer>
ok then

[04:54:00.0000] <jcranmer>
most problems emanate from devices running at less than 1 KHz clock speed :-)

[05:01:00.0000] <Philip`>
jcranmer: So it's all EDSAC's fault?

[05:03:00.0000] <Hixie>
wtf, now my plugin works in firefox but in safari only the name components are shown, not the values!

[05:04:00.0000] <Hixie>
and it works fine now.

[05:04:01.0000] <Hixie>
ok.

[05:04:02.0000] <Hixie>
fine.

[05:05:00.0000] <Hixie>
whatever.

[05:05:01.0000] <Hixie>
i don't care.

[05:05:02.0000] <Hixie>
:-P

[05:06:00.0000] <BenMillard>
krijnh, this should be an IRC logs tagline: http://krijnhoetmer.nl/irc-logs/whatwg/20081120#l-283

[05:06:01.0000] <Philip`>
s/design/party/

[05:06:02.0000] <BenMillard>
/me forgot the keg...

[05:08:00.0000] <BenMillard>
Philip`, the ".html" or ".pdf" or ".png" and suchlike portion of URIs is already a widespread convention for selecting particular content types (re: http://krijnhoetmer.nl/irc-logs/whatwg/20081120#l-420)

[05:09:00.0000] <BenMillard>
Hixie, that convention is used outside of servers: browser extensions sometimes use RegExp on URIs in the DOM to figure out what type of thing they are pointing to, and give special behaviour for them

[05:09:01.0000] <BenMillard>
I'm not saying that's a good idea, though...

[05:10:00.0000] <mookid>
no that's terible and totally impractical -_-

[05:10:01.0000] <Philip`>
You should probably use <a href type> for that

[05:11:00.0000] <Philip`>
(if it's purely client-side)

[05:11:01.0000] <Philip`>
(and if you're in control of the content)

[05:11:02.0000] <mookid>
why type and not accept?

[05:11:03.0000] <BenMillard>
Philip`, the extensions are designed to work on the web :)

[05:11:04.0000] <mookid>
what are extensions?

[05:11:05.0000] <Philip`>
mookid: Because type is advisory information indicating the expected type of the thing at the URI

[05:12:00.0000] <mookid>
so.. the accept header?

[05:12:01.0000] <BenMillard>
mookid: AKA browser add-ons

[05:12:02.0000] <Philip`>
mookid: so it's quite similar to the file extension, in that it doesn't affect the request at all

[05:12:03.0000] <mookid>
but then you're just attempting a hybrid of http conneg and URI conneg

[05:12:04.0000] <Philip`>
BenMillard: Ah, fair enough :-)

[05:13:00.0000] <mookid>
if you're admitting there are cases where you would want to indicate what type a link should be

[05:13:01.0000] <Philip`>
mookid: It's not attempting any kind of conneg at all - it's just a way for an HTML document to tell clients what might be the content-type of a URI, so they can add "warning: this link is probably a PDF and may crash your browser" to the links or whatever

[05:13:02.0000] <mookid>
you're kind of producing the 'illusive' use case

[05:14:00.0000] <Philip`>
and it doesn't involve the server at all

[05:14:01.0000] <mookid>
that's what Accept does

[05:14:02.0000] <mookid>
no not as it stands

[05:14:03.0000] <mookid>
it could

[05:14:04.0000] <mookid>
if it was connected to the Accept request

[05:14:05.0000] <mookid>
of that request implied

[05:14:06.0000] <BenMillard>
Philip`, the one's I've tried are actually quite effective as file extensions seem almost ubiquitous and the main types you'd want special behaviour for are rarely in conflict with other types

[05:14:07.0000] <mookid>
by the link

[05:15:00.0000] <BenMillard>
Philip`, it also means the feature (such as grouping links by the type of thing they point to) works without sending loads of HEAD requests and waiting for the responses

[05:15:01.0000] <mookid>
Accept header of the request implied by the link*

[05:16:00.0000] <mookid>
it woudl be really cool if we could get to a stage where HTTP was encouraged to develop a standard way of indicating available content types

[05:16:01.0000] <mookid>
then if someone sends you a URL

[05:16:02.0000] <Philip`>
mookid: But for these particular use cases, there's no need to make it send the Accept header or anything, because it's just informing the client about stuff

[05:16:03.0000] <mookid>
you right click it and get a context-menu 'open with'

[05:16:04.0000] <mookid>
and it lists your UAs for the content types available

[05:16:05.0000] <mookid>
that would be sweet :)

[05:18:00.0000] <Philip`>
That does sound like it could be quite useful

[05:18:01.0000] <Philip`>
/me goes away for a while

[05:18:02.0000] <mookid>
yeah that will never be encouraged unless browsers encourage http conneg

[05:18:03.0000] <mookid>
it's chicken and egg

[05:19:00.0000] <mookid>
it's definitely a different way of thinking about URIs though

[05:19:01.0000] <mookid>
browsers dominate HTTP applications so unless it's provisioned there it wont happen

[05:20:00.0000] <Philip`>
You could implement it by e.g. right-click "open with" triggering an HTTP request with a "Tell-Me-About-Extra-Content-Types" header, which legacy servers ignore and they send the default type, but new servers respond with a list of content-types and the client sends an extra request to get the one it wants, or something

[05:20:01.0000] <mookid>
yeah

[05:20:02.0000] <mookid>
exactly

[05:21:00.0000] <Philip`>
(But then that's a purely HTTP thing, not HTML)

[05:21:01.0000] <mookid>
well I jsut mentioned above

[05:21:02.0000] <mookid>
brwosers dominate HTTP app dev

[05:21:03.0000] <mookid>
all the frameworks are designed for them

[05:21:04.0000] <mookid>
so they focus on using URIs

[05:22:00.0000] <BenMillard>
Philip` & mookid, sounds like you want the browser to trigger a 300 Multiple Choices response: http://tools.ietf.org/html/rfc2616#section-10.3.1

[05:22:01.0000] <mookid>
the whole process is way more oslid using protocol level conneg

[05:23:00.0000] <Philip`>
"If the server has a preferred choice of representation, it SHOULD include the specific URI for that representation in the Location field" - so you need a separate URI per representation?

[05:23:01.0000] <Philip`>
Anyway, I'm going away for a while :-p

[05:24:00.0000] <mookid>
Philip` ftw?

[05:24:01.0000] <BenMillard>
Philip`, I imagine it expects a filetypeless cononical URLs like /foo with a set of filetyped URLs like /foo.bar and /foo.baz and these are what get listed in the reponse it sends

[05:24:02.0000] <BenMillard>
s/cononical URLs/cononical URL/

[05:25:00.0000] <mookid>
or is it foo?type=bar

[05:25:01.0000] <mookid>
or foo?content-type=bar

[05:25:02.0000] <mookid>
foo;type=bar

[05:25:03.0000] <BenMillard>
mookid, although which identifies the type of the resource, yeah

[05:26:00.0000] <BenMillard>
s/although/anything/

[05:26:01.0000] <BenMillard>
/me wonders how that typo happened...

[05:26:02.0000] <mookid>
yeah it's just creating ambiguity and resource blaot where its not necessary

[05:26:03.0000] <mookid>
bloat^

[05:26:04.0000] <mookid>
HTTP handles this nicely

[05:26:05.0000] <BenMillard>
huh? that *is* the HTTP spec :|

[05:26:06.0000] <BenMillard>
well, cya

[05:27:00.0000] <mookid>
well the http spec actually mentions both methods

[05:27:01.0000] <mookid>
oh, bye

[05:27:02.0000] <mookid>
ABORT ABORT!

[05:27:03.0000] <mookid>
:>

[05:38:00.0000] <mookid>
another key point here is that if users are using PUT to update resources.. how are they to know wheher PUT /document.pdf updates /document.html and /document.xml - it's ambiguous

[05:39:00.0000] <mookid>
if the URI is the same this is completely clear

[05:39:01.0000] <mookid>
added to which - as soon as the PUT is performed on that URI - the caches can pick up on this and react for all representations to be re-cached

[05:40:00.0000] <mookid>
this isn't possible if the URIs are different per resource

[05:40:01.0000] <mookid>
yoru caching mechanism has to be application aware

[05:43:00.0000] <mookid>
^ this is just an example of how doing the conneg in HTTP makes significant changes to development approach

[05:45:00.0000] <Hixie>
wow, httpbis is already up to 300 pages

[05:46:00.0000] <Hixie>
http 1.1 was only 176

[05:46:01.0000] <Hixie>
i thought http bis wasn't adding anything new?

[05:46:02.0000] <mookid>
I dont know I'm not involved with that right now

[05:46:03.0000] <mookid>
how far away is that?

[05:47:00.0000] <mookid>
Hixie: do you even accept that it is different and has benefits?

[05:48:00.0000] <mookid>
I'm not getting at you it just frustrates me that I can articulate it properly :)

[05:48:01.0000] <mookid>
can't^

[05:49:00.0000] <mookid>
I'll take that as a yes then.. -_-

[05:53:00.0000] <ehird>
wow

[05:53:01.0000] <ehird>
mookid is STILL on about conneg?!?!?!

[05:53:02.0000] <ehird>
:o

[05:53:03.0000] <jcranmer>
ehird: it's only been >48 hours

[06:03:00.0000] <mookid>
well I don't think it's as cut and dry as you seem to believe it is

[06:04:00.0000] <mookid>
other than the fact that there's no way I can get this past the protector of the internet

[06:17:00.0000] <takkaria>
PROTECTOR OF THE INTERNETES

[06:18:00.0000] <krijnh>
/me adds a new subtitle to his logs

[06:19:00.0000] <mookid>
he talked to me

[06:19:01.0000] <mookid>
I am blessed

[06:19:02.0000] <mookid>
thanks be to Neo

[06:21:00.0000] <mookid>
don't tell him about javascript it'll break his little heart

[06:21:01.0000] <mookid>
:P

[06:23:00.0000] <krijnh>
It'll wake him up now, mostly :)

[06:35:00.0000] <Dashiva>
ehird: This is nothing compared to alt, alt 2: the return of alt, alt 3: son of alt, and all the others yet to come

[06:35:01.0000] <ehird>
:D

[06:36:00.0000] <ehird>
who's gonna by connegcube.com

[06:36:01.0000] <ehird>
4-day harmonic content negotiation

[06:38:00.0000] <mookid>
lies!

[06:38:01.0000] <mookid>
what a dissapointment that was

[06:40:00.0000] <Dashiva>
ehird: Site doesn't work :<

[06:41:00.0000] <ehird>
*buy

[06:41:01.0000] <ehird>
well duh i was asking for someone to make it

[06:41:02.0000] <ehird>
:D

[06:45:00.0000] <Dashiva>
Oh snap, new last week

[06:46:00.0000] <annevk2>
poor quality lately

[07:17:00.0000] <Philip`>
"Mr last Week sez: ... tone down the insults and rhetoric, it is uncalled for" - ooh, irony

[07:37:00.0000] <webben_>
What's the current thought on how should inline stage directions/descriptions be handled in DIALOG?

[07:50:00.0000] <webben_>
or multiblock speeches...

[08:12:00.0000] <mookid>
Philip`: did you have any more thoughts on it?

[08:13:00.0000] <Philip`>
mookid: On what in particular?

[08:18:00.0000] <mookid>
whether there is a significant difference in HTTP conneg, and whether it's provision is compatible with conneg in URI

[08:37:00.0000] <Philip`>
mookid: I can't think of any more thoughts at the moment

[08:39:00.0000] <mookid>
ok just chicken

[08:58:00.0000] <Philip`>
Oh no, #html-wg got invaded

[08:59:00.0000] <Philip`>
hsivonen: Continuing my comment from there: (and making the spec intentionally complex and obscure in order to discourage people writing their own probably-buggy XML parsers when they should be using a library instead, seems kind of elitist (that's not quite the right term but I can't think of the right one))

[09:00:00.0000] <Philip`>
Wow, someone just made Gmail ugly

[09:00:01.0000] <Philip`>
Hmm, it's a bit better if I reload the whole page and it's not mixing the old and new themes...

[09:02:00.0000] <gavin>
omgchange!

[09:03:00.0000] <Philip`>
Hmm, the layout is still different, and therefore ugly

[09:04:00.0000] <Philip`>
Judging by what happened when they last upgraded everything, I will be terribly upset for about two days and then I will be used to it and it'll seem perfectly normal

[09:04:01.0000] <gavin>
heh, indeed

[09:12:00.0000] <Philip`>
Hmph, Jim Jewett seems to think my name on IRC is a typo :-(

[09:13:00.0000] <Philip`>
(when actually it's just me being far too unimaginative to come up with a decent name, and someone else having already registered "Philip" on this network)

[09:19:00.0000] <mookid>
imagine if caching was controlled just with the Vary header

[09:19:01.0000] <mookid>
that would be fricking awesome

[09:20:00.0000] <Philip`>
But impossible

[09:20:01.0000] <mookid>
not for new apps

[09:20:02.0000] <mookid>
if your GUI and your API are the same interface

[09:20:03.0000] <mookid>
one being the HTML representation and the other being JSON or XML

[09:21:00.0000] <Philip`>
If you have multiple formats for some resource, but it takes a few minutes to do the conversion (because it's expensive transcoding or OCRing or whatever), how would you handle cache invalidation once the new format has been generated?

[09:21:01.0000] <mookid>
gah? -_-

[09:23:00.0000] <Philip`>
Nicely-designed apps shouldn't block users who are uploading data while the server does a load of back-end computation - asynchrony is more user-friendly, but doesn't seem to interact well with the assumptions made by caches

[09:23:01.0000] <mookid>
why would that be the case? yuo don't have to cache the entire of your application

[09:24:00.0000] <mookid>
you should only be caching what's appropriate to cache. -_-

[09:24:01.0000] <mookid>
you can control that with your responses

[09:41:00.0000] <Dashiva>
No worries about keeping warm this winter as long as public-html stays as it is. :)

[09:43:00.0000] <Philip`>
/me wonders why Textile parses "@,@ _x_@x@" correctly, but not "@, @_x_@x@"

[09:47:00.0000] <Dashiva>
The former looks like two smileys on a seesaw, the latter looks like some weird mutant and something

[09:49:00.0000] <Philip`>
The latter parses into <code>, @&lt;em&gt;x&lt;/em&gt;@x</code>, which surely can't be intentional

[09:59:00.0000] <Dashiva>
Philip`, you just got told

[10:00:00.0000] <Philip`>
Indeedily

[10:01:00.0000] <Dashiva>
Anything you would like to tell our viewers?

[10:02:00.0000] <Philip`>
No

[10:03:00.0000] <Dashiva>
What will you do next?

[10:03:01.0000] <Philip`>
Nobody knows - that's how crazy I am

[10:04:00.0000] <Philip`>
so you better keep watching!

[10:06:00.0000] <Dashiva>
You heard him. Don't change that channel!

[10:37:00.0000] <gsnedders>
Philip` pwn'd on last week :)

[11:47:00.0000] <yecril71>
Philip`, could you look at my test page in IE8 today?

[11:47:01.0000] <yecril71>
Is there a chance for the section headers in HTML5 show up in IE7?

[11:48:00.0000] <yecril71>
It is very inconvenient that they do not

[11:48:01.0000] <yecril71>
You do not know what you are reading about

[11:52:00.0000] <yecril71>
Where is OBJECT[classid] declared?

[11:53:00.0000] <yecril71>
I cannot find it in the green summary box

[11:53:01.0000] <gsnedders>
yecril71: it's non-conforming, so nowhere

[11:55:00.0000] <yecril71>
But it is still mentioned in the text

[11:55:01.0000] <yecril71>
"If the classid attribute is present, and has a value that isn't the empty string,"

[11:56:00.0000] <gsnedders>
It has defined processing for UAs

[11:56:01.0000] <gsnedders>
Just because it has defined processing doesn't make it conforming

[11:56:02.0000] <yecril71>
Was that present perfect, or present simple?

[11:57:00.0000] <gsnedders>
present simple

[11:58:00.0000] <yecril71>
Thx

[11:58:01.0000] <gsnedders>
/me should be less ambiguous

[11:59:00.0000] <takkaria>
can someone explain to me the difference? :)

[12:00:00.0000] <gsnedders>
takkaria: Yes

[12:00:01.0000] <yecril71>
Present simple is the verb itself, inflected, followed by the object.

[12:01:00.0000] <yecril71>
Present perfect is aux "have" + past participle, followed by the object.

[12:01:01.0000] <Dashiva>
The ambiguity lies in whether defined is used as a participle or as an adjective

[12:02:00.0000] <gsnedders>
What can I put on my CV about WHATWG?

[12:02:01.0000] <Dashiva>
"Official inner-circle teamster"

[12:03:00.0000] <yecril71>
What should happen to embeeded ActiveX objects then?

[12:03:01.0000] <gsnedders>
Dashiva: Not a fanboy.

[12:03:02.0000] <yecril71>
I understand type="application/x-oleobject", but what about the CLSID?

[12:04:00.0000] <yecril71>
Where do I put it, given that classid is out?

[12:05:00.0000] <yecril71>
s/embeeded/embedded/

[12:08:00.0000] <yecril71>
classid is not mentioned as a difference from HTML4

[12:08:01.0000] <gsnedders>
Then the diff dsoc is wrong :)

[12:08:02.0000] <gsnedders>
*doc

[12:09:00.0000] <yecril71>
I would be glad to update it, but I need some information on how the decision was taken.

[12:09:01.0000] <annevk2>
feel free to blame me or submit patches to some mailing list I read

[12:10:00.0000] <yecril71>
I can do it myself, patches are excessive

[12:10:01.0000] <annevk2>
I actually think classid is still being considered for inclusion, using some hardcoded mapping table of values and plugins

[12:10:02.0000] <Philip`>
yecril71: Embedded ActiveX objects are not at all interoperable between platforms, so it wouldn't make much sense for HTML5 to provide specific syntax for them

[12:11:00.0000] <Philip`>
The conforming thing to do is to not use ActiveX :-)

[12:11:01.0000] <Philip`>
yecril71: I could look at your page, if you remind me where it is

[12:11:02.0000] <annevk2>
Philip`, you could make it interoperable if you map the classid to NPAPI

[12:12:00.0000] <yecril71>
<http://www.2a.pl/~ne01026/test.htm>

[12:12:01.0000] <Philip`>
How odd, IE8 can't even render Google search result pages non-buggily :-/

[12:12:02.0000] <gsnedders>
hmmm

[12:13:00.0000] <gsnedders>
logs missed the link to jgraham='s CV

[12:13:01.0000] <Philip`>
(It cut off after 1.5 results (but still showed the footer after that), until I selected some text and then the rest popped into existence)

[12:13:02.0000] <jgraham>
gsnedders: I didn' link to my CV :)

[12:13:03.0000] <gsnedders>
jgraham: Didn't you?

[12:13:04.0000] <yecril71>
That would be good, talking about an undefined attribute is rather strange

[12:14:00.0000] <yecril71>
It is better to say classid (deprecated) than nothing at all.

[12:15:00.0000] <yecril71>
It is possible not to use ActiveX and delegate all processing to server,

[12:15:01.0000] <annevk2>
we've taken the approach to only mention things when there's an effect or when they're allowed for convenience

[12:15:02.0000] <yecril71>
or to use application/java.

[12:15:03.0000] <annevk2>
currently neither is the case as far as HTML5 is concerned

[12:15:04.0000] <annevk2>
though as I said that might be changing

[12:15:05.0000] <yecril71>
classid has a defined effect, only itself is undefined.

[12:16:00.0000] <Philip`>
yecril71: Both columns look the same (except for the left one being narrower), and about the same as in Opera and Firefox - there's an "a" slightly above (overlapping) the horizontal line, and a "b" slightly below (also overlapping) the line

[12:16:01.0000] <yecril71>
That is rather peculiar.

[12:16:02.0000] <yecril71>
That is good.

[12:16:03.0000] <annevk2>
UAs that would let it have effect would  be non conforming per HTML5 as it stands today

[12:16:04.0000] <yecril71>
IE7 is broken in that regard.

[12:16:05.0000] <yecril71>
Thx a lot.

[12:17:00.0000] <Philip`>
yecril71: IE8 in IE7 bug compatibility mode is very different - there's no overlapping, and the left column has an extra blank line before the "b", which I assume is what IE7 does

[12:17:01.0000] <yecril71>
So it looks like, although it is not a blank line at all.

[12:18:00.0000] <Philip`>
Well, it's a line that's blank ;-)

[12:18:01.0000] <yecril71>
No, it is vertical space.

[12:18:02.0000] <yecril71>
A line is something, a space is lack of something ;-)

[12:19:00.0000] <yecril71>
The effect is described in item 1 of the embedding algorithm in HTML5.

[12:19:01.0000] <yecril71>
(of classid).

[12:20:00.0000] <yecril71>
So it takes precedence over anything else.

[12:20:01.0000] <yecril71>
And yet classid is undefined.

[12:21:00.0000] <yecril71>
The effect is conforming, only the attribute itself is nonconforming.

[12:22:00.0000] <yecril71>
ActiveX objects can be advantageous in an in-house application.

[12:22:01.0000] <annevk2>
ah, classid is already mentioned for browsers? ok

[12:22:02.0000] <Philip`>
<b>x<i>y</b>z</i> is non-conforming, but there are conformance requirements for how browsers must handle that, so it's not really any different with these nonconforming attributes

[12:22:03.0000] <yecril71>
The alternatives of using in-page java or a server-side process can be too expensive.

[12:23:00.0000] <gsnedders>
annevk2: Can we have more standards suckage?

[12:23:01.0000] <yecril71>
Except that B and I are defined and do not jump at the reader out of nowhere.

[12:24:00.0000] <yecril71>
And yet it would be nice to be able to validate the pages the components reside in.

[12:24:01.0000] <yecril71>
Without customizing the validator.

[12:25:00.0000] <yecril71>
Of course, one can always use HTML4 for that purpose,

[12:25:01.0000] <yecril71>
but what about HTML5?

[12:25:02.0000] <yecril71>
Is there a chance of getting a hint?

[12:25:03.0000] <yecril71>
(of course, IE *requires* classid now;

[12:26:00.0000] <yecril71>
assuming we can change it, what should it be changed to?)

[12:26:01.0000] <gsnedders>
yecril71: Nothing is as bad as the image element

[12:26:02.0000] <gsnedders>
(it is mentioned once, in the parser, where it gets converted to img)

[12:28:00.0000] <yecril71>
How about type="application/x-oleobject;classid={G-U-I-D}"?

[12:33:00.0000] <annevk2>
gsnedders, we'll resume shortly; not exactly sure when

[12:46:00.0000] <roc>
it's sad that some of the top-crash bugs we're seeing are caused by trojans

[12:49:00.0000] <takkaria>
ah, that sucks :(

[12:49:01.0000] <takkaria>
but I guess it speaks for the stability of your product

[12:50:00.0000] <roc>
I think for both us and Webkit the majority of top-crash bugs are actually Flash-related

[13:02:00.0000] <yecril71>
I do not understand how Tom Broyer got the need to have separate URLs for representations.

[13:03:00.0000] <Dashiva>
Maybe he doesn't agree they're representations?

[13:03:01.0000] <yecril71>
It does not follow from Fielding�s quote.

[13:06:00.0000] <hsivonen>
whose idea was it to start calling them URIs instead of URLs?

[13:11:00.0000] <yecril71>
Seems like OP�s.

[13:13:00.0000] <yecril71>
Although Fielding uses the term "identifier", so that misuse can be probably traced to the boss.

[13:14:00.0000] <yecril71>
s/misuse/unneeded generalization/

[13:21:00.0000] <ehird>
eh, it makes sense

[13:21:01.0000] <ehird>
how do you locate <tel:...>

[13:43:00.0000] <yecril71>
What is <tel:>?

[13:44:00.0000] <yecril71>
And does it implement content negotiation?

[13:47:00.0000] <gsnedders>
yecril71: tel is a uri scheme for telephone numbers

[13:57:00.0000] <takkaria>
erk.  I joined the debate 

[13:57:01.0000] <takkaria>
ehird: you locate it by phoning it, surely?

[13:57:02.0000] <takkaria>
oh, right

[13:57:03.0000] <takkaria>
I missed the context. :)

[13:58:00.0000] <yecril71>
And what about callto:?

[13:58:01.0000] <ehird>
ok then, how do you locate isbn::

[13:58:02.0000] <ehird>
("going to a bookstore and buying it" is not an answer)

[13:58:03.0000] <yecril71>
Wikipedia has a tool for that

[13:59:00.0000] <yecril71>
And isbn: does not support content negotiation either.

[13:59:01.0000] <ehird>
what has content negotiation got to do with this

[14:00:00.0000] <yecril71>
It is being asked in the context of content negotiation.

[14:00:01.0000] <takkaria>
I missed the context

[14:00:02.0000] <yecril71>
Why content negotiation is advertised for URIs in general.

[14:00:03.0000] <takkaria>
I thought that it was a "what am I meant to do with a tel:" number question, not a URI vs URL question

[14:01:00.0000] <takkaria>
URL means Universal Republic of Love, anyway, Tim Bray told me so :)

[14:01:01.0000] <yecril71>
wanna cyber? ;-)

[14:01:02.0000] <takkaria>
/me chortles

[14:02:00.0000] <yecril71>
Is tel: anything official?

[14:02:01.0000] <yecril71>
I have never seen tel:, only callto:

[14:03:00.0000] <takkaria>
I've seen tel: here and there

[14:03:01.0000] <yecril71>
Seems like by mistake

[14:03:02.0000] <yecril71>
Does ekiga register for tel:?

[14:06:00.0000] <ehird>
yecril71: timbl has it in his foaf file

[14:06:01.0000] <ehird>
somehow i doubt a mistake.

[14:17:00.0000] <gsnedders>
yecril71: http://www.iana.org/assignments/uri-schemes.html

[14:17:01.0000] <Philip`>
yecril71: RFC 3966

[14:17:02.0000] <gsnedders>
yecril71: Ns such thing as callto registered

[14:18:00.0000] <gsnedders>
/me returns to Shakespeare

[14:20:00.0000] <yecril71>
I only wanted information, not a proof :-)

[14:20:01.0000] <yecril71>
Seems the people at Skype are illiterate


2008-11-21
[16:01:00.0000] <jgraham>
http://james.html5.org/tables/table_inspector.html has some probably-horribly broken implementation of the HTML5 headers algorithm as it is currently in the spec. Note that the URL structure has changed a little so old links won't work 'till I set up a redirect

[16:01:01.0000] <jgraham>
I will fix bugs some other day 

[16:50:00.0000] <Lachy>
wow, isn't it nice when people do a little research and come up with the same result that others have tried to point out all along! http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2008-November/017367.html

[16:52:00.0000] <Dashiva>
Lachy: We should all be happy it ended well

[16:52:01.0000] <Lachy>
I am

[17:14:00.0000] <Hixie>
yeah i was impressed by that guy's humbleness

[17:14:01.0000] <Hixie>
we could all learn from that :-)

[17:19:00.0000] <Philip`>
The only problem is he seems to be missing the people who say <link rel="made" href="Mr A. Square"> and <link rel="made" href="asquare⊙ec"> :-)

[17:19:01.0000] <Philip`>
Uh

[17:19:02.0000] <Philip`>
s/rel/rev/

[22:26:00.0000] <hsivonen>
Philip`: the parsing spec is not complex for the sake of complexity

[22:26:01.0000] <hsivonen>
Philip`: would it be elitist to suggest that some Web authors can't program at all and, therefore, won't be writing parsers?

[22:28:00.0000] <hsivonen>
Philip`: I think the current way of speccing it is less elitist that e.g. one that required readers to grok lex/yacc/ocaml

[23:29:00.0000] <hsivonen>
http://lists.w3.org/Archives/Member/tag/2008Oct/0126.html (Member-only link)

[23:34:00.0000] <hsivonen>
http://lists.w3.org/Archives/Member/tag/2008Oct/0135.html

[23:38:00.0000] <annevk2>
o_O

[23:46:00.0000] <hsivonen>
Tantek's troll taxonomy lacks an anti-browser troll. http://tantek.pbwiki.com/TrollTaxonomy

[23:47:00.0000] <roc>
did everyone see that IE8's having a third beta Q12009, with a final release some time after that?

[23:48:00.0000] <hsivonen>
/me had not seen the news

[23:49:00.0000] <annevk2>
I didn't

[23:50:00.0000] <annevk2>
/me thought IE8 would come during Christmas

[23:54:00.0000] <roc>
that had been the plan

[23:57:00.0000] <annevk2>
I wonder why Gmail now has a fucked up color scheme

[23:58:00.0000] <roc>
it actually has several to choose from

[23:58:01.0000] <roc>
remember it's still in beta

[23:59:00.0000] <hsivonen>
can anyone remember if scripts inside objects and applets run?

[23:59:01.0000] <hsivonen>
(I expect them to run)

[00:05:00.0000] <zcorpan>
hsivonen: yep they do

[00:05:01.0000] <hsivonen>
thanks

[00:06:00.0000] <zcorpan>
but ie is different

[00:08:00.0000] <annevk2>
roc, beta, hah :)

[00:08:01.0000] <annevk2>
where can I pick the default one?

[00:09:00.0000] <roc>
correction: the Q12009 IE8 release will be a "release candidate", not a "beta"

[00:10:00.0000] <annevk2>
ah, themes

[00:10:01.0000] <hsivonen>
/me wonders if Release Candidate in MS vocabulary means potentially final compiler output

[00:10:02.0000] <roc>
doesn't sound like it

[00:10:03.0000] <annevk2>
/me sets it back to classic

[00:11:00.0000] <hsivonen>
/me finds that and XLink-aware tree builder may have to stop the parser

[00:12:00.0000] <hsivonen>
s/and/an/

[00:16:00.0000] <hsivonen>
Hixie: is HTML5 supposed to allow XLink autolinks?

[00:17:00.0000] <annevk2>
http://www.tbray.org/ongoing/When/200x/2008/11/20/HTML-Pain

[00:19:00.0000] <annevk2>
seems that Tim Bray hasn't really done his homework

[00:19:01.0000] <hsivonen>
yeah

[00:20:00.0000] <annevk2>
because for half of the things he likes to do the "language specification" would be utterly unusable

[00:22:00.0000] <roc>
Are people actually opposed to the existence of Mike Smith's document?

[00:25:00.0000] <hsivonen>
roc: I'm not opposed to existence but I am opposed to existence on the REC track (where some people want to put it).

[00:25:01.0000] <annevk2>
I've seen some people being opposed; I'm not exactly sure why

[00:26:00.0000] <annevk2>
I don't think it should be normative given that then it would duplicate information of another normative specification, but I suppose I could live with it

[00:26:01.0000] <roc>
yeah it shouldn't be normative

[00:27:00.0000] <roc>
but otherwise it seems like a great thing

[00:31:00.0000] <hsivonen>
roc: It's great as a reference for people who like RELAX NG Compact Syntax

[00:31:01.0000] <annevk2>
I bet that if we wrote HTML5 in RDF everyone could easily get their own version they like using SPARQL

[00:31:02.0000] <roc>
it's not safe to even joke about such things

[00:32:00.0000] <hsivonen>
annevk2: see ARIA

[00:32:01.0000] <hsivonen>
(I like RELAX NG Compact Syntax)

[00:33:00.0000] <hsivonen>
(not as normative spec, though, because it's not sufficient and it's too easy for people to ignore the parts that don't get expressed as RELAX NG Compact Syntax)

[01:05:00.0000] <annevk2>
"On our way back the TSA told me I couldn't keep my apple juice. I asked if I could drink it first. They said only if I went back outside. I asked if they meant that my apple juice was a danger in the security area but safe outside, and they confirmed that that was the case." hahaha

[02:32:00.0000] <jgraham>
BenMillard: Did you see that I have some new table inspector stuff for you to play with?

[02:35:00.0000] <BenMillard>
jgraham, yeah that's why I arrived here :)

[02:35:01.0000] <BenMillard>
why the change URL structure? are the costs really worth the gains?

[02:36:00.0000] <jgraham>
BenMillard:I set up a redirect

[02:36:01.0000] <jgraham>
So old links should work

[02:36:02.0000] <jgraham>
It is much easier to mintain with the new structure; no need to move stuff around on the server

[02:37:00.0000] <BenMillard>
oh, ok

[02:38:00.0000] <BenMillard>
I've tried a few old links and they work, but seem slow

[02:40:00.0000] <BenMillard>
jgraham, could you generate a link on the output page which points to where the table came from when inspecting via URI?

[02:40:01.0000] <jgraham>
It is possible that I could do better using symlinks or something. However if you can edit the old links, just remove the /cgi-bin/

[02:40:02.0000] <jgraham>
BenMillard: Sure

[02:41:00.0000] <BenMillard>
it seems the slowness of the links I tried was due to the complexity of the table being inspected :)

[02:41:01.0000] <BenMillard>
jgraham, oh that'll be easy enough to change on my blog...can't change archived e-mails though

[02:44:00.0000] <BenMillard>
jgraham, I take it the isHeading check means <td scope> isn't treated as a header cell in your implementation of Smart Headers? (http://code.google.com/p/html5/source/browse/trunk/tables/lib/headers/smartheaders.py#54)

[02:48:00.0000] <jgraham>
BenMillard: <td scope> isn't

[02:48:01.0000] <BenMillard>
jgraham, thanks

[02:49:00.0000] <BenMillard>
jgraham, when clicking cells in the output, they become colour-coded which lets you see what the associations are. Because that's so effective, could you avoid showing the text of the associated header cells?

[02:49:01.0000] <jgraham>
BTW I moved the source to http://hg.jgraham.webfactional.com/hgwebdir.cgi/html5_tables/

[02:49:02.0000] <BenMillard>
clicking through the cells to view associations wouldn't make the table's layout jump around

[02:49:03.0000] <jgraham>
I can resync that with the google code project if that is important to anyone

[02:52:00.0000] <BenMillard>
jgraham, the new location is fine by me

[02:57:00.0000] <Lachy>
ooh, Acid 3 changed again http://lists.w3.org/Archives/Public/www-archive/2008Nov/0015.html

[03:02:00.0000] <jgraham>
BenMillard: Changes made. Time to do other stuff for a bit :) I will try to add experimental features some other time

[03:03:00.0000] <BenMillard>
jgraham, thanks, that works great!

[03:04:00.0000] <BenMillard>
hsivonen, the IE8 release schedule (and feedback, bug fixing, hints at when code freeze happens, etc) is mentioned here: http://blogs.msdn.com/ie/archive/2008/11/19/ie8-what-s-after-beta-2.aspx

[03:25:00.0000] <hsivonen>
BenMillard: yeah, it indeed looks like release candidate doesn't mean that the last release candidate and the release are identical except for the version string (byte-for-byte same DLLs)

[03:37:00.0000] <BenMillard>
jgraham, I get a error message for Wikipedia pages which load in the browser: "The URI http://en.wikipedia.org/wiki/Ferrari_Testarossa could not be loaded." (http://james.html5.org/tables/table_inspector.py?input_type=type_uri&uri=http%3A%2F%2Fen.wikipedia.org%2Fwiki%2FFerrari_Testarossa&source=&algorithm=smartheaders)

[03:43:00.0000] <hsivonen>
is this intranet stuff? http://www.w3.org/TR/webcgm21/

[03:43:01.0000] <hsivonen>
I don't recall running into CGM content on the Web

[03:46:00.0000] <jgraham>
BenMillard: Curious

[03:51:00.0000] <jgraham>
BenMillard: Wikipedia is returning a 403. I guess it doesn't like the python UA string

[03:54:00.0000] <BenMillard>
jgraham, yeah they seem to take robots pretty seriously: http://en.wikipedia.org/robots.txt

[03:56:00.0000] <BenMillard>
jgraham, copying and pasting the timeline from that page gives me a "Request-URI Too Large"

[03:57:00.0000] <BenMillard>
it used to work, which is why I included that example...they must have started blocking tools like yours quite recently :(

[03:59:00.0000] <jgraham>
I guess at some point I could add a file upload option which would sort of solve the problem

[04:02:00.0000] <BenMillard>
jgraham, I consider this Wikipedia's problem rather than yours. :)

[04:02:01.0000] <BenMillard>
I get a big error page for this: http://james.html5.org/tables/table_inspector.py?input_type=type_uri&uri=http%3A%2F%2Fprojectcerbera.com%2Fblog%2F2008%2F09%2Funtangle&source=&algorithm=html5b

[04:03:00.0000] <BenMillard>
says "<type 'exceptions.IndexError'>: list index out of range" at the bottom

[04:03:01.0000] <BenMillard>
(original table here: http://projectcerbera.com/blog/2008/09/untangle)

[04:04:00.0000] <krijnh>
Any Opera dev in here?

[04:05:00.0000] <krijnh>
Need an explanation for http://krijnhoetmer.nl/zooi/screenshots/test-1.png :)

[04:05:01.0000] <takkaria>
hsivonen: I've seen CGM->bitmap converters before, but not for years

[04:05:02.0000] <hsivonen>
takkaria: in Web clients?

[04:07:00.0000] <takkaria>
no, just things like imagemagick

[04:10:00.0000] <MikeSmith>
hsivonen: my understanding is that WebCGM is uses in the airline industry, and perhaps also aerospace. but I don't really know the details 

[04:19:00.0000] <Lachy>
krijnh, which version of Opera is that?

[04:19:01.0000] <krijnh>
9.62

[04:19:02.0000] <krijnh>
Just updated and restarted

[04:20:00.0000] <Lachy>
ok, I'll test it and file a bug

[04:20:01.0000] <krijnh>
Well

[04:20:02.0000] <krijnh>
You can't test it

[04:20:03.0000] <Lachy>
what?

[04:20:04.0000] <krijnh>
You'll get okay results

[04:20:05.0000] <krijnh>
I've got a script from a different site

[04:21:00.0000] <krijnh>
Which is applied to _all_ other tabs as well

[04:21:01.0000] <krijnh>
Which sets the height of #content to the viewport height

[04:21:02.0000] <krijnh>
Pretty silly stuff :)

[04:21:03.0000] <Lachy>
oh, I thought that was the bug

[04:21:04.0000] <Lachy>
so what's the problem then?

[04:21:05.0000] <krijnh>
That is the bug, I guess

[04:21:06.0000] <krijnh>
Err

[04:21:07.0000] <krijnh>
http://krijnhoetmer.nl/zooi/screenshots/test-2.png

[04:22:00.0000] <krijnh>
I don't think a cached script should apply to all sites you visit

[04:22:01.0000] <krijnh>
(Unless it's a UserJS, which it is not)

[04:22:02.0000] <Lachy>
no, it shouldn't. Can you tell me how to reproduce it, or set up a page on which I can?

[04:22:03.0000] <krijnh>
I have no idea how I did it..

[04:24:00.0000] <Lachy>
well, then I can't really help you, because a screenshot doesn't give enough diagnostic information

[04:24:01.0000] <krijnh>
I know..

[04:24:02.0000] <krijnh>
I'm trying to figure out what's happening

[04:24:03.0000] <Lachy>
which page is the script on, which is applying to other tabs?

[04:25:00.0000] <krijnh>
http://krijnhoetmer.nl/zooi/screenshots/test-3.png (where it is run twice now :)

[04:25:01.0000] <krijnh>
I think it's http://wijndomein.ah.nl/_js/all.min.js

[04:25:02.0000] <krijnh>
Could it be the length that messes things up?

[04:26:00.0000] <Lachy>
what's the build number of your Opera?

[04:26:01.0000] <krijnh>
10467

[04:26:02.0000] <krijnh>
Just downloaded

[04:26:03.0000] <krijnh>
I've had the same issue with 9.52

[04:27:00.0000] <krijnh>
This is the second time I run into it btw, so it must be reproducable

[04:29:00.0000] <krijnh>
Hmm, clearing my cache doesn't solve it

[04:30:00.0000] <krijnh>
Restarting Opera neither

[04:37:00.0000] <Philip`>
You could hibernate your computer and then ship it to Opera to have someone debug it

[04:39:00.0000] <krijnh>
After clearing the cache, opera:cache should be empty, right?

[04:41:00.0000] <Lachy>
krijnh, if you backup and create a clean profile, does the problem still occur?

[04:55:00.0000] <krijnh>
Just turning off JS helps

[04:57:00.0000] <hsivonen>
/me wonders if krijnh's site has been targeted by browser.js

[04:57:01.0000] <krijnh>
It's not only my site

[04:59:00.0000] <ehird>
new gmail is awesome

[05:00:00.0000] <Philip`>
/me wonders if being targeted by browser.js should count as an honour that you're considered important enough to be worth it, or a disgrace that your site is so broken it can only be fixed via ugly hacks

[05:01:00.0000] <krijnh>
Cool, I get JS errors as well

[05:02:00.0000] <annevk2>
krijnh, maybe e-mail hallvord⊙oc with ah.nl in the subject line and a short description of the issue?

[05:02:01.0000] <krijnh>
hallvors is in here as well

[05:03:00.0000] <krijnh>
lol

[05:03:01.0000] <annevk2>
true, I believe he's not a fan of IRC though

[05:03:02.0000] <krijnh>
I'm a moron

[05:03:03.0000] <krijnh>
"Line 1 of User JS script"

[05:03:04.0000] <krijnh>
:]

[05:03:05.0000] <hsivonen>
ah. targeted by user.js then. :-)

[05:04:00.0000] <krijnh>
Not targeted

[05:04:01.0000] <hsivonen>
Philip`: perhaps browser.js hacks are beautiful

[05:04:02.0000] <krijnh>
/me checks his userjs dir..

[05:04:03.0000] <hsivonen>
well, collateral damage of user.js :-)

[05:04:04.0000] <krijnh>
Weeh, my Desktop

[05:04:05.0000] <krijnh>
Brilliant

[05:05:00.0000] <krijnh>
Fixed, sorry for wasting your time :)

[05:05:01.0000] <Lachy>
/me wonders what "Weeh, my Desktop" means

[05:05:02.0000] <krijnh>
Apparently I had a .js on my desktop

[05:05:03.0000] <Lachy>
/me will send krijnh a bill for the time spent

[05:05:04.0000] <krijnh>
Which got applied to every site

[05:05:05.0000] <krijnh>
Yeah, that's okay

[05:05:06.0000] <Philip`>
hsivonen: Having looked at browser.js, I don't think they are :-)

[05:06:00.0000] <Lachy>
wtf? Having a JS on your desktop shouldn't be used by the browser

[05:06:01.0000] <Lachy>
at least, I hope it doesn't

[05:06:02.0000] <krijnh>
It is if your user js dir is your desktop

[05:06:03.0000] <krijnh>
Hence the moron part

[05:06:04.0000] <Lachy>
oh. Why would you set your user js dir to your desktop?

[05:06:05.0000] <krijnh>
I think I've tried it out once

[05:07:00.0000] <Lachy>
That's just crazy and leaves you wide open for various kinds of attacks

[05:07:01.0000] <Philip`>
If you set your User JS directory to be the same as your cache directory, would that make every script from every page you've ever visited be applied to every page you visit in the future?

[05:07:02.0000] <krijnh>
Yes

[05:07:03.0000] <Lachy>
LOL

[05:08:00.0000] <krijnh>
Apparently it works without providing // ==UserScript== stuff in your js

[05:08:01.0000] <krijnh>
Without // @include stuff

[05:08:02.0000] <krijnh>
That's just an Opera bug, imho..

[05:08:03.0000] <krijnh>
O:)

[05:16:00.0000] <krijnh>
/me doesn't get User JS

[06:03:00.0000] <Lachy>
JohnResig, yt?

[06:13:00.0000] <BenMillard>
this basically seems to describe what already happens in that the parts which are agreed are the parts which get experimental implementations, iterating and growing over time: http://intertwingly.net/blog/2008/11/20/Half-Full

[06:18:00.0000] <BenMillard>
and the annotation system allows the maturity level of HTML5 to be labelled with section-by-section granularity

[06:24:00.0000] <mookid>
hi fans

[06:29:00.0000] <mookid>
Philip`: did you do anymore thinking? :))

[07:05:00.0000] <Philip`>
mookid: None that I can remember

[07:09:00.0000] <mookid>
meh; that email exchange was a bit of a let down

[07:21:00.0000] <hallvors>
/me thinks of the quote "removing options is evil".

[07:45:00.0000] <hsivonen>
the document organization discussion is a total bikeshed

[07:59:00.0000] <gsnedders>
Philip`: You gonna be around Cam in early Dec?

[07:59:01.0000] <gsnedders>
Like, around the 9th?

[07:59:02.0000] <Philip`>
gsnedders: Yes

[08:00:00.0000] <gsnedders>
Philip`: Want to have dinner on 8th/9th?

[08:01:00.0000] <gsnedders>
/me notes he has to fill out a form saying when he wants accommodation and when he wants meals in college

[08:01:01.0000] <gsnedders>
Interview is on the 9th

[08:01:02.0000] <gsnedders>
Sorry, interviews

[08:09:00.0000] <Philip`>
That could be possible

[08:27:00.0000] <JohnResig>
Lachy: what's up?

[08:32:00.0000] <gsnedders>
Philip`: Any preference for which day? 8th would probably be nicer for me

[08:34:00.0000] <Philip`>
gsnedders: I don't think it would make any difference to me

[08:52:00.0000] <gsnedders>
Philip`: Then the 8th. Say 19:00?

[08:53:00.0000] <gsnedders>
(i guess that'll need to be ± 5 mins to keep both of us happy)

[08:57:00.0000] <Lachy>
JohnResig, can you add some tests to your selectors api test suite to test that a NodeList is returned from querySelectorAll instead of StaticNodeList?

[09:03:00.0000] <JohnResig>
Lachy: how do you mean?

[09:04:00.0000] <Dashiva>
Oh boy, deconstructionist approach to HTML parsing

[09:08:00.0000] <Lachy>
JohnResig, var test = document.querySelectorAll("p"); pass = (test instanceof NodeList);

[09:13:00.0000] <hsivonen>
Dashiva: ?

[09:14:00.0000] <JohnResig>
Lachy: meaning that StaticNodeList is a subset of NodeList?

[09:14:01.0000] <Dashiva>
"I do not believe there is some inherent meaning in the document that exists independently of its text. The producer has no more (or less) right to infer meaning from the document than the consumer does."

[09:18:00.0000] <Lachy>
JohnResig, there is no StaticNodeList in the spec. It was removed a long time ago

[09:33:00.0000] <Philip`>
gsnedders: Sounds reasonable to me :-)

[09:33:01.0000] <Philip`>
though I'll forget unless I'm reminded

[09:36:00.0000] <gsnedders>
Philip`: I'll remind 24 hours before, at least :P

[09:36:01.0000] <gsnedders>
/me has no idea whether there will be internet for interviewees

[09:36:02.0000] <gsnedders>
/me doubts it though

[09:43:00.0000] <hsivonen>
http://groups.google.com/group/mozilla.dev.planning/browse_thread/thread/f671b48712bd956f

[09:44:00.0000] <annevk2>
fun

[09:45:00.0000] <annevk2>
hendry, hehe, you really do hear your voice in the latest Standards Suck :)

[09:45:01.0000] <annevk2>
/me thinks marcos should start publishing some more

[12:03:00.0000] <annevk2>
ooh, validator.w3.org does HTML5

[12:03:01.0000] <annevk2>
guess I should update my code

[13:35:00.0000] <Lachy>
LOL. I wonder where Dmitry will find Developers crazy enough to implement his HTML6 proposals

[13:39:00.0000] <Hixie>
mark baker's opinions are coming into focus as something quite sensible

[13:40:00.0000] <Hixie>
i don't really agree with him about wanting the DOM attributes and content attributes in different sections, but i don't think it'd be that hard to do for most attributes

[13:40:01.0000] <Hixie>
for some it'd be difficult, though, like video

[13:41:00.0000] <Dashiva>
But wouldn't a stylesheet also handle that, for any separation within the same document?

[13:42:00.0000] <Hixie>
well he wants to see the definitions as far as i can tell, just clearly separated from the content attributes

[13:57:00.0000] <Lachy>
yeah, that could be something reasonable to do for some elements

[14:13:00.0000] <Hixie>
the WebSocket and Structured client-side storage sections are reasonably self-contained, i guess i could split them out around last call time

[14:13:01.0000] <Hixie>
would have to do something about MessageEvent

[14:13:02.0000] <Hixie>
maybe we can get that into DOM3 Events

[14:14:00.0000] <Hixie>
othermaciej_: btw, did you have any opinions on chaals' reply to my comments on the progress events stuff?

[14:15:00.0000] <Hixie>
it would be helpful to know what would be most helpful for implementations

[14:15:01.0000] <Hixie>
since if his style is better then i should change html5 to match

[14:19:00.0000] <othermaciej>
Hixie: I'll take a look

[14:20:00.0000] <Hixie>
thx

[14:31:00.0000] <Hixie>
ok bbl.

[14:34:00.0000] <annevk2>
http://golem.ph.utexas.edu/~distler/blog/archives/001842.html :)

[15:59:00.0000] <Hixie>
what the heck is the url to anolis

[15:59:01.0000] <Hixie>
google can't find it anywhere


2008-11-22
[16:03:00.0000] <Lachy>
Hixie, http://hg.gsnedders.com/anolis/

[16:03:01.0000] <Hixie>
thanks

[16:03:02.0000] <Hixie>
but too late

[16:03:03.0000] <Hixie>
already sent the mail :-)

[16:11:00.0000] <Hixie>
<object data=alert.html><object data=alert.html></object></object> alert.html: <script>alert('x')</script>

[16:11:01.0000] <Hixie>
how many alerts?

[16:15:00.0000] <Hixie>
wow

[16:15:01.0000] <Hixie>
IE8's <object> actually moves the markup that's in the <object> into an altHtml="" attribute

[17:36:00.0000] <Philip`>
http://philip.html5.org/misc/spec-links.html - I don't know what that is meant to show, but maybe some information can be extracted

[17:36:01.0000] <Philip`>
...except that it shows Opera doesn't support stroke="rgba(...)", so please use something more like Firefox

[17:37:00.0000] <Philip`>
(Also it shows repaint bugs in Firefox if I scroll rightwards then leftwards since bits of text go missing)

[17:38:00.0000] <Philip`>
(but never mind that)

[17:43:00.0000] <Hixie>
nice!

[17:44:00.0000] <Hixie>
not sure what to make of it, but nice nonetheless :-)

[17:46:00.0000] <Lachy>
Philip`, why is it missing lines from the Table of Contents to every other section?

[17:47:00.0000] <Philip`>
Lachy: I intentionally omitted those lines, because they're uninteresting and would probably make it uglier

[17:47:01.0000] <Lachy>
it doesn't indicate the direction of the links

[17:48:00.0000] <Philip`>
How could it indicate the directions?

[17:49:00.0000] <Lachy>
by colouring the lines from one section linking to another, on :hover or something, so that the coloured lines mean a link from that section, and the other lines mean to that section

[17:50:00.0000] <Philip`>
That sounds kind of complicated

[17:50:01.0000] <Lachy>
I know, but your talented enough to waste time doing it :-)

[17:51:00.0000] <Philip`>
If it requires more than two lines of Perl, I probably won't bother :-)

[18:00:00.0000] <Hixie>
you could make the direction more obvious using a gradient colour

[18:00:01.0000] <Hixie>
red to black

[18:00:02.0000] <Hixie>
or some such

[18:01:00.0000] <Philip`>
Is it trivial to do gradients like that in SVG?

[18:01:01.0000] <Hixie>
is anything trivial in svg?

[18:01:02.0000] <Hixie>
other than getting a well-formedness error

[18:01:03.0000] <Philip`>
This thing has been fairly trivial SVG so far :-)

[18:02:00.0000] <Hixie>
gradients aren't that hard, and you like to them as fill:url(#id) iirc

[18:02:01.0000] <Hixie>
dunno if it'll rotate the gradient though

[18:04:00.0000] <Hixie>
i love how the response to anyone volunteering to do anything here is always more work, btw :-P

[18:05:00.0000] <Hixie>
i did consider suggesting showing teh direction but i think in general the direction doesn't matter

[18:05:01.0000] <Hixie>
since what matters it that there is a link at all

[18:06:00.0000] <Lachy>
the gradients probably won't work out too well because of all the overlapping lines

[18:07:00.0000] <Philip`>
/me just reused an old Python script he had that extracted the links between sections, and then wrote a quick Perl script to translate that into SVG

[18:08:00.0000] <Philip`>
(The line widths are 1+log(number of links))

[18:11:00.0000] <Philip`>
Google's SearchWiki has a lot of very useful comments, e.g. "wikicheese.", "This is a good article about cheese!", "sucks", "i like cheese", "Mmmmm, cheese", "Cheese is great"

[18:13:00.0000] <blooberry>
philip`: I love this quote: "If it requires more than two lines of Perl, I probably won't bother" That's almost a Perl mantra in the making

[23:25:00.0000] <annevk2>
http://crisp.tweakblogs.net/blog/1221/tweakers-punt-net-en-internet-explorer-8-compatibility.html o_O

[01:56:00.0000] <Hish>
Hixie: ping

[01:56:01.0000] <Hixie>
hey

[02:00:00.0000] <annevk2>
sound of Standards Suck recording on the Uhuru Peak is actually not that bad, even while it's done with a simple digital camera rather than typical video recording hardware

[02:00:01.0000] <annevk2>
only 20s though

[02:00:02.0000] <Hish>
Hixie, I'm trying to change my OpenOffice > html converter from html 4 to 5 by changing all divs to more meaningful tags. So, currently I'm thinking about footnotes, text frames and absolute positioned graphics. For footnotes I'm about to use the display tag. Then I want to wrap graphics and text frames  with the figure tag. does this make sense?

[02:01:00.0000] <Hixie>
display tag?

[02:01:01.0000] <Hixie>
footnotes have a section in the html5 spec describing three ways to do them depending on exactly what you want

[02:01:02.0000] <Hixie>
generally you won't be able to do a good translation from a presentational format to html5 without AI (or real I, e.g. with human intervention)

[02:02:00.0000] <Hish>
sorry, the details tag

[02:02:01.0000] <Hixie>
no, footnotes aren't like <details>

[02:02:02.0000] <Hixie>
<details> is for like disclosure triangles in dialogs

[02:02:03.0000] <Hish>
ah ok. so'

[02:02:04.0000] <Hish>
so I'll search the footnore section in the spec...

[02:02:05.0000] <Hixie>
http://www.whatwg.org/specs/web-apps/current-work/#footnotes

[02:03:00.0000] <Hish>
my typing today doesn't work...

[02:04:00.0000] <Hish>
yep. your link makes much more sense than details. 

[02:59:00.0000] <Hixie>
is there any reason i shouldn't be replying to table feedback soon?

[02:59:01.0000] <Hixie>
jgraham?

[03:34:00.0000] <jgraham>
Hixie: You don't want to spark another long discussion? ;)

[03:34:01.0000] <jgraham>
(Seriously I don't know of a reason, although Ben i working on his comparison of the current alorithm and other alternatives)

[03:35:00.0000] <jgraham>
s/i/is/

[04:02:00.0000] <Hixie>
ok i'll wait til he's done with his comparison then

[05:38:00.0000] <BenMillard>
Hixie, jgraham is right that I'm doing a comparison document

[05:38:01.0000] <jgraham>
BenMillard: I think I fixed the bugs that you found

[05:38:02.0000] <jgraham>
in the table inspector

[05:39:00.0000] <BenMillard>
jgraham, cool

[05:39:01.0000] <BenMillard>
am I right in saying both algorithms need the whole table to be present before they can run? (i.e. they don't run incrementally while the table cells come in over the wire)

[05:42:00.0000] <jgraham>
BenMillard: In principle I believe either algorithm could be written to run incrementally

[05:43:00.0000] <jgraham>
Because IIRC given a cell, the headers that apply to that cell are always before it in document order

[05:44:00.0000] <jgraham>
This would change if ltr/rtl considerations ment that we sometimes had to look to the right of a cell in order to finds its headers

[05:45:00.0000] <jgraham>
I don't know if bidi issues have been given serious consideration yet (did you look at any rtl tables?)

[05:45:01.0000] <BenMillard>
jgraham, so you'd apply the header as far as it can go given how much of the table you currently have...then some more of the table comes in so you continue applying those headers until they reach the end of their scope?

[05:46:00.0000] <jgraham>
Oh and I guess @headers breaks incremental assignment

[05:46:01.0000] <BenMillard>
jgraham, I've not looked at RTL tables but I assume the source order follows the same principles as in LTR, it's just the rendering is reversed

[05:47:00.0000] <BenMillard>
jgraham, I guess headers+id would try applying to the cells you currently have, if there are no matches then you keep receiving cells and trying again until you find a match or the table ends?

[05:48:00.0000] <jgraham>
BenMillard: Well if I were implementing this in a browser, I would invert the whole algorithm so that it would be "given data cell D which of the other cells in the table form the set of headers for C?"

[05:48:01.0000] <jgraham>
See also the mail from Aaron Leventhal

[05:48:02.0000] <BenMillard>
yeah, Aaron Leventhal sent feedback about that

[05:48:03.0000] <Philip`>
Could you invert the whole algorithm without introducing a huge pile of bugs?

[05:48:04.0000] <hsivonen>
/me just replied to Aaron's email

[05:49:00.0000] <BenMillard>
(the thread is here: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2008-October/thread.html#16915)

[05:50:00.0000] <jgraham>
Philip`: I /think/ the algorithm is not hard to invert. But I have not tried

[05:50:01.0000] <hsivonen>
I think it would make a lot of sense to invert the algorithm

[05:51:00.0000] <BenMillard>
HTML4 uses the opposite direction for it's auto algorithm

[05:51:01.0000] <hsivonen>
(this is the third time, IIRC, that implementation feedback calls for the inversion of an algorithm in HTML5, AFAICT)

[05:51:02.0000] <BenMillard>
but authors write tables from the perspective of "I want this header cell to apply in this direction"

[05:52:00.0000] <jgraham>
I am vaugely considering implementing the next iteration of the HTML5 algorithm in js in such a way that it adds aria attributes to tables 

[05:53:00.0000] <BenMillard>
Aaron is right that ATs query a cell for its related header cells, but that happens after those associations are made by the UA

[05:53:01.0000] <jgraham>
So I guess if I do that I can try inverting the algorithm

[05:53:02.0000] <jgraham>
BenMillard: I guess Aaron wants to avoid precomputing all the information when it may not be needed, which could be a serious performance hit for large tables

[05:54:00.0000] <hsivonen>
BenMillard: presumably the UA would compute the association on demand for one data cell instead or rerunning the algorithm and remembering the result every time the table is mutated

[05:54:01.0000] <jgraham>
Also I agree that an authoring guide would be better written from the same point of view as the current spec

[05:55:00.0000] <BenMillard>
jgraham & hsivonen, my impression is UAs expose everything to the accessibilty APIs once the page has loaded, or while it is being loaded, so that's what I thought header association should do

[05:55:01.0000] <BenMillard>
but if the AT is expecting to trigger the UA to do the association lookup per-cell, then inverting the algorithm makes sense

[05:56:00.0000] <jgraham>
(this is probably another problem with having a normaive language reference that is different from the main spec because it would be tempting to have the normative text describing how associaions are made, and the text describing the required UA processing be from different points of view)

[05:57:00.0000] <BenMillard>
also, from the user videos I've seen, a typical user scenario is skim-reading through a data table at high speed

[05:58:00.0000] <BenMillard>
so I thought having all the associations ready and waiting was better for the user, as well as more consistent with how such mappings are done elsewhere in the document

[05:59:00.0000] <hsivonen>
BenMillard: do you have data on how common it is that authors copy and paste tables within a page without making ids unique within the page?

[05:59:01.0000] <BenMillard>
hsivonen, I haven't seen that but Hixie has

[06:00:00.0000] <hsivonen>
ok

[06:01:00.0000] <jgraham>
hsivonen: Aaron seems to disallow arcs beween tables in his proposed rewording

[06:01:01.0000] <hsivonen>
oops.

[06:02:00.0000] <BenMillard>
Aaron Leventhal's rewording is this, right? "If there is an element in the document with a corresponding ID (via getElementById) equal to the value of /id/, and it is a header cell in the current table, then assign it to the data cell."

[06:04:00.0000] <BenMillard>
hsivonen, should your reply be listed at the end of that thread or am I looking in the wrong place? http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2008-October/thread.html#16915

[06:07:00.0000] <jgraham>
BenMillard: I think that only considers replies from the same month

[06:08:00.0000] <BenMillard>
jgraham, oh of course! lost track of the month I'm on :P

[06:16:00.0000] <BenMillard>
jgraham, does your Smart Span implementation check whether a cell is empty? Do such cells change the way it works?

[06:16:01.0000] <BenMillard>
s/Smart Span/Smart Headers/

[06:16:02.0000] <Philip`>
/me likes the craziness of how a Perl database interface module relies on the Lingua::EN::Inflect module

[06:17:00.0000] <jgraham>
BenMillard: No.

[06:18:00.0000] <jgraham>
(the HTML5 algorithm does however. It is not hard to add)

[06:18:01.0000] <hsivonen>
/me wonders if the CPAN deps are tree-like or if there are circular deps

[06:21:00.0000] <Philip`>
hsivonen: I don't think there's anything to stop circular dependencies, except convention and bug reports

[06:22:00.0000] <Philip`>
(You can always force a module to install, ignoring its dependencies and its failing tests, but it's usually a bug if you have to do so)

[06:22:01.0000] <jgraham>
Does CPAN no cope with circular deps?

[06:22:02.0000] <jgraham>
not

[06:23:00.0000] <hsivonen>
/me learns new things (to him) about C++ link errors

[06:23:01.0000] <hsivonen>
specifically, one can have two .o files that link together into an .a but not into a shared lib

[06:23:02.0000] <BenMillard>
jgraham, HTML5 says "User agents may remove empty data cells when analyzing data in a table."

[06:24:00.0000] <hsivonen>
I guess Steve Yegge is right that linker errors are a top thing of the C++ brand

[06:24:01.0000] <Philip`>
jgraham: No - when you tell it to install a module, it recursively asks whether you want to install its missing dependencies, and then it compiles and then runs the tests (and the tests will fail if you've got missing dependencies) and then installs

[06:25:00.0000] <jgraham>
Oh. Well that seems like it could be avoided

[06:26:00.0000] <jgraham>
BenMillard: HTML 5 treats <th>Foo<td><td> as equivalent to <th colspan=3>Foo

[06:27:00.0000] <jgraham>
but does not do the same vertically

[06:27:01.0000] <BenMillard>
jgraham, in a table with 3 columns, yes

[06:28:00.0000] <BenMillard>
but the "may remove empty data cells" thin seems like a bad idea because users would easily become lost in tables with many empty cells

[06:30:00.0000] <BenMillard>
in contrast, ignoring empty header cells seems like a good thing (such as an empty <th> in the top left corner of a table with both column headers and row headers)

[06:33:00.0000] <jgraham>
BenMillard: I wasn't sure what the implications of "removing empty data cells" was supposed to be which I guess is a sign that it sould be clarified

[06:33:01.0000] <jgraham>
s/cells"/cells when analysing data in a table"/

[06:35:00.0000] <BenMillard>
jgraham, my impression was it meant UAs don't have to associated header cells with empty data cells?

[06:35:01.0000] <BenMillard>
s/associated/associate/

[06:36:00.0000] <jgraham>
Also I am unsure why the wide cell thing only works if the cell extends to the edge of the table

[06:36:01.0000] <jgraham>
or raher if they empty cells extend to the edge of the table

[06:37:00.0000] <jgraham>
BenMillard: Maybe. If so, the "may" makes the algorithm non-deterministic and the scope should probably be limited to data cells that fom part of a wide cell

[06:41:00.0000] <BenMillard>
jgraham, I think the wide cell thing is for "section headers" which should use <th colspan="width_of_table">Foo</th> but instead used <th>Foo<td><td><td> to reach the edge of the table

[06:41:01.0000] <BenMillard>
it makes the gamble that a row header won't be given for a row where there was no data to report (which seems fairly safe)

[06:42:00.0000] <jgraham>
BenMillard: But that could just as easilly happen with a heading that extended across only part of the table, no?

[06:42:01.0000] <jgraham>
<th>Foo<td><td><th>Bar

[06:42:02.0000] <BenMillard>
jgraham, indeed.

[06:43:00.0000] <jgraham>
I don't understand your row header comment

[06:45:00.0000] <BenMillard>
jgraham, <tr><th>2006<td>5.6<td>8.3<tr><th>2007<td><td><tr><th>2008<td>7.3<td>9.8

[06:45:01.0000] <BenMillard>
no data is present in the row for "2007", but it's not a section header

[06:46:00.0000] <Philip`>
Hmm, Ubuntu (Gutsy) has .atom mapped to "application/atom" in mime.types, which breaks Opera and Firefox because it should have "+xml"

[06:47:00.0000] <BenMillard>
jgraham, usually a row where no data is available has a character (http://www.avert.org/aofconsent.htm) or a note but they could just as logically be empty

[06:47:01.0000] <Dashiva>
Philip`: Good thing there's extensive testing to detect things like that

[06:48:00.0000] <jgraham>
BenMillard: Oh I see. Interesting

[06:49:00.0000] <jgraham>
/me decides to delay trying to formulate an email covering all of this until later

[06:54:00.0000] <Philip`>
I thought I was being cleverly lazily by using file extensions so I didn't have to look up the Apache documentation on how to override content-types, but sadly I was thwarted

[06:54:01.0000] <Philip`>
s/lazily/lazy/

[06:54:02.0000] <BenMillard>
jgraham, it's worth noting that an AT can announce "Empty" or somesuch for truely empty cells but has to say "Question" or "Dash" if there's a random character, just in case that character is significant

[06:55:00.0000] <hsivonen>
Philip`: is that an Apache bug, a Debian bug or an Ubuntu bug?

[06:55:01.0000] <hsivonen>
Philip`: bad brokenness in any case

[06:55:02.0000] <BenMillard>
jgraham, so it's actually better for users if empty rows are truely empty, which is incompatible with the "wide header cell" thing in HTML5 when such rows have a row header cell

[06:55:03.0000] <hsivonen>
me wonders how static array of ints can reference a static pointer

[06:57:00.0000] <hsivonen>
the linker is telling me that the pointer thingy is an undefined symbol referenced from what is a static array of ints

[06:58:00.0000] <BenMillard>
jgraham, it's also worth noting that it would compatible with some real tables, such as Table 4.1 here, if they used <th> for their header cell: http://www.federalreserve.gov/generalinfo/basel2/docs2005/potentialimpact.htm

[06:58:01.0000] <BenMillard>
jgraham, although if they retrofitted <th> to their header cells, they could retrofit appropriate colspan at the same time to make these proper section headers. :)

[07:02:00.0000] <BenMillard>
jgraham, I think implying <td><b> and <td><strong> as aliases of <th> messes up reasonable tables, so we should drop that heuristic.

[07:05:00.0000] <BenMillard>
jgraham, most frustrating case I've found is <td><b> both for row headers and for important data, meaning the <td><b> heuristic would get the row headers right but would erroneously apply "5.8 mm" as a column header as well: http://flickr.com/photos/adactio/1383011861/meta/in/photostream/

[07:06:00.0000] <BenMillard>
jgraham, could the multi-table support only pick the innermost tables? the above page causes "Table 3" to have the layout table around the data table: http://james.html5.org/tables/table_inspector.py?input_type=type_uri&uri=http%3A%2F%2Fflickr.com%2Fphotos%2Fadactio%2F1383011861%2Fmeta%2Fin%2Fphotostream%2F&source=&algorithm=smartheaders

[07:07:00.0000] <BenMillard>
jgraham, it's available further down at Table 4 and Table 5, so removing Table 3 due to it containing another table seems possible

[07:09:00.0000] <Philip`>
hsivonen: The file says "This file is part of the "mime-support" package.  Please send email (not a bug report) to mime-support⊙pdo if you would like new types and/or extensions to be added." which makes it sound like Debian

[07:10:00.0000] <hsivonen>
Philip`: I suggest sending email to that address, then 

[07:10:01.0000] <Philip`>
(Maybe it's fixed in newer versions, but I can't update this Ubuntu because apt-get gives errors while trying to download stuff)

[07:17:00.0000] <Philip`>
/me is surprised that code he wrote two years ago still actually runs, without any library incompatibilities or similar problems, having reinstalled it on a more modern system

[07:17:01.0000] <Philip`>
despite using the Catalyst framework and various crazy modules

[07:36:00.0000] <hsivonen>
http://twitter.com/asmitter/statuses/1018146131

[08:21:00.0000] <Lachy>
http://www.alistapart.com/articles/thisishowthewebgetsregulated

[08:22:00.0000] <BenMillard>
ALA's URLs get worse with each issue

[08:23:00.0000] <jgraham>
BenMillard: Tables with descendant tables are no longer shown. If this causes problems in practice, I will revert the change

[08:33:00.0000] <BenMillard>
jgraham, that seems to work. :)

[08:35:00.0000] <Lachy>
I completely disagree with Joe Clark's recommendation to use open captioning instead of closed captioning

[08:36:00.0000] <Philip`>
Why?

[08:37:00.0000] <Lachy>
what's really needed is a decent, widely supported, standardised closed captioning/subtitling format that solves all the technical problems with the existing mess of formats

[08:38:00.0000] <Dashiva>
srt!

[08:38:01.0000] <Lachy>
so that one closed caption/subtitle file will work regardless of the video codec or device

[08:38:02.0000] <Lachy>
Dashiva, SRT is a very limited format

[08:39:00.0000] <Philip`>
Lachy: Won't that still have the same problems about unpredictability of fonts and of requiring extra UI to enable the captions?

[08:41:00.0000] <Lachy>
we also need a better selection of fonts available, which can be done by providing some guidelines about what fonts that media player vendors/device manufacturers should use

[08:41:01.0000] <Lachy>
and get them all to standardise on using UTF-8

[08:42:00.0000] <Lachy>
a decent subtitle format needs to be able to specifiy fonts, colours, position and timing of text

[08:44:00.0000] <Lachy>
the problem with open captions is that it requies creating 2 or more copies of the same video. One without captions and one or more with captions or subtitles in various languages

[08:47:00.0000] <BenMillard>
Lachy, http://screenfont.ca/fonts/today/ and http://screenfont.ca/tech/ are Joe Clark's professional opinion on fonts and Unicode in captioning and subtitles (http://screenfont.ca/about/) which seem close to your own ("New fonts", "A single Unicode encoding")

[09:48:00.0000] <Lachy>
http://lists.w3.org/Archives/Public/www-archive/2008Nov/0028.html

[09:48:01.0000] <Lachy>
http://lists.w3.org/Archives/Public/www-archive/2008Nov/0029.html

[09:50:00.0000] <Lachy>
it's interesting to see how much Dean's opinion has changed so dramatically in the past 12 months

[10:39:00.0000] <yecril71>
I consider a major disadvantage of HTML (and XML also) that locality of identifiers is not supported.

[10:40:00.0000] <yecril71>
It effectively forces page authors to use GUIDs for identifiers in templates.

[10:40:01.0000] <yecril71>
And those GUIDs have to be regenerated each time a template is used,

[10:41:00.0000] <yecril71>
if only there is a chance that a template is used twice on the same page.

[10:41:01.0000] <yecril71>
The situation is mildly alleviated by the fact that a NAME is local to a FORM.

[10:42:00.0000] <yecril71>
Not all problems can be handled in this way though.

[11:18:00.0000] <gsnedders>
Hixie: Why the asking about Anolis?

[11:19:00.0000] <gsnedders>
Hixie: http://anolis.gsnedders.com fwiw

[11:19:01.0000] <gsnedders>
(though it was all ready too late when Lachy gave the wrong answer)

[12:27:00.0000] <gsnedders>
Why do banks use things for security that are entirely forgettable?

[12:29:00.0000] <Dashiva>
Because they predate the internet?

[12:30:00.0000] <gsnedders>
I'm expected to give the 5th, 3rd, and 2nd digit (in that order) of "my" passnumber

[12:30:01.0000] <gsnedders>
Like I can remember it at all

[12:33:00.0000] <Philip`>
Maybe it's your date of birth

[12:35:00.0000] <gsnedders>
Philip`: no

[12:39:00.0000] <jwalden>
easy!  5, 3, 2

[12:41:00.0000] <Philip`>
gsnedders: Oh, okay; my passnumber is your date of birth, so I thought it'd be a good guess

[12:41:01.0000] <gsnedders>
Philip`: Big/little endian?

[12:42:00.0000] <Philip`>
gsnedders: If I told you that, you'd be able to steal all my money, so I think I'll pass

[12:43:00.0000] <gsnedders>
/me thinks he knows the first two digits

[13:15:00.0000] <theanxy>
wii Lachy 

[13:17:00.0000] <theanxy>
sorry.

[15:07:00.0000] <Lachy>
"It is easy to make the case that the general video uploader would find it an undue hardship to caption their videos. Undue hardship is a legitimate exemption in every relevant piece of legislation. So let’s assume that you the video uploader wouldn’t have to do it. Fine.

[15:07:01.0000] <Lachy>
"But what about your host? Google could afford to do it for you, though they wouldn’t want to. Should they have to?"

[15:07:02.0000] <Lachy>
http://www.alistapart.com/comments/thisishowthewebgetsregulated?page=1#10

[15:07:03.0000] <Lachy>
Clearly, Joe Clark doesn't have a clue about just how much video content is uploaded to YouTube and Google Video

[15:08:00.0000] <Lachy>
last I heard, it was something like 10 minutes worth of video uploaded every minute

[15:09:00.0000] <Lachy>
somehow I think captioning 240 hours of video every day would be impossible for any organisation

[15:14:00.0000] <Philip`>
http://news.bbc.co.uk/1/hi/uk/7730679.stm - "13 hours every minute"

[15:14:01.0000] <Philip`>
You're off by a few orders of magnitude :-)

[15:20:00.0000] <Lachy>
woah

[15:21:00.0000] <Philip`>
They must have quite a few servers doing video transcoding

[15:22:00.0000] <Philip`>
It would work fine if Google just farmed out the captioning jobs to the Amazon Mechanical Turk

[15:23:00.0000] <Philip`>
Offer $0.02 to caption a minute of video, and it'll only cost ~$8M/year

[15:25:00.0000] <Lachy>
yeah, cause captioning 1 minute of video would take someone about 10-15 minutes worth of work for an expert (including transcribing it and syncing up the text, etc.) and someone's time is really only worth $0.10/hour anyway

[15:27:00.0000] <jgraham>
I'm not sure you could find ~10,000 people to caption video for that much

[15:27:01.0000] <Philip`>
These are not just "someone" - they're someone who's decided to scrounge for money on Mechanical Turk, so they're not going to have very high expectations

[15:28:00.0000] <jgraham>
Even if you could pay 100 times that, which would make the cost prohibitive

[15:28:01.0000] <Philip`>
And Google could cut down the work hugely by detecting videos with pretty much the same audio track as others, since they're mostly songs or clips from TV shows posted a zillion times

[15:29:00.0000] <Philip`>
and you don't really need to get a thousand people to independently caption Rick Astley if you can figure out that the repeats

[15:29:01.0000] <Lachy>
Google should just develop speech recognition and produce captions on the fly

[15:30:00.0000] <Philip`>
Hmm, but those would only be subtitles and not captions

[15:31:00.0000] <svl>
Google should just heal every possible kind of auditory impairment people could suffer from, so captioning wouldn't be necessary anymore.

[15:31:01.0000] <Lachy>
except for people who need to keep the volume turned down low for some reason

[15:31:02.0000] <jgraham>
I wonder what fraction of the content on youtube is duplicates

[15:32:00.0000] <Philip`>
/me has probably already mentioned several times that he played Portal with no speakers or headphones, and its captioning worked pretty nicely

[15:33:00.0000] <Philip`>
(complete with textual versions of gameplay-relevant sound effects)

[15:36:00.0000] <jgraham>
/me wonders why Rob Sayre thinks that his issue with the parser spec indicates a faliure of process

[15:42:00.0000] <Philip`>
It seems he is saying that since it is not perfect, that invalidates claims that the way it is written is better than other ways

[15:46:00.0000] <Lachy>
he must be completely ignoring the fact that in its current state, it's still better than the state of any other prior spec written using alternative methods

[15:46:01.0000] <Philip`>
http://philip.html5.org/misc/portal-caption.jpg

[15:47:00.0000] <Philip`>
although, oddly, the audio says "and then there will be" in place of "[garbled]"


2008-11-23
[18:26:00.0000] <sayrer>
jgraham, all I meant was that it's not stable

[18:28:00.0000] <sayrer>
"the parsing work is stable to the extent that it can be without having implementations in major-marketshare browsers"

[18:28:01.0000] <sayrer>
I do not think this is true

[18:29:00.0000] <sayrer>
and if someone wanted to work on that piece as a separate document, that might be reasonable

[18:29:01.0000] <sayrer>
not automatically dismissed with something along the lines of "our way is the best way, move along"

[19:22:00.0000] <sayrer>
hmm, looks like html5lib doesn't parse the contents of the noscript element

[19:28:00.0000] <Hixie>
i don't understand this interest in splitting out html5 using this new scheme

[19:28:01.0000] <Hixie>
i've never seen a spec do this before

[19:29:00.0000] <Hixie>
(the "implementation" vs "authoring" split that some people want)

[19:30:00.0000] <sayrer>
that one doesn't make much sense to me either. DOM initialization vs. post DOM initialization is more sensible (parser/script dependencies notwithstanding). another more sensible split would be old stuff vs. new stuff

[19:35:00.0000] <Hixie>
sayrer: which "consistency and stability claims" are overblown and how are they overblown?

[19:36:00.0000] <Hixie>
sayrer: also, how is the HTML5 change control unsatisfactory?

[19:37:00.0000] <sayrer>
Hixie, well jgraham is claiming that the parser section is so stable that nothing further can happen without major browser implementation

[19:37:01.0000] <sayrer>
yet html5lib doesn't even do what the spec says regarding noscript

[19:37:02.0000] <Hixie>
sayrer: i doubt he thinks that _nothing_ further can happen

[19:38:00.0000] <sayrer>
"the parsing work is stable to the extent that it can be without having implementations in major-marketshare browsers"

[19:38:01.0000] <Hixie>
but it's certainly true that we're mostly just waiting for wide-scale implementation experience

[19:38:02.0000] <Hixie>
well that's true

[19:38:03.0000] <Hixie>
it's as stable as it can ever be without wide-scale implementation experience

[19:38:04.0000] <Hixie>
that doesn't mean nothing further can happen

[19:38:05.0000] <Hixie>
even a completely stable spec can change

[19:39:00.0000] <Hixie>
just not much

[19:39:01.0000] <Hixie>
even a shipped product (the epitomy of stable) can change, e.g. security fixes

[19:39:02.0000] <Hixie>
CSS2 was as stable as it gets, yet look at the amount of change _it_ went through

[19:41:00.0000] <sayrer>
just saying there's an obvious delta between the spec and a library that most major html5 contributors have commit access to

[19:41:01.0000] <Hixie>
so?

[19:42:00.0000] <sayrer>
surprised it could happen

[19:42:01.0000] <Hixie>
really?

[19:42:02.0000] <Hixie>
dude, all software has bugs

[19:42:03.0000] <Hixie>
all specs have bugs

[19:42:04.0000] <sayrer>
well, surprised it's never even been noticed

[19:42:05.0000] <Hixie>
i hate to be the one to break it to you

[19:44:00.0000] <Hixie>
sayrer: also, how is the HTML5 change control unsatisfactory?

[19:44:01.0000] <sayrer>
Hixie, I don't like the explicit private feedback challenges

[19:44:02.0000] <sayrer>
er

[19:44:03.0000] <sayrer>
channels

[19:45:00.0000] <Hixie>
yeah, nor do i

[19:45:01.0000] <sayrer>
like, I am guessing you did not invent MessagePort

[19:45:02.0000] <sayrer>
maybe I am wrong about that

[19:45:03.0000] <Hixie>
sadly, some people don't want to say anything in public because their products are secret, or whatnot

[19:45:04.0000] <Hixie>
invent how?

[19:45:05.0000] <sayrer>
it was quite different from the Apple and Mozilla proposals as I recall

[19:46:00.0000] <Hixie>
proposals for what?

[19:46:01.0000] <Hixie>
do you mean the worker stuff?

[19:46:02.0000] <sayrer>
yeah

[19:46:03.0000] <Hixie>
workers are unrelated to messageport really

[19:46:04.0000] <Hixie>
though i did reuse them for that purpose since it seemed useful

[19:47:00.0000] <sayrer>
well, at any rate, I never proposed a course of action

[19:47:01.0000] <Hixie>
(messageport though, the api, was basically something i wrote up based on a series of proposals related to capabilities APIs through messages)

[19:47:02.0000] <sayrer>
I just don't think other editorial splits are obviously bad

[19:48:00.0000] <Hixie>
oh some are definitely good

[19:48:01.0000] <Hixie>
like we should split out the storage stuff eventually

[19:48:02.0000] <sayrer>
especially if someone is willing to do a document

[19:48:03.0000] <Hixie>
since that's pretty self-contained

[19:48:04.0000] <sayrer>
even if you don't suggest it

[19:48:05.0000] <sayrer>
anyway, gtg

[19:48:06.0000] <Hixie>
later

[22:56:00.0000] <Hixie>
there's no way to install ubuntu that doesn't require installing X?

[22:59:00.0000] <hdh>
i hear the server edition doesn't install X

[23:02:00.0000] <Hixie>
man that's hard to see on their site

[23:03:00.0000] <Hixie>
thanks

[23:03:01.0000] <Hixie>
that's what i want

[03:01:00.0000] <jgraham>
Yeah, we should probably have provided feedback from html5lib that we wanted to (at least optionally) behave like a scripting UA even with no scripting support

[03:49:00.0000] <takkaria>
Hubbub has a scripting/no-scripting flag too

[03:55:00.0000] <Dashiva>
I wonder how much all this noise about reaching rec faster is contributing to slowing down rec :)

[04:25:00.0000] <gsnedders>
SNOW!

[07:21:00.0000] <maikmerten>
Question about the media API: Seeking works by setting currentTime to a new value (in seconds). Is there a special reason why currentTime isn't readonly and seeking works e.g. with a method seek(), taking a relative value (zero to one), which would cover the case when the duration of the stream isn't known (but it's filesize is) and rough seeking shall still be available.

[07:21:01.0000] <maikmerten>
s/./?

[07:22:00.0000] <maikmerten>
(there are media formats without a header containing duration, so all you got is the byte position if the client shouldn't try to find a timestamp at the end of the bytestream)

[07:23:00.0000] <gsnedders>
zero to one relative to the filesize?

[07:23:01.0000] <gsnedders>
(seeming the length isn't known)

[07:24:00.0000] <maikmerten>
yeah, if the duration in "proper time" isn't known one can still seek to an approximate position using the position of the bytestream

[07:25:00.0000] <maikmerten>
this, however, only works if seeking is not taking absolute time. The client can then find a fitting representation

[07:26:00.0000] <maikmerten>
if duration is known: Seek to the position (in seconds or whatever)

[07:26:01.0000] <maikmerten>
if only the length of the bytestream is known: Well, just jump to a position in there

[07:27:00.0000] <maikmerten>
I'm currently trying to implement a subset of the HTML5 media specs for a Java playback applet, which only knows bytepositions and thus always seeks relative

[07:29:00.0000] <maikmerten>
I *could* estimate the total duration in seconds given the current byteposition and the current timestamp the media stream has

[07:30:00.0000] <maikmerten>
but that means the duration will change during playback (and given how dynamic video bitrate is this estimate can be like 1000% off)

[07:30:01.0000] <maikmerten>
(the estimate will first be horrible and be more accurate at the end)

[07:31:00.0000] <maikmerten>
this isn't just specific to this applet thingie, but for some formats the clients really have to seek to the end of the file and find a timestamp befor the duration is known

[07:32:00.0000] <nessy>
it's why doublec has asked for a duration attribute

[07:32:01.0000] <maikmerten>
ah, I see :)

[07:32:02.0000] <maikmerten>
same solution as for my applet ;-)

[07:32:03.0000] <nessy>
not sure it will come though - and whether it's a good solution, because not every author may include that info

[07:32:04.0000] <maikmerten>
most won't

[07:33:00.0000] <nessy>
yeah, so the estimate is probably the best you can do

[07:33:01.0000] <maikmerten>
quite frankly I'd hope that currentTime would be currentPosition, which can either be seconds *or* byte position

[07:33:02.0000] <maikmerten>
and a flag indicating what it is

[07:33:03.0000] <maikmerten>
same for duration

[07:34:00.0000] <maikmerten>
but this smells like a hack

[07:34:01.0000] <nessy>
or a value marker (e.g. sec,B)

[07:34:02.0000] <BenMillard>
/me has radically simplified his Public-HTML message filter: it now marks everything without me in either the To or CC line as Read.

[07:34:03.0000] <maikmerten>
right

[07:35:00.0000] <nessy>
well, width/height take pixels, em and other entities, too

[07:35:01.0000] <nessy>
why not currentPosition

[07:35:02.0000] <Philip`>
BenMillard: You could simply unsubscribe from the group :-)

[07:35:03.0000] <maikmerten>
http://people.xiph.org/~maikmerten/demos/bigbuckbunny-applet.html <-- applet without duration given. Will still allow seeking (relative to byteposition) and display a progress bar (also byteposition). Works out okay in practice

[07:35:04.0000] <nessy>
yeah

[07:35:05.0000] <nessy>
write an email to the list suggesting this, I'd say

[07:36:00.0000] <nessy>
/me has to get some sleep

[07:36:01.0000] <maikmerten>
sweet dreams!

[07:36:02.0000] <maikmerten>
and yeah, will write a mail

[07:36:03.0000] <nessy>
yeah - 2:30am here :)

[07:36:04.0000] <nessy>
cool, cya

[07:37:00.0000] <BenMillard>
Philip`, http://lists.w3.org/Archives/Public/public-html/2007Jul/0096.html

[07:38:00.0000] <gsnedders>
BenMillard: Lazy so and so

[07:38:01.0000] <hsivonen>
Lachy: FWIW, .srt goes almost all the way considering the feature set of subtitling in Northern Europe (it lacks italics and placing text sometimes at the top of the frame)

[07:52:00.0000] <BenMillard>
jgraham, it seems column headers are applied to the column headers on their right and row headers are applied to the rows below them in the HTML5 (November 2008) mode: http://james.html5.org/tables/table_inspector.py?input_type=type_uri&uri=http%3A%2F%2Fprojectcerbera.com%2Fblog%2F2008%2F09%2Funtangle&source=&algorithm=html5b

[07:53:00.0000] <BenMillard>
Step 1.5.5 and Step 1.5.10 seem to be the parts of HTML5 which are supposed to prevent that, by block associations coming from headers of the same height or width (depending on the axis)

[07:53:01.0000] <BenMillard>
s/block/blocking/

[07:54:00.0000] <BenMillard>
(re: the "steps" in here: http://www.whatwg.org/specs/web-apps/current-work/multipage/tabular-data.html#header-and-data-cell-semantics)

[08:00:00.0000] <BenMillard>
jgraham, only highlighting when clicking around a table makes checking the associations a dream

[08:02:00.0000] <BenMillard>
jgraham, Smart Headers works correctly across column headers and down row headers: http://james.html5.org/tables/table_inspector.py?input_type=type_uri&uri=http%3A%2F%2Fprojectcerbera.com%2Fblog%2F2008%2F09%2Funtangle&source=&algorithm=smartheaders

[08:08:00.0000] <maikmerten>
(my proposals on relative seeking: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2008-November/017387.html )

[08:09:00.0000] <maikmerten>
(oh, nessy did leave the room)

[08:14:00.0000] <jgraham>
BenMillard: OK, I will have a look at what I did wrong

[08:21:00.0000] <jgraham>
BenMillard: Does that do what you expect now?

[08:28:00.0000] <BenMillard>
jgraham, that fixes the bug I reported without regressions in the data cells

[08:29:00.0000] <jgraham>
BenMillard: I had forgotten to prevent header cells from ever forming associations

[08:29:01.0000] <BenMillard>
right, so in HTML5 headers don't associate with headers, ever, even when they have different spans and are next to each other?

[08:30:00.0000] <BenMillard>
(that's my reading of HTML5, btw, which your inspector now follows perfectly)

[08:30:01.0000] <jgraham>
I believe that is correct

[08:31:00.0000] <gsnedders>
Why can we not have headers associating with headers?

[08:31:01.0000] <BenMillard>
gsnedders, I think HTML5 will change in this respect (Smart Headers has always allowed header cells to have header cells)

[08:32:00.0000] <jgraham>
gsnedders: Hixie doesn't think that it makes sense for some reason I always forget

[08:32:01.0000] <gsnedders>
jgraham: silly Hixie.

[08:32:02.0000] <jgraham>
I think he is worried about UI issues (too much information being reported)

[08:33:00.0000] <jgraham>
But I think removing that information is premature optimisation and that UAs should be free to compete on which of the associations they read out in different circumstances

[08:34:00.0000] <jgraham>
e.g. travelling across a table they might only read the headers that changed, but on a keypress might read everything

[08:34:01.0000] <BenMillard>
jgraham, indeed

[08:34:02.0000] <jgraham>
Which I think largely eliminates the problem

[08:35:00.0000] <BenMillard>
in particular, it's needed when there are multiple levels of header cells and the lower levels have repetitve text which is disambiguated by the upper levels. such as: http://lists.w3.org/Archives/Public/www-archive/2007Aug/att-0003/offset-mess.htm

[08:36:00.0000] <BenMillard>
in HTML5, reading across level 2 of the column headers doesn't report whether you are in the "Quirks Mode" or "Standards Mode" section of the table: http://james.html5.org/tables/table_inspector.py?input_type=type_uri&uri=http%3A%2F%2Flists.w3.org%2FArchives%2FPublic%2Fwww-archive%2F2007Aug%2Fatt-0003%2Foffset-mess.htm&source=&algorithm=html5b

[08:36:01.0000] <BenMillard>
(similarly, level 3 doesn't report whether you are in <body> or <html> for either "Quirks Mode" or "Standards Mode")

[08:37:00.0000] <BenMillard>
in Smart Headers, the headers do associate with each other (so, for example, a user could query the final "scroll" to header "Standards Mode, <html>" which disambiguates where they are): http://james.html5.org/tables/table_inspector.py?input_type=type_uri&uri=http%3A%2F%2Flists.w3.org%2FArchives%2FPublic%2Fwww-archive%2F2007Aug%2Fatt-0003%2Foffset-mess.htm&source=&algorithm=smartheaders

[08:37:01.0000] <BenMillard>
s/to header/to hear/

[08:40:00.0000] <BenMillard>
(I discussed this with Hixie over IRC a couple of months ago, which is why I think it will change: http://krijnhoetmer.nl/irc-logs/whatwg/20080926#l-238)

[08:40:01.0000] <BenMillard>
gsnedders, ^^^

[08:45:00.0000] <BenMillard>
perhaps I should include a diagram to explain the terminology my comparison document uses

[10:10:00.0000] <Hixie>
maikmerten: the browser is required to work out the duration

[10:11:00.0000] <maikmerten>
Hixie, that is a pretty heavy requirement

[10:12:00.0000] <Hixie>
yup

[10:12:01.0000] <maikmerten>
heh :)

[10:13:00.0000] <maikmerten>
what about concatenated files?

[10:13:01.0000] <maikmerten>
seeking to "the end" once is a reasonable requirement, I guess

[10:14:00.0000] <maikmerten>
I just wonder what to do in the nastier cases

[10:14:01.0000] <maikmerten>
(x files concatened, total duration would have to determined by determining "the end" for each segments, basically meaning a binary search for each)

[10:15:00.0000] <maikmerten>
oh the nastyness of media containers - seeking tables in the headers suck, not having those suck, too ;)

[10:20:00.0000] <Hixie>
you have to work out the time corresponding to the start and end of every set of bytes you have

[10:20:01.0000] <Hixie>
to populate the bufferedTimes attribute

[10:25:00.0000] <maikmerten>
pretty ambitious, but if all browser vendors implement this properly this'll be pretty cool stuff

[10:28:00.0000] <maikmerten>
anyway, one question about API design. I have no doubt it has received considerable... consideration. Why are some actions (let's take seeking) triggered by altering an attribute, while others (starting playback etc) are methods?

[10:36:00.0000] <Hixie>
play() does more than set paused to false

[10:36:01.0000] <Hixie>
it might also call load()

[10:37:00.0000] <Hixie>
which is why it's a method

[10:37:01.0000] <Hixie>
currentTime just seeks and sets currentTime

[10:42:00.0000] <maikmerten>
hmmm... won't it also alter other things? The video stops while the buffers need to be refilled etc. etc.

[10:42:01.0000] <maikmerten>
but anyway, I assume this just doesn't matter

[10:51:00.0000] <Hixie>
it's not a great worry to me, no :-)

[11:08:00.0000] <maikmerten>
:)

[13:16:00.0000] <BenMillard>
just noticed this in my feeds: http://www.w3.org/QA/2008/11/w3c_validator_now_with_html5.html

[13:17:00.0000] <BenMillard>
the "More Options" link opens an area with a "Document Type" dropdown, where "HTML5 (experimental)" is the 2nd option: http://validator.w3.org/#validate_by_uri+with_options

[13:18:00.0000] <KrocCamen>
http://validator.w3.org/check?uri=http%3A%2F%2Fcamendesign.com&charset=(detect+automatically)&doctype=Inline&group=0 :)

[13:20:00.0000] <BenMillard>
krijnh, parentheses inside a URI works better in the logs than in Opera. well done B)

[13:21:00.0000] <BenMillard>
KrocCamen, I think this one is more interesting: http://validator.w3.org/check?uri=http%3A%2F%2Fwww.google.com%2Fprivacy.html&charset=%28detect+automatically%29&doctype=Inline&group=0&user-agent=W3C_Validator%2F1.606

[13:22:00.0000] <KrocCamen>
It’s only HTML5 by doctype though.

[13:22:01.0000] <KrocCamen>
No section / header / footer tags &c.

[13:23:00.0000] <KrocCamen>
Everybody should switch to the HTML5 doctype really though. I’m never writing another HTML4 site again, even if I don’t use HTML5 tags in order to support IE (which my site doesn’t).

[13:23:01.0000] <BenMillard>
KrocCamen, it's still conforming HTML5 on a google.com property :)

[13:23:02.0000] <KrocCamen>
Many of the Mozillians and W3C editors are at Google.

[13:23:03.0000] <KrocCamen>
Now Microsoft - that’s a different story

[13:23:04.0000] <KrocCamen>
Silverlight Ahoy!

[13:27:00.0000] <BenMillard>
KrocCamen, you wrote "For a 12 year old piece of software, the VB6 IDE still does it for me. Pure bliss." and I totally, if rather guiltily, agree! :D (http://camendesign.com/tweet/vb6-ide)

[13:28:00.0000] <KrocCamen>
It’s an odd piece of software. Not much software ages well.

[13:29:00.0000] <KrocCamen>
Everything else from that era is unusable. TextMate might age quite well. It is perfectly usuable and doesn’t need to do anything more.

[13:30:00.0000] <BenMillard>
the syntax highlighting is very conservative and all the better for it; the auto-indention does what I expect; I use Shift+F2 all the time; setting tab orders in the Forms view Works with a big W

[13:30:01.0000] <gsnedders>
/me dislikes TextMate

[13:30:02.0000] <gsnedders>
/me can never remember why

[13:30:03.0000] <BenMillard>
(my comment was about the VB6 IDE, btw)

[13:30:04.0000] <gsnedders>
Whenever I try it I like for around five minutes, then get really annoyed at it

[13:31:00.0000] <KrocCamen>
Hmm Mac text editors are important to users. It’s not like Windows where everything sucks equally.

[13:31:01.0000] <KrocCamen>
People rag on about BBedit, and I can’t personally bear it.

[13:31:02.0000] <gsnedders>
Urgh

[13:31:03.0000] <gsnedders>
BBEdit was good on OS 9

[13:31:04.0000] <jgraham>
/me couldn't get the indentation to work in textmate

[13:31:05.0000] <gsnedders>
But now, it still seems as if it should be in BBEdit

[13:32:00.0000] <gsnedders>
*OS 9

[13:32:01.0000] <gsnedders>
/me is too tired

[13:32:02.0000] <BenMillard>
I eventually cobbled together my own MDI text editor in VB6 so I could see how hard it is to make a text editor I like...turns out to be harder than I thought

[13:32:03.0000] <jgraham>
First rule of text editors: Unless you indent as well as emacs it is really hard for me to like you

[13:32:04.0000] <BenMillard>
especially since I'm plain awful at programming :P

[13:32:05.0000] <KrocCamen>
Before I switched to Mac, I used my own text editor for four years because I couldn’t find anything adequate, commercial or not, on Windows.

[13:34:00.0000] <KrocCamen>
There’s a feature I had though, that I’ve never seen in any other editor, and that I still pine for in TextMate.

[13:34:01.0000] <KrocCamen>
--Split Folder / File view.

[13:35:00.0000] <KrocCamen>
Screenshot here http://camendesign.com/InlineASP.png

[13:36:00.0000] <jgraham>
Just the directory tree thing on the left?

[13:37:00.0000] <jgraham>
Isn't that quite common?

[13:37:01.0000] <BenMillard>
(that Google Privacy page uses <meta charset="utf-8"> as well as <doctype html>, omitting a lot tags which are also optional in HTML4)

[13:37:02.0000] <KrocCamen>
Yeah. If a folder contains 1’000 files, you don’t want the folder tree that far apart.

[13:38:00.0000] <gsnedders>
jgraham: I could never get into emacs or vim. Both are just horrid for a UX POV

[13:38:01.0000] <KrocCamen>
I’ve never seen it in another editor. It’s always combined folder/file view.

[13:38:02.0000] <jgraham>
Oh, I see you mean the way that the files are shown below the directories rather than in the same tree

[13:39:00.0000] <BenMillard>
KrocCarmen, I don't see the point in that. Windows Explorer provides that, from which you can drag and drop multiple files, folders, mixtures of both, etc

[13:39:01.0000] <jgraham>
gsnedders: emacs may not be pretty and may have weird keyboard shortcuts but it has awesome functionality and I am only a very trivial user

[13:39:02.0000] <KrocCamen>
It’s different in a web project, really. I used it for four years and I swear by it.

[13:41:00.0000] <jgraham>
gsnedders: For example the indentation in python-mode is better than anything else I have ever tried, AucTeX is the only thing that makes writing LaTeX bearable, js2 mode is wonderful

[13:42:00.0000] <jgraham>
This makes up for the weird modal interface and all the things that make UI people cry

[13:42:01.0000] <jgraham>
I have ever learnt more vi than how to quit it when I accidentially launch it though :)

[13:43:00.0000] <gsnedders>
/me just uses Ctrl + C to quit vim :)

[13:49:00.0000] <jgraham>
BenMillard: Do you have an eta on the tables comparison thing?

[13:50:00.0000] <jgraham>
(don't worry if not, I'm just interested)

[13:51:00.0000] <BenMillard>
jgraham, my deadline is 1st December 2008

[13:52:00.0000] <BenMillard>
I expect to make it available before then so I gather feedback and correct it

[13:52:01.0000] <jgraham>
BenMillard: OK. I may not be around much next week, epecially at the end but if you email me and I see it, I will have a look

[13:53:00.0000] <jgraham>
(assuming you are interested in my feedback of course :)

[13:55:00.0000] <BenMillard>
jgraham, of course I am!

[13:55:01.0000] <BenMillard>
jgraham, by next week do you mean the next 7 days, or the 7 days after that?

[13:55:02.0000] <BenMillard>
s/next 7 days/next 6 days/

[13:55:03.0000] <jgraham>
I mean the 7 days leading up to the 1st

[13:55:04.0000] <jgraham>
I should have said this week :)

[13:56:00.0000] <gsnedders>
jgraham: that's still ambiguous

[13:56:01.0000] <BenMillard>
jgraham, in that case I'll throw it online right now (in it's very rough state) so you can take a look

[13:56:02.0000] <jgraham>
Well the week starts on a Sunday, right?

[13:56:03.0000] <gsnedders>
jgraham: Depends who you ask :)

[13:56:04.0000] <jgraham>
/me expects gsnedders to quote some ISO standard at him now

[13:57:00.0000] <gsnedders>
ISO 8601 defines what first day of the week is

[13:58:00.0000] <hsivonen>
jgraham: ISO weeks start on Monday

[13:58:01.0000] <jgraham>
Yeah, I just looked that up...

[13:58:02.0000] <gsnedders>
/me just looked that up too

[13:58:03.0000] <hsivonen>
It's confusing when airline or hotel booking UIs use non-ISO weeks

[13:59:00.0000] <gsnedders>
I might not know what the standard says, but I know what standard to look at :P

[13:59:01.0000] <jgraham>
OK, Ben, I will not be around much at the end of the nex ISO 8601 week

[13:59:02.0000] <gsnedders>
:D

[13:59:03.0000] <jgraham>
is that clear? :)

[13:59:04.0000] <hsivonen>
(when the UI has a date picker with a week per row)

[14:00:00.0000] <gsnedders>
jgraham: I guess :P

[14:00:01.0000] <jgraham>
hsivonen: Week starts Sunday is common in colloquial usage in the UK

[14:00:02.0000] <gsnedders>
jgraham: No, it isn't.

[14:00:03.0000] <jgraham>
gsnedders: Yes it is

[14:00:04.0000] <gsnedders>
jgraham: It's common in England. I can tell you Scotland is different.

[14:00:05.0000] <hsivonen>
jgraham: then your colloquial usage is non-conforming :-)

[14:01:00.0000] <jgraham>
gsnedders: In Scotland men wear what, and let's not beat about the bush here, is effectively a skirt

[14:01:01.0000] <jgraham>
so I'm not sure your opinion counts for much

[14:01:02.0000] <jgraham>
;)

[14:01:03.0000] <gsnedders>
It is not!

[14:01:04.0000] <hsivonen>
on a more serious note, the really crazy part is that there are countries whose week numbering is non-conforming to ISO 8601

[14:02:00.0000] <jgraham>
/me has a Scotish father

[14:02:01.0000] <jgraham>
Now non-conforming week numbering would indeed suck

[14:02:02.0000] <Dashiva>
hsivonen: Well, we wouldn't need two ways if nobody used the other way

[14:02:03.0000] <gsnedders>
/me has a Scottish grandfather

[14:03:00.0000] <jgraham>
Although week numbering strikes me as decidedly odd anyway

[14:03:01.0000] <hsivonen>
my parents had to deal with a Portuguese company that made up their own week numbers, which sucked

[14:03:02.0000] <BenMillard>
jgraham, gave you the link (it's not ready for public review yet)

[14:03:03.0000] <jgraham>
BenMillard: Thanks :)

[14:04:00.0000] <gsnedders>
jgraham: http://flickr.com/photos/jgraham/3053429409/ — That's awesome

[14:05:00.0000] <jgraham>
gsnedders: I had forgotten that I had just uploaded those

[14:05:01.0000] <gsnedders>
:D

[14:05:02.0000] <jgraham>
I need to write descriptions

[14:05:03.0000] <jgraham>
:)

[14:05:04.0000] <gsnedders>
Needless to say, it is still awesome

[14:07:00.0000] <jgraham>
Thanks :)

[14:09:00.0000] <hsivonen>
jgraham: are you still in NZ?

[14:16:00.0000] <jgraham>
hsivonen: No, sadly not

[14:16:01.0000] <jgraham>
/me is just being rather slow at uploading photos

[14:17:00.0000] <gsnedders>
/me has been going through photos from August today

[14:26:00.0000] <hsivonen>
http://www.robertnyman.com/2008/11/21/ie-6-now-coming-to-a-cell-phone-near-you/

[14:30:00.0000] <KrocCamen>
It’s almost like they’re unaware of the iPhone.

[14:30:01.0000] <jgraham>
Interesting. I guess it might not be popular enough for people to target it 

[14:30:02.0000] <jgraham>
So there may be no practical effect

[14:30:03.0000] <gsnedders>
http://blogs.msdn.com/windowsmobile/archive/2008/11/11/internet-explorer-mobile-6.aspx which is what that cites says no such thing

[14:31:00.0000] <hsivonen>
gsnedders: http://blogs.msdn.com/windowsmobile/archive/2008/11/11/internet-explorer-mobile-6.aspx#9063740

[14:31:01.0000] <KrocCamen>
If anybody were to enter the mobile IE team’s office with an iPhone, I think they’d die.

[14:32:00.0000] <gsnedders>
hsivonen: Nowhere does it say that it is based on WinIE6

[14:32:01.0000] <KrocCamen>
Why would anybody make that assumption? Mobile IE is a different product.

[14:33:00.0000] <hsivonen>
gsnedders: well, the comment I linked to sure implies that it's mainly IE6-based except for some components of which jscript is mentioned specifically as coming from IE8

[14:33:01.0000] <KrocCamen>
That’s like saying Opera Mini 4 is based on Opera v4.

[14:33:02.0000] <gsnedders>
hsivonen: I don't read it like that

[14:33:03.0000] <hsivonen>
KrocCamen: rumor has it that the currently shipping Mobile IE is based on Desktop IE 4

[14:33:04.0000] <hsivonen>
which is really sad

[14:33:05.0000] <jgraham>
KrocCamen: http://news.cnet.com/8301-13860_3-10039152-56.html says it is based on IE6, which seems to be the source cited for that claim

[14:34:00.0000] <KrocCamen>
It would make sense in a way. IE4 was a major structural stabilisation of IE for modern content.

[14:34:01.0000] <jgraham>
s/ KrocCamen / gsnedders /

[14:37:00.0000] <hsivonen>
gsnedders: if it isn't based on IE6, the naming is a huge PR blunder

[14:37:01.0000] <gsnedders>
Yeah, totally.

[14:37:02.0000] <KrocCamen>
IE Mobile Explorer Pro Win 2008 Version.

[14:37:03.0000] <KrocCamen>
Totally.

[14:39:00.0000] <KrocCamen>
Spot-quiz. Name one good Microsoft product name.

[14:39:01.0000] <BenMillard>
I've used Mobile IE and it's less capable than IE6

[14:39:02.0000] <jgraham>
Word

[14:39:03.0000] <doublec>
Bob

[14:39:04.0000] <BenMillard>
Messenger

[14:39:05.0000] <KrocCamen>
That’s just generic, it’s not a name.

[14:39:06.0000] <BenMillard>
Explorer

[14:39:07.0000] <BenMillard>
oh...

[14:40:00.0000] <hsivonen>
Word, Excel, Windows, Internet Explorer seem like great product names

[14:40:01.0000] <BenMillard>
so what's the name of Explorer?

[14:40:02.0000] <jgraham>
The success that Microsoft has had in taking generic words and turning them into trademarks is astonishing

[14:40:03.0000] <hsivonen>
particularly Word, Windows and Internet Explorer

[14:40:04.0000] <hsivonen>
jgraham: indeed

[14:40:05.0000] <KrocCamen>
A good name is a brand of it’s own that people feel a affinity to. Firefox is one modern hero for example.

[14:40:06.0000] <jgraham>
I think windows less so, but word and internet explorer for sure

[14:41:00.0000] <KrocCamen>
Internet Explorer is a job description, not name XD

[14:41:01.0000] <jgraham>
Internet Explorer makes people think that the Blue E is the same as The Internet

[14:41:02.0000] <KrocCamen>
It may as well be called "Internet Browser". Or "The Blue E".

[14:42:00.0000] <hsivonen>
/me has wondered if Windows XP was meant to subliminally exploit the brand image of Jesus

[14:42:01.0000] <jgraham>
I think Powerpoint is likely also a good name although it does sound like it came out of a focus group

[14:43:00.0000] <jgraham>
hsivonen: ?

[14:43:01.0000] <KrocCamen>
I’ll give you that one Powerpoint is a reasonably good name. Excel too.

[14:43:02.0000] <hsivonen>
jgraham: http://en.wikipedia.org/wiki/Chi_Rho

[14:44:00.0000] <BenMillard>
jgraham, an ex-girlfriend who used Firefox would say things like "wow...the Internet takes ages to start these days" so I think that's a mental model problem rather than a triump of branding :)

[14:45:00.0000] <jgraham>
hsivonen: Ah. World religions would not be my chosen topic on Mastermind

[14:45:01.0000] <jgraham>
:)

[14:45:02.0000] <jgraham>
BenMillard: Sure, but having "Internet" in the name must help surely

[14:46:00.0000] <KrocCamen>
Mental model is spot on. If anything, these brands and names are the problem.

[14:46:01.0000] <KrocCamen>
The LISA just had "text editor", it was more document centric than application centric.

[14:47:00.0000] <KrocCamen>
A truly good OS/UI wouldn’t have applications, it’d have psuedo applications via folder views.

[14:47:01.0000] <KrocCamen>
In a spatial browser, the "Music" folder would be a music player. The "Pictures" folder would be the picture app. That’s how I would imagine it.

[14:49:00.0000] <jgraham>
KrocCamen: I have several "Picture" apps and I find the OSX behaviour of trying to have magic folders tied to applications irritating

[14:49:01.0000] <KrocCamen>
Magic folders? Oh you mean bundles.

[14:49:02.0000] <jgraham>
So whilst I used to drink the document centric koolaid, I am not sure I am a believer anymore

[14:49:03.0000] <KrocCamen>
No, I don’t mean bundles that *launch* an app. I mean the folder view is the app. You open Music, and the folder view has extra buttons for play &c.

[14:49:04.0000] <jgraham>
KrocCamen: I mean things like the iphoto pictures folder

[14:50:00.0000] <jgraham>
Not the application bundles which are a fine idea

[14:52:00.0000] <KrocCamen>
I would combine the two. Take BeOS, and have application folders that are themselves the data folders. I would imagine in OS X terms, that you’d double click on ~/Music and you’d get finder still, but it’d be essentially iTunes. No database file, all searching/metadata done through the FS.

[14:52:01.0000] <KrocCamen>
(Like BeOS mail, a truly elegant app)

[14:54:00.0000] <BenMillard>
KrocCamen, IE uses actual files and folders for its Favourites and History. In IE8, address bar autocomplete (and similar features) use Windows Search

[14:54:01.0000] <KrocCamen>
With xattr / Spotlight, we really shouldn’t need these db files for each app. iTunes / Mail &c.

[14:54:02.0000] <jgraham>
How does it work if I have a file that I want to edit in either Lightroom or Photomatix? How does it work if, like Photomatix, I need to work on multiple files at the same time

[14:54:03.0000] <KrocCamen>
BenMillard - a pretty good example in a ways.

[14:55:00.0000] <KrocCamen>
jgrahm - that’s what needs to be debated. Where does brand / app and document divide.

[14:56:00.0000] <KrocCamen>
In an ideal world where we didn’t have competition or profit goals we wouldn’t need brands. Of course, that’s going to never happen. But for a very brief period, it did exist. Right back with the LISA/Mac.

[14:58:00.0000] <takkaria>
ideal worlds would have plenty of competition for me. :)

[14:58:01.0000] <KrocCamen>
(and XEROX Parc before that of course. All the pioneers of UI never thought in term of brands/products. I find that particularly fascinating)

[15:02:00.0000] <KrocCamen>
One good argument against the application-centric model is Adobe Reader.

[15:02:01.0000] <KrocCamen>
I feel like I could roll over and die having to use that app. It offends on so many levels.

[15:02:02.0000] <KrocCamen>
But Apple preview is a very document-centric app, being mostly transparent to the user

[15:03:00.0000] <KrocCamen>
Having a brand forced down your throat just to view a simple document is anti-user.

[15:09:00.0000] <BenMillard>
KrocCramen, this is the shabby homebrew VB6 text editor I made and use (based on standard MDIForm/MDIChild and RichTextBox): http://projectcerbera.com/tools/textstudio/

[15:09:01.0000] <BenMillard>
jgraham has a point about automatic indention...it doesn't have that but I should add it

[15:10:00.0000] <BenMillard>
and syntax highlighting, and UTF-8 support, but those 2 are beyond my abilities

[15:11:00.0000] <KrocCamen>
Looks alright. I used an OCX control for indentation and syntax highlighting

[15:11:01.0000] <KrocCamen>
I forget it’s name though

[15:12:00.0000] <BenMillard>
given your strong views on UI, I feel blessed to have escaped with a "looks alright" :)

[15:17:00.0000] <KrocCamen>
Art is personality.

[15:19:00.0000] <KrocCamen>
UX for the masses is not the same as UX for solving problems :)

[15:25:00.0000] <jgraham>
BenMillard: Good indentation is a hard problem: http://steve-yegge.blogspot.com/2008_03_01_archive.html (just under the orange heading that says "Missing highlighting")

[15:43:00.0000] <BenMillard>
jgraham, oh yeah doing it properly like that is WAY beyond me!

[15:43:01.0000] <BenMillard>
jgraham, I would have followed the VB6 IDE convention, which always assumes the "same indentation as the previous line" when you press Return

[15:44:00.0000] <BenMillard>
jgraham, with Tab increasing indention by 1 level (and applies to multiple lines of you select multiple lines) and Shift+Tab decreasing in the same way

[15:45:00.0000] <KrocCamen>
Sometimes computer’s shouldn’t overthink. It prevents the user expressing their artistic nature.

[15:52:00.0000] <jcranmer>
s/r's/rs/

[15:53:00.0000] <jcranmer>
er, s /r’s/rs/


2008-11-24
[23:52:00.0000] <zcorpan>
i wonder why validator.w3.org just says "HTML5" rather than "(X)HTML5+ARIA"

[23:57:00.0000] <zcorpan>
"The DOCTYPE Declaration for "HTML5" has been inserted at the start of the document, but even if no errors are shown below the document will not be Valid until you add the new DOCTYPE Declaration." it says for XHTML5

[23:59:00.0000] <zcorpan>
hmm i guess the proper label would be "HTML5+ARIA / XHTML5+ARIA, SVG 1.1 plus MathML 2.0 (experimental)"

[00:01:00.0000] <Hixie>
"html5" seems like a more understandable label :-)

[01:40:00.0000] <annevk3>
http://standardssuck.org/uhuru-peak

[01:47:00.0000] <hsivonen>
annevk3: how does the clothing stuff work when one climbs to Uhuru peak? is there documentation on how warm clothes you need to bring and then you add layers as you go higher? You start walking in a *very* warm climate, I presume.

[01:51:00.0000] <annevk3>
there's no real documentation, but it basically involves adding layers once you go higher :)

[01:52:00.0000] <annevk3>
above 4000 or so starts freezing at night and above 4000-5000 it will always be cold and there's a possibility for snow

[01:53:00.0000] <annevk3>
(in Europe the heights would be lower, but you still have the same effect)

[02:01:00.0000] <zcorpan>
my public-html folder had 100 unread emails this morning and only 3 or so were worth reading

[02:01:01.0000] <zcorpan>
why are everyone so obsessed about splitting the spec

[02:02:00.0000] <annevk3>
it's an easy thing to have an opinion about

[02:02:01.0000] <hsivonen>
zcorpan: spec organization seems like the ultimate bikeshed

[02:02:02.0000] <annevk3>
I have never actually tried to find out if it's true, but I've the feeling that simple matters attract way more comments than complex ones

[02:09:00.0000] <Philip`>
Maybe people care because the organisation of the spec has a fundamental impact on its readability for them

[02:10:00.0000] <zcorpan>
the discussion about spec organisation has a fundamental impact on my readability of the list

[02:10:01.0000] <zcorpan>
or well, not spec organisation discussions in particular but bikesheds in general

[02:11:00.0000] <hsivonen>
Philip`: maybe I'm just cynical, but I think some people might be discussing the organization without testing the readability by trying to read the spec

[02:12:00.0000] <annevk3>
very annoying that Flickr doesn't store the date the video was actually taken

[02:58:00.0000] <virtuelv>
JohnResig: you around?

[05:59:00.0000] <Philip`>
Gmail's automatic service suggestion thing is great

[05:59:01.0000] <Philip`>
For the thread "setting HTMLMediaElement volume and playbackRate attributes", it suggests "Would you like to... Track DHL package 0000000000"

[07:43:00.0000] <Philip`>
If I write a table with headings along the top and other headings along the left edge, and the top-left cell is empty because there's nothing for it to say, should that top-left cell be a <td> or a <th> or something else?

[07:44:00.0000] <gsnedders>
Philip`: td

[07:44:01.0000] <Philip`>
gsnedders: But it's not a data cell

[07:45:00.0000] <gsnedders>
/me opens the damned spec

[07:46:00.0000] <gsnedders>
It should be neither tr nor td.]

[07:46:01.0000] <gsnedders>
*th or td

[07:46:02.0000] <gsnedders>
Which is rather problematic.

[07:47:00.0000] <Philip`>
/me chooses th, for no good reason (other than the necessity of at least choosing something)

[07:49:00.0000] <Philip`>
Oh, maybe td would be better because then it's an #empty-data-cell

[07:50:00.0000] <takkaria>
the spec should provide feedback on that

[07:58:00.0000] <jgraham>
It should be designed to make no practical difference either way

[07:58:01.0000] <jgraham>
Since authors will choose at random raher than based on the spec

[07:59:00.0000] <gsnedders>
ooo…

[07:59:01.0000] <gsnedders>
new standards suck!

[07:59:02.0000] <Philip`>
gsnedders: So do old ones

[07:59:03.0000] <Philip`>
only moreso

[08:00:00.0000] <jgraham>
Does anyone know anything about parking in copenhagen? Like whether there are any reasonably secure car parks?

[08:00:01.0000] <jgraham>
/me realises this is something of a long shot

[08:01:00.0000] <gsnedders>
/me forwards the question

[08:02:00.0000] <Philip`>
jgraham: The obvious place to look is http://www.parking.dk/

[08:02:01.0000] <jgraham>
Philip`: That didn't provide an obvious answer to my quesion alhough I may not have been looking hard enough

[08:03:00.0000] <Philip`>
parking.co.uk is much less helpful :-(

[08:04:00.0000] <gsnedders>
jgraham: "define: reasonable…"

[08:04:01.0000] <gsnedders>
jgraham: "there are underground parks, and I would assume that they have CCTV"

[08:04:02.0000] <gsnedders>
/me pokes fangel, ahrd

[08:04:03.0000] <gsnedders>
*hard

[08:04:04.0000] <fangel>
<- copenhagen'er..

[08:04:05.0000] <gsnedders>
/me runs away

[08:04:06.0000] <gsnedders>
/me has carried out his usefulness by bullying fangel in here

[08:05:00.0000] <fangel>
yes.. that is all you're good for really, gsnedders..

[08:05:01.0000] <jgraham>
Oh, well that sounds quite promising

[08:05:02.0000] <gsnedders>
/me realizes he ought to get train tickets to Cambridge

[08:06:00.0000] <Philip`>
/me suddenly discovers Firefox 3.1 has a fancy ctrl-tab menu, but also discovers it has a stupid unusable ordering of tabs and can't find a way to correct it :-(

[08:07:00.0000] <fangel>
jgraham: so.. I heard rumours of questions.. horrible questions regarding my lovely hometown of Copenhagen..

[08:08:00.0000] <blooberry>
philip`: is that new in 3.1?

[08:08:01.0000] <jgraham>
fangel: Yeah, I have to stay somewhere overnight located roughly where Copenhagen happens to be and will have a reasonable amount of stuff in the car that I would prefer not to be stolen

[08:09:00.0000] <jgraham>
Not hugely valuble, but it would be a pain

[08:09:01.0000] <jgraham>
So one possibility is to stay in Copenhagen and park in some kind of secure car park, if such a thing exists

[08:10:00.0000] <jgraham>
Another possibility would be to stay somewhere not in Copenhagen and assume that not being in a city mitigates the risk

[08:10:01.0000] <Philip`>
blooberry: I assume so, since I haven't seen it in 3.0

[08:11:00.0000] <jgraham>
Philip`: Which build are you running? It seems to have disappeared a few days ago for me

[08:12:00.0000] <Philip`>
jgraham: 20081111

[08:12:01.0000] <fangel>
sorry... phone

[08:13:00.0000] <fangel>
there are underground parkingplaces which I would assume has CCTV..

[08:13:01.0000] <fangel>
parking lots.. parking.. whats the word?

[08:13:02.0000] <jgraham>
in en-gb carparks

[08:13:03.0000] <fangel>
and maybe some of the larger hotels has their own underground carpark..

[08:14:00.0000] <jgraham>
The larger hotels tend to be rather expensive...

[08:14:01.0000] <fangel>
but as long as whatever stuff your hauling is in the boot, it should be reasonable safe to keep it on the street overnight..

[08:14:02.0000] <jgraham>
I guess the whole back of the car will be full

[08:15:00.0000] <jgraham>
Are the underground carparks the ones shown on http://www.p-henvisning.dk/

[08:16:00.0000] <fangel>
yeah.. if you can see it though the windows, it might be a bit of a temptation.. but I don't think car-break-ins are that bad if you park it on a random street.. but well - there are pay-by-the-hour underground carparks..

[08:16:01.0000] <fangel>
two sec..

[08:17:00.0000] <fangel>
iirc, the P with the ^ over it signifies that it is a covered parking space (that is, multistory or underground)

[08:18:00.0000] <jgraham>
Ah, interesting

[08:18:01.0000] <fangel>
and so you have to pay, and then they tend to have cctv..

[08:19:00.0000] <jgraham>
fangel: thanks

[08:20:00.0000] <fangel>
no problem..

[08:20:01.0000] <fangel>
long car journey, or what?

[08:21:00.0000] <jgraham>
Yeha

[08:21:01.0000] <jgraham>
*Yeah

[08:22:00.0000] <gsnedders>
jgraham: Why not take a ferry from somewhere like Newcastle to Goteberg?

[08:23:00.0000] <jgraham>
gsnedders: Harwich-Ejsberg (sp?) made the most sense

[08:23:01.0000] <gsnedders>
Ejsberg?

[08:23:02.0000] <gsnedders>
Esbjerg?

[08:23:03.0000] <gsnedders>
/me is surprised that made the most sense

[08:25:00.0000] <fangel>
I've taken a return trip with the Harwich-Esbjerg ferry once upon a time..

[08:25:01.0000] <fangel>
where are you going from and too jgraham?

[08:25:02.0000] <gsnedders>
fangel: to Oslo

[08:26:00.0000] <gsnedders>
fangel: not sure quite where from

[08:26:01.0000] <gsnedders>
Not to Oslo

[08:26:02.0000] <gsnedders>
/me is too tired

[08:26:03.0000] <Philip`>
Is there a way to do something like onhashchange that works in all current browsers?

[08:26:04.0000] <gsnedders>
fangel: to Linköping

[08:26:05.0000] <gsnedders>
(the other place where Opera has major offices)

[08:27:00.0000] <fangel>
;)

[08:28:00.0000] <fangel>
it's a 5 hour drive from Copenhagen to Linköping.. 

[08:28:01.0000] <fangel>
and a 3 hour drive from Esbjerg to Copenhagen..

[08:29:00.0000] <jgraham>
Yeah, the ferry gets in at about 1pm so doing it all in a day is not possible but splitting it over two days is easy

[08:43:00.0000] <BenMillard>
Philip`, either gets used for that in the wild (re: http://krijnhoetmer.nl/irc-logs/whatwg/20081124#l-144)

[08:43:01.0000] <BenMillard>
Philip`, I prefer using <td> since it shouldn't act as a header for other cells

[09:02:00.0000] <nlg>
Im writing a script to traverse a webpage to change all links. Is LXML enough for this task or would you recommend using html5lib as well?

[09:03:00.0000] <gsnedders>
nlg: There are issues with lxml's serializer, but it's parser works fine on most conforming pages

[09:04:00.0000] <nlg>
gsnedders: Is there a way to "cleanup" mallformed pages?

[09:04:01.0000] <gsnedders>
nlg: "cleanup" in what way

[09:04:02.0000] <BenMillard>
in these sense of HTML Tidy?

[09:04:03.0000] <BenMillard>
s/these/the/

[09:06:00.0000] <gsnedders>
(for xref and toc and the like of the spec lxml's parser and html5lib's serializer are used)

[09:06:01.0000] <nlg>
Hm, I´ll try the lxml. Thanks

[09:11:00.0000] <Philip`>
Is there a term for something that's kind of like JSON but that totally ignores the standard, and just relies on Python repr and JS eval being compatible in typical cases?

[09:11:01.0000] <Philip`>
(Preferably some term more technical than "laziness")

[09:12:00.0000] <gsnedders>
:D

[09:12:01.0000] <gsnedders>
Philip`'s dream?

[09:20:00.0000] <gsnedders>
/me wonders what the hell to put on skills section of CV

[09:20:01.0000] <jgraham>
Philip`: PYON? It sounds catchy

[09:32:00.0000] <takkaria>
and before now, I thought I'd seen some fairly bad bikeshedding

[09:35:00.0000] <Philip`>
jgraham: That sounds like a good name, particularly since it reminds me of Peon and hence of Warcraft

[09:35:01.0000] <Philip`>
As far as I can tell, it's quite like JSON except it uses single quotes instead of double quotes

[09:49:00.0000] <Dashiva>
Philip`: Source code? :)

[09:59:00.0000] <gsnedders>
in CSS21, clear only applies to block-level elements, but is it conforming to be set on an inline element?

[10:00:00.0000] <BenMillard>
gsnedders, if it works then just do it. :)

[10:00:01.0000] <gsnedders>
BenMillard: When I'm writing a CV it probably would be better for it to be conforming :)

[10:01:00.0000] <Dashiva>
Just because it doesn't apply doesn't mean it isn't legal

[10:01:01.0000] <gsnedders>
Dashiva: That's my thought.

[10:02:00.0000] <Philip`>
/me would hope potential employers would care about more substantial issues than how pedantically you wrote your CV's markup :-p

[10:02:01.0000] <Dashiva>
My CV uses only very basic CSS

[10:03:00.0000] <Dashiva>
left margin on body, bottom margin on dd :)

[10:04:00.0000] <Dashiva>
I did close optional tags, though, just in case the reader is a twit

[10:07:00.0000] <BenMillard>
gsnedders, there's a note here: http://www.w3.org/TR/CSS21/visuren.html#propdef-clear

[10:08:00.0000] <gsnedders>
/me spots an RFC2119 keyword inside an informative note

[10:08:01.0000] <gsnedders>
/me implodes

[10:08:02.0000] <BenMillard>
"may" is present as well :)

[10:08:03.0000] <BenMillard>
have you searched www-style?

[10:09:00.0000] <hsivonen>
what's the deal with wanting to use GET for ping?

[10:10:00.0000] <Dashiva>
As I recall, different interpretations of vague parts of the HTTP spec

[10:12:00.0000] <Dashiva>
JR in particular talked a lot about POST not being "safe"

[10:13:00.0000] <hsivonen>
I find if baffling that people working on HTTP suggest using GET for something that isn't idempotent

[10:13:01.0000] <hsivonen>
"safe" is a rathole

[10:13:02.0000] <Dashiva>
Well, that's what peopel tried to tell him

[10:13:03.0000] <hsivonen>
idempotent is a more tractable criterion

[10:13:04.0000] <Dashiva>
It starts here: http://lists.w3.org/Archives/Public/public-html/2007Oct/0337.html

[10:15:00.0000] <BenMillard>
gsnedders, I've skimed through some search results but not found people discussing the conformance (only discussing what works versus what makes sense to specify): http://www.w3.org/Search/Mail/Public/search?keywords=%22clear+property%22&hdr-1-name=subject&hdr-1-query=&index-grp=Public__FULL&index-type=t&type-index=www-style

[10:16:00.0000] <Dashiva>
hsivonen: One of his arguments is that ping has side effects on two third parties, not on the user itself, so it's idempotent as far as that user is concerned

[10:17:00.0000] <gsnedders>
hit counters aren't idempotent, yet normally use GET

[10:17:01.0000] <Dashiva>
gsnedders: That's a bug in the world

[10:18:00.0000] <BenMillard>
gsnedders, so it might be worth you sending an e-mail to them, if you can't find an answer (now dinnertime for me)

[10:28:00.0000] <hsivonen>
Dashiva: but idempotence from the user POV isn't what matters for HTTP working. It's idempotence from the POV of server and caches

[10:29:00.0000] <Dashiva>
Preaching to the choir :)

[11:16:00.0000] <gsnedders>
#a-phrase-or-paragraph-with-an-alternative-graphical-representation:-charts,-diagrams,-graphs,-maps,-illustrations

[11:16:01.0000] <gsnedders>
That @id is too long

[11:55:00.0000] <Philip`>
gsnedders: You should trim all ids to the shortest unambiguous prefix

[12:00:00.0000] <gsnedders>
Philip`: That makes them more liable to change, though

[12:03:00.0000] <Philip`>
gsnedders: You should map each id onto a 64-bit prefix of its MD5

[12:48:00.0000] <gsnedders>
Philip`: That makes them quite unguessable

[13:06:00.0000] <hsivonen>
soo... in terms of content models, is there any case left where a container only takes embedded content?

[13:07:00.0000] <hsivonen>
did the concept of embedded content become irrelevent for content models?

[13:07:01.0000] <Hixie>
i believe so but i'm not sure. click on the definition of the term and see what refers to it

[13:08:00.0000] <hsivonen>
Hixie: can you think of any reason of keeping "embedded content" in the schema structure?

[13:08:01.0000] <Hixie>
not off-hand

[13:10:00.0000] <gsnedders>
Hixie: How about off-foot?

[13:11:00.0000] <hsivonen>
I guess I'll flatten it out then

[13:11:01.0000] <hsivonen>
this should fix the bug that showed up on anne's fronteers presentation

[13:25:00.0000] <hsivonen>
/me mumbles about ant builds sucking compared to letting Eclipse do its thing

[13:33:00.0000] <gsnedders>
for fragments does @id or @name take precedence?

[15:31:00.0000] <jgraham>
All these emails about splitting the spec are beyond a joke

[15:37:00.0000] <hsivonen>
denial of productivity attack by bureaucracy trolling

[15:37:01.0000] <Dashiva>
It wouldn't be w3c without it, would it?

[15:38:00.0000] <jgraham>
hsivonen: Seriously

[15:40:00.0000] <Hixie>
luckily it happened around the same time i was feeling like a break from editing

[15:40:01.0000] <jgraham>
I really wish it wouldn't make things worse if I wrote to the mailing list that this discussion is an enormous bikeshed and that the very people most often touted as beneficiries from splitting the spec are also those most likely to be put off from contributing due to the volume of email created by this discussion

[15:40:02.0000] <jgraham>
i.e. authors

[15:40:03.0000] <Hixie>
personally i just wish that the chairs would stop in and make some sort of process statement

[15:40:04.0000] <Dashiva>
Not to mention that most people seem to ignore the fact that authors only read specs if they want the actual spec, otherwise they read authoring guides

[15:41:00.0000] <Hixie>
either putting it to vote, or rejecting the proposals, or picking one, or something

[15:42:00.0000] <jgraham>
The point about authors and authoring guides has been made innumerable times

[15:42:01.0000] <Philip`>
Are desktop browser developers unusually competent relative to other UA developers? It seems nobody else believes that it's actually not that hard to learn how to read the spec and to ignore the bits that aren't relevant to you

[15:43:00.0000] <jgraham>
Parse error: Too many negatives

[15:44:00.0000] <Dashiva>
Implementation limit exceeded: Doing more than one thing simultaneously

[15:44:01.0000] <jgraham>
Oh I see

[15:44:02.0000] <jgraham>
I was thinking you were saying the opposite thing which is why it didn't make sense

[15:44:03.0000] <hsivonen>
Philip`: well, given e.g. Roy Fielding's posts, one might think that browser developers are considered unusually *in*competent :-)

[15:44:04.0000] <Dashiva>
It could be a worry that if the browser parts get to be part of the main document, they'll get undue weight

[15:45:00.0000] <jgraham>
(that UA developers much be unusually dim because they can only cope with things that are all lumped together)

[15:45:01.0000] <ohtermaciej>
we need all the help we can get

[15:45:02.0000] <Philip`>
jgraham: I wouldn't dare insult UA developers in this channel - I'd go elsewhere and do it behind their backs

[15:45:03.0000] <Philip`>
Uh

[15:45:04.0000] <ohtermaciej>
top stop writing such crappy software that everyone hates

[15:45:05.0000] <Philip`>
s/UA/desktop browser/

[15:46:00.0000] <Hixie>
i was utterly baffled by roy's e-mail where he point blank stated that he's cleverer than everybody else and that's why i should listen to him and ignore everyone else (http://lists.w3.org/Archives/Public/public-html/2008Nov/0316.html)

[15:46:01.0000] <Philip`>
So I'm just wondering why while desktop browser developers who've read and implemented parts of the spec seem to be saying it's not a problem, other people say it's too hard for them to write non-browser UAs when the spec is so large and mixed up, or something

[15:47:00.0000] <Hixie>
("Because I have more experience...")

[15:47:01.0000] <jgraham>
Hixie: Yeah at that point I gave up on him

[15:47:02.0000] <Dashiva>
/me expects a darn good lastweek episode from this chat

[15:48:00.0000] <hsivonen>
as a non-browser UA developer, I wouldn't trust any spec split that claimed that some part didn't apply to validators

[15:48:01.0000] <hsivonen>
I'd read all the parts of the spec to be sure

[15:48:02.0000] <Philip`>
Is it still considered trolling if you're trying to troll not the people you're talking to, but anonymous observers who are reading the conversation hours later?

[15:48:03.0000] <jgraham>
I mean clearly he is a bright guy butyou would have thought that bright people would realise that argument by authority is relly weak

[15:49:00.0000] <ohtermaciej>
Philip`: at least some non-browser UA developers  don't seem to have a problem with the HTML5 spec as currently arranged

[15:49:01.0000] <takkaria>
the point that needs to be made is that if someone actually finds the spec organisation hindering in practice, then that's better grounds for doing it than lots of people arguing that it's hindering in theory

[15:49:02.0000] <ohtermaciej>
namely the ones who have worked on non-browser implementations

[15:50:00.0000] <Hixie>
Lachy: yt?

[15:50:01.0000] <ohtermaciej>
I don't believe we have heard from anyone who attempted to implement any conformance class of HTML5 and was in practice hindered by the organization of the spec

[15:50:02.0000] <hober>
I suspect Roy is a believer in the vast, browser-wing conspiracy, which makes me sad.

[15:51:00.0000] <ohtermaciej>
yeah, the Web would be so nice if only it weren't for all these browsers

[15:51:01.0000] <jgraham>
weeee the return of the vast browser-wing conspiracy

[15:51:02.0000] <Philip`>
One person [I'm too lazy to check in the archives] indicated that there was a practical hindrance, because he couldn't read and understand the real spec but he could provide feedback on the simplified markup spec

[15:51:03.0000] <ohtermaciej>
we should just get rid of those, and then the Web will be fine

[15:51:04.0000] <jgraham>
Can we have a secret cabal please?

[15:51:05.0000] <ohtermaciej>
Philip`: did the person name the implementation they were working on?

[15:51:06.0000] <Hixie>
hober: i've been trying to find out what his beliefs are in the thread i'm having with him -- so far all i've established is that he thinks he, quote, has more experience [...] than just about anyone else, unquote.

[15:52:00.0000] <Philip`>
ohtermaciej (sic): I think they were speaking as an author rather than as an implementor

[15:52:01.0000] <Hixie>
Philip`: iirc that person also said that the plan to have filtered views at last call would work for him

[15:53:00.0000] <othermaciej>
(sorry, not my usual computer today)

[15:53:01.0000] <Philip`>
jgraham: https://www.blogger.com/comment.g?blogID=7538849950414416888&postID=7494409664460936334 - the problem is it's not really a secret cabal when we muse aloud about getting cabalistic T-shirts :-(

[15:54:00.0000] <Philip`>
(By the way, why is Blogger's comment system so antiquated?)

[15:55:00.0000] <Dashiva>
About Roy, there was a comment during the URI thing that LEIRI would meet a "wall of flames" if it ever was proposed formally. That kind of culture might consider it not argument from authority, but something else. :)

[15:56:00.0000] <othermaciej>
Philip`: that wouldn't be a counterexample to my statement, then (though it is possible that for a large number of people the organization of the HTML5 spec is so harmful that they are unable to even start an implemetnation)

[15:57:00.0000] <Philip`>
I remember some Microsoft people (who count as implementors) complaining that the spec was killing too many trees when they printed out a copy for each developer

[15:58:00.0000] <BenMillard>
Philip`, think I heard that at TPAC 2008 from one of them, too.

[15:58:01.0000] <BenMillard>
it seemed more like a jest than a serious complaint, though

[15:58:02.0000] <Philip`>
(Maybe they print it out because IE can't load it without freezing?)

[15:58:03.0000] <Dashiva>
Interesting that the IETF considers HTML 2.0 a moving target


2008-11-25
[16:02:00.0000] <jgraham>
I wonder if users really do distinguish "safe" and "unsafe" UI elements. I seriously doubt it but the HTTP people seem to believe it religiously. It would be nice to see actual evidence

[16:03:00.0000] <Dashiva>
jgraham: Maybe someone should tell them about CSS

[16:03:01.0000] <jgraham>
Dashiva: It might break their hearts

[16:04:00.0000] <BenMillard>
jgraham, I typically act on what I expect to happen rather than how it was presented or how it works.

[16:04:01.0000] <BenMillard>
so if I click something which says "Confirm My Order" then I expect it to do that, whether it's an <a href> using GET or an <input type="submit"> using POST

[16:05:00.0000] <jgraham>
BenMillard: Agreed. I tend to assume that links marked "Delete" will be non-idempotent...

[16:06:00.0000] <Philip`>
My bank uses buttons that don't look like real buttons, though they are quite buttony - they're just black text on a white background, with a single-pixel black border around them

[16:06:01.0000] <jgraham>
Which it turns out that my university webmail uses...

[16:07:00.0000] <jgraham>
As does squirrelmail

[16:09:00.0000] <jgraham>
flickr uses identical buttons for things that are links as for things that are actions

[16:09:01.0000] <Philip`>
It's a bit peculiar how that webmail system inserts an incrementing ID value into every request via the link URLs

[16:09:02.0000] <jgraham>
/me concludes that any user who believed HTTP rather than using their common sense would quickly come to grief

[16:11:00.0000] <takkaria>
Philip`: what, squirrelmail?

[16:12:00.0000] <Philip`>
takkaria: No, jgraham's university webmail

[16:12:01.0000] <jgraham>
Philip`: You mean the last number in the link?

[16:13:00.0000] <Philip`>
(I think the university wrote the whole thing themselves)

[16:13:01.0000] <Philip`>
(Also they wrote Exim themselves)

[16:13:02.0000] <jgraham>
s/last/first/

[16:14:00.0000] <Philip`>
jgraham: If by "number" you mean "string of letters", then yes

[16:14:01.0000] <jgraham>
"Number base 26"

[16:15:00.0000] <Dashiva>
base 26, rot 10?

[16:15:01.0000] <Philip`>
(The AAAD in /session/abc123//AAAD@list@12345@13456)

[16:15:02.0000] <jgraham>
Yeah

[16:15:03.0000] <Philip`>
(where all links on the page go to URLs with "AAAE" in them)

[16:16:00.0000] <Philip`>
It seems a quite server-stateful model

[16:17:00.0000] <Dashiva>
Gotta love sites that keep state in URLs

[16:17:01.0000] <Dashiva>
I tried shopping on such a site

[16:18:00.0000] <Dashiva>
If you open a new tab to look at two items, you create two separate shopping baskets

[16:18:01.0000] <Philip`>
Has anyone at the IETF noticed yet that not all their RFCs are US-ASCII?

[16:19:00.0000] <Philip`>
(e.g. http://tools.ietf.org/rfc/rfc303.txt seems to be ISO-8859-1)

[16:21:00.0000] <Dashiva>
Philip`: That's just a bug

[16:21:01.0000] <Dashiva>
As long as it's a bug, you don't have to care about it

[16:34:00.0000] <Philip`>
I hate how Google changes link hrefs to track me when I click on search results, because if I open a link in a tab then close it then copy the link's URL, it copies the mangled tracking version :-(

[16:58:00.0000] <takkaria>
IMAP requires an excessive amount of state-keeping

[17:00:00.0000] <roc>
I hate copying and pasting links from a search results page and embarrassing myself

[17:00:01.0000] <roc>
that's why I support <a ping>

[17:02:00.0000] <Philip`>
Is there a solution to the problem that you'd want to use ping for users with browsers that support it, and old-fashioned tracking solutions for everyone else?

[17:04:00.0000] <Dashiva>
You could "solve" that by doing UA sniffing, but the real problem would also include people with ping support, but disabled

[17:05:00.0000] <Philip`>
No it wouldn't, since you'd want to Not Be Evil and would send ping to users whose have explicitly disabled ping, so all that matters is whether their browser supports the feature

[17:05:01.0000] <Philip`>
s/whose/who/

[17:06:00.0000] <Dashiva>
You could also do both with a local @ping the first time the user visits, and then store "supports @ping" in a cookie if you get that

[17:06:01.0000] <roc>
so few people would disable it, they don't matter

[17:07:00.0000] <Philip`>
Dashiva: That sounds like a horrid hack

[17:08:00.0000] <Hixie>
the search engine results page links were the source of the first request

[17:11:00.0000] <Philip`>
roc: IE8 goes to some lengths to prevent third-party sites tracking you when it's in its private browsing mode, so I suppose Microsoft thinks people sometimes care about being tracked, and so those people would care about disabling ping too

[17:11:01.0000] <Philip`>
(This is separate from the don't-let-my-parents-know-what-I'm-looking-at aspects of the private browsing mode)

[17:12:00.0000] <Hixie>
disabling ping="" in the porn browsing mode seems quite reasonable

[17:13:00.0000] <roc>
yeah, and again, this is where ping enhances privacy

[17:13:01.0000] <Dashiva>
Philip`: Well, you could simplify it a lot by having opt-in ping support instead, but then most users wouldn't benefit :)

[17:15:00.0000] <Philip`>
So it does matter that sites can detect users who have browsers that support ping but who have chosen to disable ping because they don't want to be tracked, so the sites can intentionally not track them (instead of using their legacy fallback tracking mechanisms which they use for IE6/etc users)

[17:36:00.0000] <Hixie>
/me wonders how best to respond to this feedback:

[17:36:01.0000] <Hixie>
> I also believe that issuing an unsafe HTTP request in the context of

[17:36:02.0000] <Hixie>
> page navigation is in conflict with the web architecture.

[17:51:00.0000] <Lachy>
Hixie, there are some issues with the way iframe is defined that does make it slightly harder for authors who are uninterested in processing requirements to understand...

[17:51:01.0000] <Lachy>
e.g. the definition of the name attribute.: "The name  attribute, if present, must be a valid browsing context name."

[17:51:02.0000] <Hixie>
yeah iframe certainly isn't ideal

[17:51:03.0000] <Hixie>
i expect to find a bunch of things that need phrasing better when i go through and annotate everything for the views work

[17:51:04.0000] <Lachy>
That should say something about what the name attribute means from an authoring perspective and what it should be used for.

[17:52:00.0000] <Lachy>
but I intend to document everything like that in the authoring guide anyway

[17:52:01.0000] <Lachy>
and since I am writing that guide, I find these requests for splitting the spec to be quite redundant

[17:53:00.0000] <Lachy>
especially because it was essentially the same arguements last year that led to the creation of the authoring guide

[18:03:00.0000] <Lachy>
one thing that is really bugging me about all these requests to split the spec up into numerous specs for each feature is that it makes significantly harder to figure out which spec to look in for a particular feature; instead of just going to one spec and searching within it for the feature

[18:04:00.0000] <Lachy>
I used to have that problem all the time with the DOM specs, although with experience I've learned where to look for most things

[18:05:00.0000] <Hixie>
fwiw, the whatwg version of the spec will be one document for all of html5 regardless of what contortions i have to go through to generate the w3c documents

[18:05:01.0000] <Lachy>
to some extent, I still have that problem with the multipage version of this spec, which is why I always use the single page version

[18:05:02.0000] <Hixie>
(with exception for the storage and websockets bits)

[18:05:03.0000] <Lachy>
yeah, that's fair enough. I don't have a problem with shifting those to their own spec in the webapps group

[18:07:00.0000] <Lachy>
but I must say as an author, having the DOM APIs along side the element definitions makes it significantly easier and is why, in many cases, despite the claims to the contrary on the list

[18:07:01.0000] <Lachy>
s/and is why//

[18:09:00.0000] <Hixie>
yeah i personally think it's awesome

[18:10:00.0000] <Lachy>
for the authoring guide, I'm going to include the DOM APIs, but I'm going to include a view feature that shows information based on what the user wants and their skill level

[18:11:00.0000] <Lachy>
much like what you've said you'll make for the spec

[18:11:01.0000] <Lachy>
(we could probably even use the same script once it's written)

[18:13:00.0000] <Hixie>
yeah

[18:23:00.0000] <Lachy>
I don't understand what problem including PING in the request body solves. It seems like an unnecessary extra 4 bytes to send for no reason and the need to define a new text/ping MIME type

[18:24:00.0000] <Hixie>
reply on the thread

[18:34:00.0000] <Lachy>
this reply only says that you included it, but doesn't really give any justification for it other then perhaps Julian's spec purity concerns http://lists.w3.org/Archives/Public/public-html/2008Nov/0473.html

[18:35:00.0000] <Hixie>
the reasoning for adding it was to allow hosts to recognise them based on the content-type

[18:35:01.0000] <Lachy>
but I guess fighting the uselessness of it would be more work than accepting it and implementing it

[18:35:02.0000] <Lachy>
ok

[18:35:03.0000] <Hixie>
if we get a PING method it'll become moot and i'll remove the body

[22:00:00.0000] <Hixie>
so, of all the talk about keygen, there isn't a single person who has said how you would use this ke

[22:00:01.0000] <Hixie>
key

[23:51:00.0000] <hsivonen>
MikeSmith++ for moratorium

[23:52:00.0000] <MikeSmith>
hsivonen: sorry I didn't get around to it sooner

[23:54:00.0000] <MikeSmith>
We had a three-day weekend here, and I was offline for most it, trying to actually enjoy it.

[23:54:01.0000] <MikeSmith>
we seem to have an unfortunate pattern of some of the most contentious discussions happening on weekends

[23:55:00.0000] <MikeSmith>
anyway, I'll try to a much better job of staying on top of things (as far as keeping the discussions productive)

[23:59:00.0000] <MikeSmith>
hsivonen: speaking of productive discussions, the discussion with Karl Berry about the GNU error format seems like a bit of a clash of cultures

[00:00:00.0000] <hsivonen>
oops. I have an unreplied email in my inbox

[00:01:00.0000] <MikeSmith>
given the decision-making regime at GNU, and the constraints that Karl has expressed, I'm wonder how useful what we end up with will be

[00:03:00.0000] <MikeSmith>
hsivonen: anyway, it seems like you've volunteered to try to do their work for them. I hope you can manage to put together something that can make it past the RMS veto.

[00:15:00.0000] <MikeSmith>
/me reads hsivonen's reply

[00:16:00.0000] <MikeSmith>
hsivonen: it might be useful for further discussions to move back to whatever GNU mailing list it was where we started this

[00:17:00.0000] <hsivonen>
/me really doesn't like the concept that encoding is a property of the computer

[00:17:01.0000] <hsivonen>
MikeSmith: which list was it?

[00:18:00.0000] <MikeSmith>
/me is looking now

[00:18:01.0000] <hsivonen>
s/computer/system installation/

[00:18:02.0000] <MikeSmith>
hsivonen: bug-standards⊙go

[00:18:03.0000] <MikeSmith>
I'll send a reply to Karl suggesting that

[00:19:00.0000] <hsivonen>
MikeSmith: ok

[00:42:00.0000] <hsivonen>
Hixie: do you intend to write spec language for an iframeless conformance class of consumers?

[00:42:01.0000] <hsivonen>
Hixie: if not, do we really care about conformance requirements for iframe text content?

[00:43:00.0000] <Hixie>
no, and it's possible that we can get away with saying it has to be blank, i don't know how many people want to allow iframe to fallback

[00:44:00.0000] <hsivonen>
considering that Netscape 4 is dead, are there iframeless browsers except Opera when configured to be one?

[00:45:00.0000] <hsivonen>
I wonder how many people use Opera with iframes disabled now that Web apps use iframes for non-advertising purposes

[00:50:00.0000] <Hixie>
it certainly would be convenient for us to be able to say it must be empty. :-)

[00:51:00.0000] <hsivonen>
Hixie: have you checked if it really is empty virtually all the time on real sites?

[00:51:01.0000] <Hixie>
no

[00:58:00.0000] <hsivonen>
HTML5 support in the W3C Validator seems to be of a lot of interest on twitter

[01:16:00.0000] <MikeSmith>
hsivonen: the availability of it at the W3C site definitely seems to have created some buzz

[01:17:00.0000] <MikeSmith>
I don't know if people just didn't know about validator.nu before, or what

[01:36:00.0000] <hsivonen>
MikeSmith: I think support in the W3C Validator is perceived to give more legitimacy and a feeling of almost-readyness to HTML5 itself

[01:40:00.0000] <MikeSmith>
hmm, the "give more legitimacy" part is a clear win for us, but not sure the "feeling of almost-readyness to HTML5 itself" part is. I would hope it doesn't lead people to get disappointed and start commenting that any element that the validator tells them is HTML5-valid should be supported in browsers already.

[02:18:00.0000] <gsnedders>
r2433 ftw.

[02:19:00.0000] <gsnedders>
/me waits for people to complain that there really was a 0th year

[02:23:00.0000] <Hish>
hi. can anybody give me a hint what tag would best describe a tabulator as used in MS word etc.? currently I use a <span class="tab"> for this...

[02:34:00.0000] <Hixie>
Hish: i believe you want a css feature

[02:35:00.0000] <Hish>
Hixie: I thought about this as well but at the end of the day, a tabulator is very close to a table structure. So, I'm really not sure if a tabulator really is only presentation...

[03:02:00.0000] <Hish>
Hixie: one more question: is there a tag which lets me define a data range, i.e. row/col 2/2 to 6/6 of a table so that I can use this range within the output tag?

[03:02:01.0000] <Hixie>
no

[03:03:00.0000] <Hish>
thx

[03:04:00.0000] <hsivonen>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cp%3E%3Cinput%20placeholder%3D'text%0Atext%0Atext'%3E%0A

[03:04:01.0000] <hsivonen>
interesting in WebKit

[03:05:00.0000] <Hixie>
hah

[03:25:00.0000] <Hixie>
so... <input type=file multiple> would result in different actual fields in the field data set per file

[03:26:00.0000] <Hixie>
but with <input type=email multiple>

[03:26:01.0000] <Hixie>
what do we do with input.value?

[03:32:00.0000] <Hixie>
maybe we should introduce input.values

[03:33:00.0000] <Hixie>
i'd make it comma-separated but that would suck because addr-spec can contain commas

[03:39:00.0000] <Hixie>
hmm...

[03:39:01.0000] <Hixie>
maybe make .value contain the value if multiple="" is absent, and be empty otherwise

[03:40:00.0000] <Hixie>
and have .values be null except for this case, and in this case have it return some sort of mutable DOMStringList

[03:40:01.0000] <Hixie>
not that i really want to have to invent a whole new API just for this

[03:41:00.0000] <Hixie>
i could just disallow the quoted-string part of addr too

[03:43:00.0000] <Hixie>
aw man and domain-literal has crazy ass syntax too

[03:43:01.0000] <hsivonen>
Hixie: what's your expected fallback behavior?

[03:43:02.0000] <Hixie>
who the hell invented this

[03:43:03.0000] <Hixie>
hsivonen: using .value would have nice fallback behavior

[03:44:00.0000] <Hixie>
but if i do that i really have to redefine addr-spec to be radically simpler

[03:44:01.0000] <hsivonen>
/me notes that Mail.app allows the user to type a comma after an email address

[03:44:02.0000] <hsivonen>
/me wonders how that works if the addresses can contain commas

[03:45:00.0000] <Philip`>
Gmail splits addresses on commas too

[03:45:01.0000] <Hixie>
they can only contain commas in quoted bits

[03:45:02.0000] <hsivonen>
Hixie: can the addresses be whitespace-separated?

[03:45:03.0000] <hsivonen>
good XML and SGML vocabularies separate tokens with whitespace

[03:46:00.0000] <Hixie>
   "foo,bar"@[\,].com   appears to be a valid e-mail address

[03:46:01.0000] <Hixie>
as does   " "@[\ ]

[03:46:02.0000] <hsivonen>
that's crazy

[03:46:03.0000] <hsivonen>
Does email delivery with addresses like that work?

[03:46:04.0000] <Hixie>
i doubt it

[03:47:00.0000] <Hixie>
i'm just gonna redefine addr-spec to be what users expect

[03:47:01.0000] <Hixie>
dot-atom "@" dot-atom

[03:47:02.0000] <hsivonen>
/me expects more controversy with HTML5 redefining things from the IETF realm

[03:48:00.0000] <Philip`>
Why does anything have to be redefined, rather than just defining new terms?

[03:49:00.0000] <hsivonen>
Philip`: you mean defining a new term like "Web email address"?

[03:49:01.0000] <Hixie>
well i didn't redefine addr-spec

[03:50:00.0000] <Hixie>
i just defined "Valid email address"

[03:50:01.0000] <hsivonen>
Hixie: are you coordinating this with mailto URIs?

[03:50:02.0000] <Hixie>
no.

[03:51:00.0000] <Hixie>
mailto: is defined in rfc2368 and allows the whole gamut of batshit crazy stuff like comments and quoted strings and backslash-escaped characters and everything.

[03:51:01.0000] <Hixie>
we're already (as of wf2) narrower than that

[03:51:02.0000] <Hixie>
i'm just making it even narrower.

[03:52:00.0000] <Philip`>
hsivonen: I mean defining a new term like "sensible-addr-spec", rather than redefining addr-spec, since it sounded like that was possible what people were talking about

[03:52:01.0000] <hsivonen>
Philip`: ok

[03:56:00.0000] <Lachy>
Hixie, how will a <input type=email name=name multiple> element be submitted? Will it be like this: ?name=a⊙x&name=b⊙x

[03:56:01.0000] <hsivonen>
Lachy: that would be backwards-incompatible

[03:57:00.0000] <Lachy>
so it would need to be ?name=<some delimted list of emails>

[04:09:00.0000] <Lachy>
Hixie, AUDITNAV is a bad name to use. If PING isn't acceptable, now about something like NOTIFY?

[04:10:00.0000] <Philip`>
TRACK!

[04:10:01.0000] <hsivonen>
BIKE

[04:10:02.0000] <Philip`>
As a bonus, people couldn't fake it with XHR

[04:11:00.0000] <Hixie>
Lachy: tell julian, not me

[04:11:01.0000] <Hixie>
i'll use whatever he gets approved

[04:13:00.0000] <Philip`>
/me agrees that PING is a bad name, because it's got pretty much nothing to do with what everyone else in the computer networking world means by 'ping' (i.e. a reachability/RTT-measurement tool)

[04:16:00.0000] <Lachy>
/me isn't motivated enough to start another naming debate on the list.

[04:16:01.0000] <Hixie>
i agree that for the http name PING is suboptimal

[04:17:00.0000] <Hixie>
i don't care if it's AUDITNAV or something else

[04:17:01.0000] <Hixie>
NOTIFY seems a bit vague for what we're defining, but it could be defined more vaguely too

[04:17:02.0000] <Lachy>
AUDIT would be better

[04:17:03.0000] <Hixie>
i just don't really care

[04:17:04.0000] <Hixie>
i have enough troubles with html, http can take care of itself :-)

[04:19:00.0000] <hsivonen>
http://en.wikipedia.org/w/index.php?title=Document_Type_Declaration&diff=253488659&oldid=prev

[04:19:01.0000] <hsivonen>
typical wikipedia

[04:22:00.0000] <broquaint>
Anyone know if there will be a JS interface to form data handling in the upcoming specs? Specifically I'm thinking of something like: var fData = someForm.toPostData(); window.location = '/path?' + fData;

[04:23:00.0000] <broquaint>
Because currently there's no native way of doing it and any naive implementation will vary from browser to browser AFAICT.

[04:29:00.0000] <Hixie>
can't you just set the form's action to "/" and post the form normally?

[04:31:00.0000] <broquaint>
Sure, but if you've got existing data (be it form based or otherwise) there's no programmatic way to get a multipart/form-data representation.

[04:32:00.0000] <broquaint>
I'm not sure if this is relevant to #whatwg's goals but I figure I can be pointed in the appropriate direction if not :)

[04:32:01.0000] <Hixie>
why would you want that representation?

[04:33:00.0000] <broquaint>
For AJAX mainly.

[04:33:01.0000] <Hixie>
it's definitely something that would be in our goals if you can convince us there's a good reason for it :-)

[04:33:02.0000] <broquaint>
Ooh, interesting :)

[04:34:00.0000] <Lachy>
it's interesting how wikipedia is incorrectly using the abbreviation "DTD" to mean Document Type Declaration

[04:34:01.0000] <Lachy>
/me will fix it

[04:35:00.0000] <Lachy>
hmm, well, perhaps their not. It's just ambiguous since their using headings like "HTML 4.01 DTDs"

[04:36:00.0000] <Hixie>
nn

[04:36:01.0000] <broquaint>
So, for example, if you wanted to do an XHR with the POST method using multipart/form-data there's no standard way of stitching the data together such that it is indistinguishable from the browser's equivalent.

[04:37:00.0000] <broquaint>
Perhaps I should put a mailing list message together ...

[04:49:00.0000] <hsivonen>
does wikipedia have something like cvs blame?

[04:51:00.0000] <hsivonen>
/me finds this piece of encyclopedic writing: "There is currently no indication as to whether HTML 5 will support RDFa, or be limited just to microformats."

[05:20:00.0000] <Dashiva>
Hey, new last week

[05:21:00.0000] <Philip`>
I think Mr Last Week doesn't quite understand the concept of "week", given the variation in posting frequency

[05:24:00.0000] <Philip`>
contenteditable would be great if it was just slightly less buggy in Firefox

[05:36:00.0000] <hsivonen>
contenteditable is also underdefined

[05:37:00.0000] <hsivonen>
yesterday, I learned that Gecko's editor has a dependency on the HTML containment rules of the HTML parser

[05:38:00.0000] <hsivonen>
whoa. since when has contenteditable accepted the empty string

[05:39:00.0000] <Philip`>
Since at least IE6

[05:39:01.0000] <Philip`>
which is nice because you can do <span contenteditable>...</span>

[05:40:00.0000] <Philip`>
/me is just using contenteditable for plain text, so he doesn't care how it handles HTML markup

[05:40:01.0000] <Philip`>
(though I wouldn't mind if there were some way I could force it to be plain markup and prevent users inserting fancier things)

[05:42:00.0000] <Philip`>
Oh, hmm, paste is really badly broken in my contenteditable spans in FF3 :-(

[05:43:00.0000] <yecril71>
The suggestion to provide an explicit baseline for time stamps is recursive.

[05:44:00.0000] <yecril71>
The baseline would have to be denoted somehow as well,

[05:44:01.0000] <yecril71>
so it would possibly depend on its own baseline,

[05:44:02.0000] <yecril71>
and so on.

[05:58:00.0000] <Philip`>
Gmail wins again! "Would you like to... Add to calendar - scheduled for Weekly on Tuesday, Thu..." - yes, thank you for letting my add the hypothetical garbage collection on the paved form login cowpath into my calendar

[05:59:00.0000] <Philip`>
Someone needs to teach Gmail to understand metaphors

[06:00:00.0000] <Dashiva>
/me blinks

[06:01:00.0000] <Lachy>
Philip`, what?

[06:01:01.0000] <Lachy>
nothing you just said makes sense to me

[06:01:02.0000] <Dashiva>
Is that referring to a message on public-html?

[06:01:03.0000] <Dashiva>
And it picked up some sentence as being a potential calendar item?

[06:02:00.0000] <Philip`>
Lachy: Gmail applies some heuristics to email messages so it can offer useful services, like adding meetings mentioned in the email into your calendar

[06:02:01.0000] <Philip`>
or like tracking DHL packages that were mentioned in the email

[06:03:00.0000] <Philip`>
and in every single case I've seen, it's got it totally wrong and misinterpreted the emails (in this case one of Hixie's recent messages to the WHATWG list)

[06:03:01.0000] <Lachy>
yeah, I understand that. But "paved form login cowpath" doesn't make sense at all

[06:03:02.0000] <Philip`>
See "Solving the login/logout problem in HTML"

[06:04:00.0000] <Philip`>
around the 11th paragraph

[06:04:01.0000] <Philip`>
"The form login cowpath is so commonly frequented that not only has someone already gone and paved it but it has also been tree-lined, has garbage collection scheduled for Tuesdays and Thursdays, and will be electing a representative at the next general election."

[06:05:00.0000] <Dashiva>
What happens if there are two (or more) forms with the name given in the www-authenticate header?

[06:05:01.0000] <Philip`>
Dashiva: Or none?

[06:06:00.0000] <Dashiva>
Philip`: none is simple, means you don't get to login :)

[06:08:00.0000] <Philip`>
Dashiva: "The form parameter, if present, indicates that the first form element in the entity body whose name is the specified string, in tree order, if any, is the login form." sounds pretty clear on that

[06:09:00.0000] <Dashiva>
Huh, missed that

[06:10:00.0000] <Dashiva>
So I guess allowing several different auths on the same login page wouldn't work

[06:12:00.0000] <Philip`>
/me isn't quite sure why you'd want to tell the UA where the login form is

[06:13:00.0000] <Philip`>
because the UA can't be a bot that logs in automatically, since it doesn't know the names of the username and password fields in the identified login form

[06:13:01.0000] <Philip`>
and the UA can't highlight the form to the user because highlights are ugly and disturb the page design, so authors would want control of that themselves instead

[06:15:00.0000] <Dashiva>
Finding the password field isn't that hard :)

[06:16:00.0000] <Philip`>
Oh, that's a good point

[06:21:00.0000] <Lachy>
gsnedders, yt?

[06:34:00.0000] <yecril71>
Is there a chance someone fixes the negative margin on DL in the specification?

[06:34:01.0000] <yecril71>
It makes the element section headers invisible.

[06:35:00.0000] <Philip`>
/me sees that http://planet.intertwingly.net/ now includes Mr Last Week

[06:43:00.0000] <hsivonen>
now I can get a dose of workplace stalking from my favorite mobile destination

[06:46:00.0000] <jgraham>
/me notes that the time of th big bang is constrained to much better than 1 billion years, assuming scientists aren't hopelessly wrong about everthing

[06:47:00.0000] <yecril71>
Why are SCRIPT elements not allowed within TABLE elements?

[06:47:01.0000] <yecril71>
It seems if the table rows are written by JavaScript, so must be the TABLE tags.

[06:47:02.0000] <Philip`>
jgraham: It only requires scientists to be hopelessly wrong about one thing, not about everything

[06:47:03.0000] <Philip`>
jgraham: and I trust scientists to be hopelessly wrong about several things

[06:48:00.0000] <Philip`>
(though I don't know which several things those are, and I trust the scientists to sort it out in the end)

[06:48:01.0000] <jgraham>
Philip`: Well it would require them to be wrong in specific ways about specific things. 

[06:50:00.0000] <Philip`>
I suppose the scientists could always cheat by just redefining the concept of time so that they are correct

[06:50:01.0000] <jgraham>
(fwiw the uoted uncertianty is about 1% i.e. about 100 million years rather than 10% i.e. 1 billion years)

[06:50:02.0000] <jgraham>
s/uo/quo/

[06:52:00.0000] <jgraham>
(However the use case of <time>5 seconds after the big bang</time> is clearly silly

[06:52:01.0000] <jgraham>
)

[07:28:00.0000] <Philip`>
Hmm, it seems Julian took seriously my idea to use the TRACK method

[07:28:01.0000] <Philip`>
Maybe it is actually a sensible idea, though that wasn't my intention at all

[07:30:00.0000] <yecril71>
Why isn�t form structure fully reflected in the DOM?

[07:30:01.0000] <yecril71>
A FORM knows its elements, a FIELDSET does not.

[07:30:02.0000] <yecril71>
An INPUT control does not know its LABEL.

[07:31:00.0000] <yecril71>
Or a FIELDSET it belongs to.

[07:31:01.0000] <Philip`>
/me isn't sure what yecril71 means

[07:31:02.0000] <yecril71>
It would be nice to have these properties (read-only).

[07:31:03.0000] <Philip`>
Oh, like having an HTMLInputElement.label property or something?

[07:32:00.0000] <yecril71>
/me means introducing DOM properties

[07:32:01.0000] <yecril71>
Right.

[07:32:02.0000] <Dashiva>
A control can have many labels

[07:32:03.0000] <yecril71>
.label can return a collection

[07:34:00.0000] <yecril71>
or just the first label, in document order

[07:34:01.0000] <Philip`>
Obvious question: what are the use cases that would be solved by adding these properties?

[07:34:02.0000] <Dashiva>
And you might want to check the spec

[07:34:03.0000] <Dashiva>
Some of them are already in

[07:34:04.0000] <yecril71>
/me thinks having multiple lables is bad design

[07:35:00.0000] <Philip`>
http://www.whatwg.org/specs/web-apps/current-work/multipage/forms.html#dom-lfe-labels

[07:38:00.0000] <yecril71>
Thanks.  I see a fieldset has elements and an input has labels.

[07:39:00.0000] <yecril71>
How about adding INPUT.fieldset?

[07:39:01.0000] <Philip`>
What for?

[07:40:00.0000] <yecril71>
To differentiate response to an event by the fieldset the control is in.

[07:41:00.0000] <yecril71>
Since we have the other two, .fieldset would just complete the picture.

[07:41:01.0000] <Dashiva>
Completing the picture isn't a use case

[07:42:00.0000] <yecril71>
If course, one can go up the DOM to find the nearest fieldset.

[07:42:01.0000] <Philip`>
'Completeness' sounds like it's arguing for theoretical purity, which usually isn't considered a worthwhile goal

[07:42:02.0000] <yecril71>
Completing the picture was an additional comment.

[07:43:00.0000] <Philip`>
Is differentiating responses to events based on the fieldset a problem that occurs in practice, enough to justify the cost of adding the feature to HTML5?

[07:44:00.0000] <Philip`>
s/enough/often enough/

[07:45:00.0000] <yecril71>
I do not know; I have just happened to find it useful so I decided to ask.

[07:46:00.0000] <yecril71>
I think .fieldset was left out because there exists an easy workaround,

[07:46:01.0000] <yecril71>
unlike the other cases.

[07:46:02.0000] <Philip`>
It seems like something that could be useful and isn't supported yet, but it's not clear whether it's worth supporting in the language rather than telling people to navigate the DOM themselves

[07:47:00.0000] <yecril71>
It is not clear to me either.

[07:49:00.0000] <Philip`>
Just use jQuery and then all this is trivial :-)

[07:49:01.0000] <Philip`>
$(input).parents('fieldset:first') or whatever

[07:52:00.0000] <yecril71>
Since OBJECT is submittable now, it needs to be associated with a FORM.

[07:53:00.0000] <yecril71>
The association is not historical any more.

[07:57:00.0000] <Dashiva>
Oh snap, I got told

[07:58:00.0000] <Dashiva>
But since Lachy works for Opera, the point is somewhat lessened :)

[07:59:00.0000] <yecril71>
And what is the use case for FIELDSET.name?

[07:59:01.0000] <Lachy>
Dashiva, what?

[07:59:02.0000] <Dashiva>
Last last week

[08:02:00.0000] <gsnedders>
"philip the WHAT WG teamster rebel"

[08:02:01.0000] <gsnedders>
Philip`'s stealing my title!

[08:03:00.0000] <Dashiva>
yecril71: Fieldsets appear in the form elements collection

[08:03:01.0000] <Lachy>
oh, wow. I love how Mr Last Week is defending me from MikeSmith in such a sarcastic way

[08:04:00.0000] <yecril71>
Right, thanks.

[08:04:01.0000] <gsnedders>
haha.

[08:04:02.0000] <gsnedders>
That is nice.

[08:05:00.0000] <Philip`>
gsnedders: Since when were you a rebel?

[08:05:01.0000] <gsnedders>
Philip`: I'm constantly going against Hixie's opinions, and getting killed for it

[08:05:02.0000] <jgraham>
/me wonders if gsnedders is confusing real life with a FPS

[08:06:00.0000] <jgraham>
Becuse in real life it is hard to constantly get killed

[08:06:01.0000] <gsnedders>
jgraham: FPS!? Third-person shooters is where it's at1

[08:06:02.0000] <gsnedders>
*!

[08:06:03.0000] <Philip`>
Why does nobody write second-person shooters?

[08:06:04.0000] <jgraham>
gsnedders: Just put it this way, if you respawn in RL people ill get freaked out

[08:07:00.0000] <gsnedders>
I think those would be harder to control than third-person shooters, Philip` 

[08:07:01.0000] <jgraham>
Or start some sort of major world religion I guess

[08:07:02.0000] <gsnedders>
jgraham: I do freak people out, though

[08:07:03.0000] <Dashiva>
Philip`: I've considered it

[08:07:04.0000] <Dashiva>
There is a second-person horror game on the playstation, I believe

[08:07:05.0000] <Lachy>
Philip`, probably because no-one has figured out how to market a game where the aim is to get yourself shot

[08:08:00.0000] <jgraham>
Lachy: That would be awesome

[08:08:01.0000] <Lachy>
yeah, but it would be too easy to just stand out in the open

[08:08:02.0000] <jgraham>
Maybe you could only control people indirectly like you had to annoy the neighbourghs enough that they eventually went postal

[08:09:00.0000] <Lachy>
maybe it would work if the other characters in the game were on your side and intentionally trying not to hit you, so you had to position yourself in such a way that you would get hit by friendly fire

[08:11:00.0000] <yecril71>
And why are fieldsets listed?

[08:11:01.0000] <yecril71>
They are not listed in HTML4,

[08:11:02.0000] <yecril71>
and they are not form controls,

[08:12:00.0000] <yecril71>
so the name HTMLFormControlsCollection is not really appropriate.

[08:13:00.0000] <yecril71>
I understand it is practical to ask for all fieldsets,

[08:13:01.0000] <yecril71>
but it is already covered by getElementsByTagName.

[08:43:00.0000] <Philip`>
If I'm using lxml, how do I iterate over all h1, h2, and h3 elements?

[08:43:01.0000] <Philip`>
node.iter('h1') works, but node.iter('h1|h2') doesn't and I can't find any syntax that does

[08:43:02.0000] <jgraham>
you could use xpath

[08:44:00.0000] <jgraham>
(alhough it may not be the most efficient way and returns a lsit not an iterator)

[08:44:01.0000] <BenMillard>
"I very much hope we can (through W3C and MWI) drive user awareness of the mobile Web as a platform and start to create a mindset among developers and users that it's "OK" to have a different user experience between different devices / platforms - that this can be consistent with there being One Web." -- http://www.w3.org/QA/2008/11/_as_part_of_a.html

[08:44:02.0000] <BenMillard>
doesn't sound like One Web to me...

[08:45:00.0000] <Dashiva>
Different user experience as in different styling and device-specific widgets, or as in different pages entirely?

[08:46:00.0000] <BenMillard>
"'Full Web' to me is a code word for 'web pages designed exclusively for PCs.'" ouch, guess he never used a good mobile browser...

[08:47:00.0000] <jgraham>
Philip`: Or I guess just do [x for x in node.iter() if x.tag in ('h1','h2', 'h3')]

[08:47:01.0000] <jgraham>
(but that is kind of obvious)

[08:49:00.0000] <BenMillard>
this makes sense: "The idea is to allow the proxies to do their work on pages that have not taken mobile users into account, while stepping out of the way when a content provider has developer a service that adapts for mobile anyway."

[08:49:01.0000] <BenMillard>
such as webpages using HTML 4.01 Strict, with a handheld stylesheet, I assume (lol)

[08:49:02.0000] <BenMillard>
"I will say that the growth of flat-rate data plans is a positive trend in the industry right now and I hope that trend continues." me too!

[08:51:00.0000] <BenMillard>
"I sincerely hope the HTML Working Group considers our work in CDF before they re-invent the wheel." I hadn't heard of CDF before now...

[08:54:00.0000] <smedero>
BenMillard: CDF has been brought up many times in the SVG in HTML discussions, search public-html. 

[08:56:00.0000] <smedero>
(though, not in a particularly compelling way...)  

[08:57:00.0000] <BenMillard>
smedero, CDF comes from here? http://www.w3.org/2004/CDF/

[08:57:01.0000] <smedero>
yep

[08:58:00.0000] <gsnedders>
What's a good human readable and computer parsable biblio format?

[08:59:00.0000] <jgraham>
gsnedders: Bibliography?

[08:59:01.0000] <gsnedders>
jgraham: yeah

[08:59:02.0000] <jgraham>
bibtex is pretty standard for LaTeX applications

[09:00:00.0000] <jgraham>
Dunno if it meets your needs

[09:00:01.0000] <gsnedders>
I just want something for Anolis that isn't Refer

[09:00:02.0000] <gsnedders>
And it needs to be easy to parse in Python :P

[09:00:03.0000] <BenMillard>
smedero, this seems quite a stretch from the Mobile Web's Default Delivery Context: http://www.w3.org/TR/2007/CR-WICD-20070718/#child-objects

[09:01:00.0000] <BenMillard>
/me concludes CDF is FAIL.

[09:02:00.0000] <jgraham>
http://pybliographer.org/

[09:02:01.0000] <gsnedders>
jgraham: Can you import a GPL module into a MIT licensed project?

[09:03:00.0000] <jgraham>
Incidentially I _really_ have no idea if bibtex is sutiable for what you want

[09:03:01.0000] <gsnedders>
BibTeX is suitable

[09:03:02.0000] <jgraham>
gsnedders: IANAL

[09:03:03.0000] <gsnedders>
My IANAL opinion would be it is questionable

[09:06:00.0000] <jgraham>
http://groups.google.com/group/comp.lang.python/browse_thread/thread/cd8646451aacfcc9

[09:06:01.0000] <Lachy>
gsnedders, you can, but it means that any distribution that includes the GPL'd code is effectively licenced under the GPL

[09:08:00.0000] <gsnedders>
There appears to be other bibtex scripts in Python

[09:08:01.0000] <gsnedders>
like, MIT licensed ones

[09:11:00.0000] <jgraham>
gsnedders: That sounds altogether like a better idea

[09:13:00.0000] <jgraham>
gsnedders: Where are the MIT licensed ones?

[09:13:01.0000] <rubys>
Lachy's correct

[09:14:00.0000] <gsnedders>
jgraham: http://code.google.com/p/bibstuff/ looks best

[09:14:01.0000] <rubys>
http://www.gnu.org/licenses/gpl-faq.html#WhatDoesCompatMean

[09:17:00.0000] <gsnedders>
woah.

[09:18:00.0000] <gsnedders>
The W3C doesn't have any exportable format of TRs, as far as I can see!

[09:18:01.0000] <jgraham>
exportable?

[09:18:02.0000] <gsnedders>
like, more than just http://www.w3.org/TR/

[09:18:03.0000] <gsnedders>
like, that I can use to build a bibtex file of all W3C TRs

[09:21:00.0000] <jgraham>
http://www.w3.org/2002/01/tr-automation/tr.rdf

[09:21:01.0000] <Philip`>
gsnedders: http://www.w3.org/2002/01/tr-automation/tr.rdf ?

[09:21:02.0000] <Philip`>
Oops, I lose

[09:21:03.0000] <gsnedders>
Where did you find any link to that!?

[09:22:00.0000] <Philip`>
http://google.com

[09:22:01.0000] <Philip`>
and http://www.w3.org/2002/01/tr-automation/

[09:22:02.0000] <Philip`>
(I searched for "w3c tr list")

[09:22:03.0000] <jgraham>
In the <head> (I assumed that the W3C couldn't have a list without having accompnying RDF)

[09:23:00.0000] <gsnedders>
Philip`: I failed to google for the right thing

[09:23:01.0000] <gsnedders>
jgraham: So did I. Which was why I was amazed to not find it in two seconds.

[09:25:00.0000] <jgraham>
"Benefits of using Semantic Web technologies

[09:25:01.0000] <jgraham>
@@@"

[09:26:00.0000] <rubys>
gsnedders: view source on http://www.w3.org/TR/

[09:26:01.0000] <rubys>
There's a meta link

[09:28:00.0000] <rubys>
oh, that's probably what jgraham was referring to.  nm.

[09:28:01.0000] <Philip`>
/me wonders how to efficiently implement keyframe-like animation in SVG, where each frame is 100KB of independently-generated SVG code and there's typically (but not always) a lot of similarity between adjacent frames

[09:30:00.0000] <Philip`>
(In particular, animating http://philip.html5.org/misc/spec-links.html over the lifetime of the spec)

[09:44:00.0000] <gsnedders>
problem: I can't include a converted copy of <http://www.w3.org/Style/Group/css3-src/biblio.ref> in Anolis, yet I need the same IDs to remain compat.

[09:46:00.0000] <Philip`>
Why can't you include that?

[09:46:01.0000] <gsnedders>
licensing, etc.

[09:46:02.0000] <Philip`>
Why do you need the same IDs?

[09:46:03.0000] <gsnedders>
compat. with existing documents

[09:47:00.0000] <Philip`>
Are there practical compatibility considerations?

[09:47:01.0000] <gsnedders>
seeming people like Lachy want to use it for docs they already edit, yes

[09:47:02.0000] <Philip`>
(i.e. do documents unavoidably rely on those specific IDs, or something?)

[09:47:03.0000] <gsnedders>
The IDs isn't the problem.

[09:48:00.0000] <gsnedders>
It's including everything in that file when it is rather long that is.

[09:48:01.0000] <gsnedders>
It contains really random things.

[09:48:02.0000] <gsnedders>
not that you can see that :)

[09:56:00.0000] <BenMillard>
gsnedders, maybe you could contact relevant staff at W3C about what you want?

[09:56:01.0000] <gsnedders>
BenMillard: I don't really want to use Refer anyway, so it isn't particularly helpful

[10:14:00.0000] <gsnedders>
BenMillard: Also, I shouldn't have this document at all, so contacting the relevant staff may not be the best move :)

[10:17:00.0000] <Philip`>
But mentioning in public logged IRC that you have illicitly gained access to it is a good move? :-)

[10:17:01.0000] <gsnedders>
:P

[10:18:00.0000] <gsnedders>
Philip`: I've implied often enough that I have got at the css3 postprocessor stuff :P

[11:20:00.0000] <Lachy>
gsnedders, it's not really a problem that you have a copy of that stuff. Bert really wouldn't mind.

[11:20:01.0000] <gsnedders>
Lachy: He knows I do.

[11:21:00.0000] <Lachy>
ok

[12:11:00.0000] <Lachy>
it would be nice if some people read the spec before making absurd, non-sensical comments about it: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2008-November/017432.html Not after: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2008-November/017434.html

[12:12:00.0000] <Lachy>
(FYI, in that first mail, he totally screwed up the quoting, but I think his comments start after the " ~TJ" signature. Everything before that is quoted.)

[12:18:00.0000] <jgraham>
Lachy: The first email suggests that he did read the spec. He may even have a legitimate piece of feedback

[12:23:00.0000] <Hixie>
the idea of pointing out the login form is that a bot that knows the name-value mapping of the fields in the form can then fill it in and submit it without needing to know the current url of the login form itself

[12:23:01.0000] <Hixie>
it may not be particularly useful

[12:29:00.0000] <Lachy>
Hixie, if the form only contains 1 text input and 1 password input, then a bot could take a fairly reasonable guess as to what each one is for

[12:30:00.0000] <Hixie>
yup

[12:31:00.0000] <Lachy>
but still, the value of the feature is still questionable since the bot could just as easily be manually told which form to submit when given the credentials

[12:33:00.0000] <Lachy>
jgraham, I couldn't find anything Pentasis wrote in the first email that made any sense, let alone suggests he read the spec

[12:33:01.0000] <jgraham>
"But, I will concede to the fact that this may make things more difficult. But in that case I would ask for the definition of the spec to be changed from "The time element represents a date and/or a time." into something more restricting and exact which at least represents the limitations of this element that are presented in this discussion."

[12:35:00.0000] <jgraham>
So his acionable feedback is "change the definition of the <time> element to something like 'The time element represents a date and/or time that can be expresed on a gregorian calendar'

[12:37:00.0000] <jgraham>
except that is not a sufficient constraint

[12:38:00.0000] <jgraham>
anyway I don't know if it is worth making the change but it is not obviously crazy

[12:43:00.0000] <Lachy>
ok, fair enough. I guess I missed that bit after being stunned by his suggestions for setting the "base time" (whatever that is) to the big bang

[12:44:00.0000] <Hixie>
that does seem a little premature

[12:45:00.0000] <jgraham>
Oh that bit was crazy. AFAICT he was suggecting something like <time base="Death of Joan of Arc">5 seconds</time> to represent 5 seconds after the moment of death of Joan of Arc 

[12:46:00.0000] <jgraham>
/me wonders if users actually do have problems with <input placeholder> type designs

[12:47:00.0000] <jgraham>
/me has always found that they are used in situtations where the purpose of the conrol can be stored in short term meory for long enough to use the in put correctly

[12:54:00.0000] <BenMillard>
jgraham, I've seen placeholder text used instead of labelling text

[12:54:01.0000] <BenMillard>
when you enter several values in a form which does that throughout, it makes for a fun memory exercise :)

[12:55:00.0000] <jgraham>
Ah, OK I can see that it can be abused, but I have only encountered it in situations where it was easy to remember the point of the field

[12:56:00.0000] <Hixie>
yeah it's easily abusable but definitely has cases where it is useful

[12:56:01.0000] <Hixie>
it's mostly for discovery

[12:57:00.0000] <gsnedders>
See <http://habariproject.org/en/user/login>, and the markup of that

[12:57:01.0000] <BenMillard>
jgraham, I've been noodling with the Table Inspector interface and came up with this: http://projectcerbera.com/!dev/table-inspector/

[12:57:02.0000] <BenMillard>
(I broke parts of the JS by changing the types of elements)

[12:59:00.0000] <gsnedders>
BenMillard: You came up with the default Apache file listing? Wow.

[12:59:01.0000] <BenMillard>
gsnedders, new.html is the actual interface

[13:00:00.0000] <gsnedders>
BenMillard: :)

[13:00:01.0000] <jgraham>
BenMillard: It's nicer than anyhing I managed

[13:00:02.0000] <BenMillard>
jgraham, take as much/little as you want from it if you decide to update the real inspector

[13:00:03.0000] <jgraham>
BenMillard: I will 

[13:01:00.0000] <BenMillard>
jgraham, my /!dev folder is blocked in robots.txt, so it shouldn't start outranking you in search engines :)

[13:01:01.0000] <BenMillard>
reducing the amount of vertical height was my main goal in the new interface

[13:03:00.0000] <BenMillard>
oh, and I moved the CSS and JS to external resources so they can be cached for frequent users

[13:05:00.0000] <Lachy>
Hixie, in the definition of Interactive Content, it says "Certain elements in HTML have an activation behavior, which means the user agent should allow the user to manually trigger them in some way, for instance using keyboard or voice input (though not mouse clicks, which are handled above)"

[13:06:00.0000] <Lachy>
can you make "though not mouse clicks, which are handled above" a link to wherever that's referring to?

[13:06:01.0000] <Hixie>
can you send mail? i'm about to be afk

[13:06:02.0000] <Lachy>
ok

[13:07:00.0000] <gsnedders>
Hixie: Lies! You're typing on your keyboard!

[13:08:00.0000] <Lachy>
gsnedders, he said "about to be", meaning soon, not presently

[13:11:00.0000] <gsnedders>
Oh, I can't read.

[14:14:00.0000] <Hixie>
/me ponders <input type=submit novalidate>

[14:36:00.0000] <Hixie>
i don't know whether to do <input type=submit novalidate> yet or whether to tell people to use form.submit() for now and add it later

[14:36:01.0000] <Hixie>
any opinions either way?

[14:37:00.0000] <Dashiva>
So there are people asking for it right now?

[14:40:00.0000] <jgraham>
What is the use case?

[14:41:00.0000] <Dashiva>
Saving a partially completed form was one

[14:43:00.0000] <jgraham>
Can you do that entirely without scipt assuming a novlidate attribute?

[14:44:00.0000] <Dashiva>
No script on the client side, at least

[14:45:00.0000] <jgraham>
So you have two buttons, one marked save and one marked submit? 

[14:47:00.0000] <jgraham>
I guess it is a rther low cost feature since it is not doing something rather than doing something extra

[14:51:00.0000] <Hixie>
true

[14:51:01.0000] <Hixie>
the use case is saving the form, yeah

[14:53:00.0000] <Hixie>
also a "preview" button, and buttons that do server-side filling of fields without submitting the whole thing

[14:53:01.0000] <Hixie>
i guess i'll add novalidate

[14:53:02.0000] <Hixie>
you have about an hour to come up with a better name than "novalidate". :-)

[14:53:03.0000] <Hixie>
bbl

[15:50:00.0000] <BenMillard>
Hixie, calling it skipvalidity (which could be authored as skipValidity in the document) seems more consistent with the use of "validity" and "fooValidity()" and "ValidityFoo" for things in The Constraint Validation API: http://www.whatwg.org/specs/web-apps/current-work/multipage/forms.html#the-constraint-validation-api

[15:53:00.0000] <BenMillard>
ignorevalidity (ignoreValidity) fits the above and fits the naming of "ignoreCase" and the convention for saying "UAs must ignore foo"


2008-11-26
[16:10:00.0000] <Hixie>
i think i prefer novalidate, for consistency with nohref, noresize, noshade, and nowrap (though the irony that all four of those are obsolete in html5 is not lost on me)

[16:15:00.0000] <BenMillard>
so many conventions to choose from :P

[16:15:01.0000] <BenMillard>
I'm unlikely to use the feature so I don't mind what you call it

[16:24:00.0000] <Hixie>
BenMillard: :-)

[17:35:00.0000] <Hixie>
ew

[17:35:01.0000] <Hixie>
MouseEvent makes a mess of the init* methods

[17:58:00.0000] <Hixie>
hmm

[17:58:01.0000] <Hixie>
so <input type=color>

[17:58:02.0000] <Hixie>
should it be possible for the user to set it to no color?

[18:01:00.0000] <BenMillard>
transparent?

[18:10:00.0000] <Hixie>
no, nothing at all

[18:11:00.0000] <Hixie>
like, type=number allows any number, as well as ""

[18:11:01.0000] <Dashiva>
Can it have no color as initial value?

[18:11:02.0000] <Hixie>
but type=range doesn't allow ""

[18:11:03.0000] <Hixie>
Dashiva: that's another way of phrasing the same question, effectively

[18:12:00.0000] <Lachy>
what are the use cases for colour selection that we're trying to address and which ones are we not?

[18:13:00.0000] <Hixie>
in: selecting the colour of a label in gmail, selecting a color in a paint program

[18:13:01.0000] <Hixie>
can't think of any that are "out" offhand, but i'm sure there are many

[18:14:00.0000] <Hixie>
http://images.google.com/images?client=safari&rls=en-us&q=color%20pickers&ie=UTF-8&oe=UTF-8&um=1&sa=N&tab=wi suggests that color pickers don't have a "no color" mode

[18:14:01.0000] <Lachy>
ok, so this is basically a simple RGB colour palette

[18:15:00.0000] <Dashiva>
Hixie: But many of them have a cancel button

[18:16:00.0000] <Hixie>
Dashiva: yeah but that doesn't unset the previously selected color

[18:16:01.0000] <Lachy>
I'd assume things like paint colour selectors on interior decoration sites wouldn't be adequately covered by this, so that'd be out

[18:17:00.0000] <Hixie>
right

[18:18:00.0000] <Hixie>
out: pantone color selector

[18:19:00.0000] <Dashiva>
I suppose we aren't addressing "How do you ensure the user selects a color and doesn't just go ahead with the default without considering it" either

[18:22:00.0000] <Hixie>
not really

[18:29:00.0000] <Lachy>
yes, I think no colour should be possible, because if someone wants to make a paint application, that's one way of letting the user select transparent

[18:32:00.0000] <BenMillard>
Hixie, I think Word has a no colour mode.

[18:33:00.0000] <Hixie>
Lachy: opacity should be separate from this anyway

[18:41:00.0000] <Lachy>
Hixie, the colour palettes in apps like Adobe Fireworks allow the user to select a no-colour option.  it's useful when, e.g., you want to set the border colour of a shape to one colour and leave the fill colour transparent

[18:47:00.0000] <Hixie>
Lachy: makes sense

[18:51:00.0000] <Hixie>
hmm

[18:51:01.0000] <Hixie>
should we make the form <input type=color> parse colors in a simple way, or using the wacky <font color> algorithm?

[18:52:00.0000] <Hixie>
or using the css algorithm...

[18:52:01.0000] <Hixie>
so many options...

[18:52:02.0000] <Hixie>
can't use css, as it might return an alpha!=1.0 color

[18:58:00.0000] <Lachy>
could you use a subset of the CSS colour, including #FFF, #FFFFFF, rgb(...) and hsl(...), but excluding rgba() and hsla()?

[18:58:01.0000] <Lachy>
also, the colour keywords

[19:01:00.0000] <Hixie>
that seems unnecessarily complex given that we'd still want to serialise everything to #rrggbb for submission

[19:02:00.0000] <heycam>
that's pretty much the same as strokeStyle/fillStyle on the canvas context object though

[19:02:01.0000] <Lachy>
specifically, what conditions are you trying to address? Is this for when the user types in a colour manually or when the value is set by a script?

[19:03:00.0000] <Hixie>
strokeStyle and fillStyle are full CSS colors

[19:03:01.0000] <Hixie>
Lachy: value of the value="" attribute

[19:03:02.0000] <Hixie>
(and form submission)

[19:04:00.0000] <Lachy>
oh, ok.

[19:14:00.0000] <Lachy>
with other input types, I don't think there is any precedent for the value of the value attribute being automatically normalised prior to submission, and so it would probably be best to require the format #rrggbb

[19:15:00.0000] <Lachy>
plus, if we allowed other types, then that would seem to create complications when the input.value property is set by scripts

[19:16:00.0000] <Lachy>
s/other types/other formats/

[19:17:00.0000] <Hixie>
i'm still normalising it before submission btw (from uppercase to lowercase)

[19:17:01.0000] <Lachy>
ok

[20:01:00.0000] <Lachy>
Hixie, is the order of the input types in the table in any particular order? It seems rather random. Could you put them into alphabetical order?

[20:02:00.0000] <Lachy>
well, I guess there sort of grouped by category

[20:07:00.0000] <Hixie>
the order is the order used so that there are the fewest differences from type to type in terms of what cells say "yes"

[20:07:01.0000] <Hixie>
except that password isn't before text

[20:07:02.0000] <Lachy>
Hixie, a valid simple colour should be A to F, not A to Z

[20:08:00.0000] <Hixie>
oops

[20:10:00.0000] <Hixie>
ok fixed

[21:42:00.0000] <sayrer>
Hixie, so, I thought there was a feature freeze? but now we have this new 401 form...

[21:51:00.0000] <Hixie>
sayrer: i said i wasn't adding anything new that hadn't already been requested as of the feature freeze (last december)

[21:51:01.0000] <Hixie>
the recent additions are from requests from 2006/2007

[21:52:00.0000] <sayrer>
that doesn't seem like a useful freeze to me 

[21:52:01.0000] <sayrer>
thanks

[21:52:02.0000] <Hixie>
(or, in the case of workers, from requests from browser vendors who said that without a spec they'd just make up stuff)

[21:52:03.0000] <Hixie>
well the freeze is only intended to land us on schedule

[21:52:04.0000] <sayrer>
well, you are a browser vendor just making stuff up :)

[21:52:05.0000] <Hixie>
i mean implementors

[21:53:00.0000] <sayrer>
/me shrugs

[22:44:00.0000] <hsivonen>
Hixie: seems more like a suggestion freeze than a feature freeze :-)

[22:48:00.0000] <Hixie>
yeah, that'd be a better term

[22:48:01.0000] <Hixie>
i don't recall exactly how i phrased it

[22:56:00.0000] <hsivonen>
http://intertwingly.net/blog/2008/11/20/Half-Full#c1227667144

[22:57:00.0000] <hsivonen>
Is there a definition for hixie:LC, hixie:CR, w3c:LC and w3c:CR?

[22:58:00.0000] <hsivonen>
(and are there URIs to bind hixie and w3c to?)

[23:14:00.0000] <hsivonen>
Hixie: btw, what happened to the OpenID integration idea that sicking mentioned at TPAC?

[23:16:00.0000] <Hixie>
i wonder how hixie:LC, hixie:CR, w3c:LC and w3c:CR differ

[23:16:01.0000] <Hixie>
hsivonen: no idea

[23:17:00.0000] <Hixie>
what's zcorpan's e-mail address?

[23:17:01.0000] <Hixie>
specifically, his webkit bugzilla account address

[23:18:00.0000] <hsivonen>
I'd try searching webkit bugzilla for zcorpan and simonp

[23:19:00.0000] <Hixie>
tried that

[23:43:00.0000] <hsivonen>
Hixie: was "no idea" to the OpenID idea? that is, did you examine the feasibility of moving bits of the OpenID experience to browser chrome in a backwards-compatible way?

[23:54:00.0000] <hsivonen>
http://lists.w3.org/Archives/Public/www-validator/2008Nov/0044.html

[23:55:00.0000] <Hixie>
no idea was to openid. not really sure what to do about it.

[00:25:00.0000] <Hixie>
when you substitute a for b in c, which one is left in c? a, or b?

[00:26:00.0000] <takkaria>
a?

[00:45:00.0000] <gavin>
a

[00:45:01.0000] <gavin>
a is a substitue for b

[00:59:00.0000] <Hixie>
wikitionary agrees

[01:04:00.0000] <hsivonen>
why are SVG and Canvas "extensions"? http://www.extremetech.com/article2/0,2845,2335251,00.asp

[01:08:00.0000] <mookid>
Why am I reading a thread about 'the login/logout problem' ?

[01:09:00.0000] <mookid>
there isn't a problem.. it works fine

[01:09:01.0000] <virtuelv>
hsivonen: that article is, in general, rather uninformed

[01:11:00.0000] <Hixie>
ie8 is adding svg and canvas?

[01:11:01.0000] <Hixie>
that's news to me

[01:17:00.0000] <Hixie>
wtf, screen sharing just doesn't work anymore to this computer

[01:17:01.0000] <Hixie>
i don't get it

[01:18:00.0000] <Hixie>
no error message, nothing

[01:18:01.0000] <Hixie>
afp, too

[01:18:02.0000] <Hixie>
just doesn't connect

[01:29:00.0000] <Lachy>
have you tried restarting the machine?

[01:29:01.0000] <Lachy>
I mean the one you're trying to connect to

[01:30:00.0000] <Hixie>
yes

[01:33:00.0000] <virtuelv>
hsivonen: the telltale sign in that article would be

[01:33:01.0000] <virtuelv>
«What's notable here is the margin. Chrome's winning margin is huge, even though Firefox 3.04, Opera and Safari have incorporated V8»

[01:33:02.0000] <Hixie>
o_O

[01:34:00.0000] <hsivonen>
virtuelv: whoa

[01:34:01.0000] <hsivonen>
/me didn't actually read the whole article

[01:35:00.0000] <virtuelv>
«We tested the version of Firefox (called Minefield) that does include the V8 code and listed those results below our "official" findings.»

[01:38:00.0000] <Hixie>
if I add remainingSpacePercentage, should I add it to sessionStorage, localStorage, and Database, or should I add it to Navigator and assume shared storage?

[01:39:00.0000] <virtuelv>
Hixie: is the percentage really relevant?

[01:40:00.0000] <Hixie>
microsoft want a feature to say how much space is remaining, and bytes don't work

[01:40:01.0000] <Philip`>
Does it make sense on sessionStorage? I'd assume that'd be stored in RAM, and browsers don't have fixed limits on how much RAM a page can use

[01:40:02.0000] <virtuelv>
a percentage is equally useless

[01:41:00.0000] <virtuelv>
if I'm trying to store a DOMString of length 1231, a percentage isn't going to help me

[01:41:01.0000] <Philip`>
virtuelv: Bytes wouldn't help you either, since you don't know how many bytes it'll take to store that string

[01:41:02.0000] <Hixie>
the only thing that it would allow is showing a UI saying how close you are to running out

[01:41:03.0000] <Hixie>
but i guess the UA could do that better anyway

[01:42:00.0000] <Philip`>
virtuelv: so you should just try to store it, and watch for exceptions

[01:53:00.0000] <yecril71>
Allowing to execute hidden commands from the keyboard does not seem to be a good idea at all.

[01:53:01.0000] <yecril71>
Although it would make implementing Vi in HTML pretty hard,

[01:53:02.0000] <yecril71>
I do not think HTML should explicitly provide for that.

[01:54:00.0000] <Hixie>
zcorpan needs to be online more. someone hook him up with screen(1) and irssi(1), please. :-P

[01:55:00.0000] <hsivonen>
the HTML article in wikipedia gets vandalized all the time and isn't protected. the XHTML article is semi-protected, though.

[01:56:00.0000] <yecril71>
Should I bring up the issue with fieldsets and HTMLControlsCollection to the list?

[02:01:00.0000] <Hixie>
what's the issue?

[02:11:00.0000] <yecril71>
The issue is a fieldset is not a control.

[02:11:01.0000] <yecril71>
And it does not belong to form.elements, as of HTML4.

[02:12:00.0000] <yecril71>
So there is a major incompatibility and a semantical flaw.

[02:12:01.0000] <Hixie>
browsers put fieldsets in form.elements, so there's not much we can do about that

[02:12:02.0000] <Hixie>
html5 defines it in a way that solves the "semantical flaw"

[02:12:03.0000] <Hixie>
i.e. it doesn't have a contradiction as best i can tell

[02:12:04.0000] <yecril71>
According to MSDN, a fieldset does not have a name.

[02:12:05.0000] <Hixie>
msdn is rarely accurate

[02:13:00.0000] <Hixie>
i wouldn't pay much attention to it

[02:14:00.0000] <yecril71>
Thanks, I shall evaluate it and leave a note there if it works anyway.

[02:14:01.0000] <mookid>
Hixie: what is all this stuff abuot forms?

[02:14:02.0000] <Hixie>
mookid: ?

[02:14:03.0000] <Hixie>
which stuff?

[02:15:00.0000] <yecril71>
Now that OBJECT is submittable, the note about legacy reasons should go.

[02:15:01.0000] <yecril71>
Because it is a full member of the FORM.

[02:16:00.0000] <Hixie>
could you provide more context? i'm not psychic, i've no idea what note you are talking about

[02:16:01.0000] <yecril71>
"OBJECT belongs to FORM.elements for legacy reasons".

[02:17:00.0000] <yecril71>
(quoting from memory)

[02:17:01.0000] <Hixie>
could you quote from the spec?

[02:17:02.0000] <Hixie>
i can't find that string anywhere

[02:17:03.0000] <yecril71>
Have to find it again first.

[02:19:00.0000] <yecril71>
For historical reasons, the object element, which is not otherwise considered to be related to forms, is also a form-associated element.

[02:19:01.0000] <yecril71>
That is the text.

[02:20:00.0000] <yecril71>
I would also note that IE7 gets awfully slow when displaying the specification,

[02:20:01.0000] <yecril71>
event the multipage version.

[02:20:02.0000] <yecril71>
And the section headers are not visible at all.

[02:20:03.0000] <yecril71>
They are hidden under the green background.

[02:21:00.0000] <yecril71>
That makes it somehow hard to know what you are reading about.

[02:21:01.0000] <Hixie>
yeah, IE has all kinds of bugs

[02:21:02.0000] <Hixie>
i recommend using another browser

[02:21:03.0000] <Hixie>
/me fixes the line in question

[02:22:00.0000] <yecril71>
Thats all right, except that the cost of maintenance doubles.

[02:23:00.0000] <yecril71>
And you cannot get rid of IE7 in Windows.

[02:23:01.0000] <yecril71>
Of course, you can get rid of Windows, but that is quite an operation.

[02:24:00.0000] <yecril71>
I think it would be best for everybody to take that into account.

[02:24:01.0000] <hsivonen>
cost of maintenance doubles? Firefox, Chrome and Safari autoupdate themselves on Windows

[02:25:00.0000] <yecril71>
Only if run with administrative privileges, something I shall never do.

[02:25:01.0000] <Hixie>
ok, fixed the object/form line

[02:27:00.0000] <yecril71>
I think it would not be so much harm to get rid of the negative top margin for now.

[02:28:00.0000] <yecril71>
A browser window is not short of sheet space, and PDF can be used for printing.

[02:30:00.0000] <yecril71>
Borrowing the header backround from the following element seems like a dirty hack.

[02:30:01.0000] <Hixie>
it's pretty. and standards compliant. If IE can't handle it, that's not my problem or the spec's problem.

[02:31:00.0000] <yecril71>
You can make the PDF as pretty as you wish.

[02:31:01.0000] <Hixie>
i don't read the pdf

[02:31:02.0000] <yecril71>
This is not a beauty contest.

[02:32:00.0000] <hsivonen>
I think that Ubuntu is a better solution against Windows malware than running a personal Windows box without admin privileges

[02:32:01.0000] <yecril71>
It is better to be ugly than to be unreadable.

[02:32:02.0000] <Hixie>
so don't use IE

[02:33:00.0000] <Hixie>
making it slightly more readable isn't going to make the spec work in IE anyway

[02:33:01.0000] <Hixie>
IE doesn't handle the size of the page

[02:33:02.0000] <Hixie>
not much we can do about that

[02:33:03.0000] <yecril71>
It can read the multipage version.

[02:33:04.0000] <yecril71>
I do.

[02:34:00.0000] <Hixie>
Basically, the problem is with IE, not the spec. The solution is to have IE be fixed, not have the spec work around bugs in IE.

[02:34:01.0000] <yecril71>
I cannot have the IE fixed.

[02:34:02.0000] <hsivonen>
yecril71: doesn't IE8 work, either?

[02:34:03.0000] <yecril71>
You have to ask Philip`.

[02:34:04.0000] <Hixie>
you can't have the spec fixed either. :-)

[02:35:00.0000] <yecril71>
And borrowing the backround for the next element is far from being good markup.

[02:35:01.0000] <Hixie>
it's quite acceptable css

[02:35:02.0000] <hsivonen>
http://www.w3.org/TR/mobile-bp/#d0e704

[02:36:00.0000] <yecril71>
If you want the backround to be green, you have to choices:

[02:36:01.0000] <yecril71>
2 choices:

[02:36:02.0000] <yecril71>
wrap in a common ancestor, which does not apply here,

[02:36:03.0000] <Hixie>
hsivonen: as someone who has worked on browser vendors, i hate it when sites work around bugs in browsers

[02:36:04.0000] <yecril71>
or use the same class.

[02:36:05.0000] <Hixie>
or use a negative margin :-)

[02:36:06.0000] <Hixie>
which is fine :-)

[02:37:00.0000] <hsivonen>
Hixie: well, the mobile-bp doc doesn't acknowledge people from the companies you've worked for...

[02:38:00.0000] <hsivonen>
I think it's quite telling that Opera and Apple aren't acked in the Mobile BP stuff

[02:39:00.0000] <Hixie>
chaals wrote part of it

[02:39:01.0000] <Hixie>
iirc

[02:39:02.0000] <yecril71>
The only workaround for Internet Explorer is to disable CSS.

[02:40:00.0000] <yecril71>
That makes the page readable and the performance is much better.

[02:40:01.0000] <hsivonen>
Hixie: whoa. indeed. I was looking at the acks and thought the editor was from vodafone

[02:40:02.0000] <hsivonen>
I must have mixed up the editorships of different docs

[02:41:00.0000] <Hixie>
actually i thought a googler worked on that doc too, but i don't see that in the acks anywhere

[02:41:01.0000] <Hixie>
might be another one

[02:41:02.0000] <Hixie>
there are so many

[02:41:03.0000] <yecril71>
However, even if I add whatwg.org to restricted sites, that will not disable CSS by default.

[02:42:00.0000] <yecril71>
And I am sorry to see Hixie behave so arrogantly.

[02:43:00.0000] <Hixie>
if there was a bug in the spec's style sheet, would you ask the browsers to work around it?

[02:43:01.0000] <Hixie>
simple software engineering. you fix the bug, you don't work around the bug.

[02:44:00.0000] <yecril71>
Not a single browser is fully CSS-compliant.

[02:45:00.0000] <yecril71>
Your attitude is unrealistic, however you may not like it.

[02:45:01.0000] <yecril71>
The publisher should aim at the intersection of what is supported.

[02:45:02.0000] <hsivonen>
It's still baffling that a W3C REC doesn't include PNG support as part of the assumed image format support

[02:46:00.0000] <hsivonen>
how did *that* get past *principles*?

[02:46:01.0000] <hsivonen>
but then the markup language support is specced as XHTML Basic 1.1 [XHTML-Basic] delivered with content type application/xhtml+xml.

[02:48:00.0000] <yecril71>
And I cannot fix the bug in IE, as I already said.

[02:48:01.0000] <Hixie>
you also cannot fix the spec, so i don't see how that is different or relevant

[02:49:00.0000] <yecril71>
Well, but you can.

[02:49:01.0000] <yecril71>
With a very tiny amount of work.

[02:49:02.0000] <Hixie>
if you insist on using IE, then i recommend using http://dev.w3.org/html5/spec/Overview.html instead.

[02:51:00.0000] <yecril71>
That document does not have a multipage version.

[02:52:00.0000] <Hixie>
ah well, i tried

[02:53:00.0000] <gsnedders>
Hixie: Then why have work-arounds for IE at all?

[02:53:01.0000] <gsnedders>
like the /* that last decl is for IE6. Try removing it, it's hilarious! */ one

[02:54:00.0000] <Hixie>
i thought i'd gotten rid of that one

[02:54:01.0000] <yecril71>
Hixie�s theory about fieldset.name does not hold.

[02:54:02.0000] <yecril71>
The MSDN is correct here.

[02:54:03.0000] <gsnedders>
I don't see it in <http://www.whatwg.org/specs/web-apps/current-work/header-whatwg>, but I see it in <http://www.whatwg.org/specs/web-apps/current-work/>

[02:55:00.0000] <Hixie>
gsnedders: fixed

[02:55:01.0000] <yecril71>
So it is not really "already implemented", and it is different from HTML4.

[02:55:02.0000] <gsnedders>
/me heads off

[02:57:00.0000] <Hixie>
yecril71: do you have a testcase demonstrating the error in the spec?

[02:57:01.0000] <Hixie>
i'm pretty sure i tested this

[02:57:02.0000] <Hixie>
but i could be wrong!

[02:57:03.0000] <yecril71>
Just a minute.

[02:58:00.0000] <yecril71>
(That is, I have it, but I have to publish it.)

[03:00:00.0000] <hsivonen>
I think a lot of this mobile "best" practice stuff could go away if Opera Mini had serious competition

[03:00:01.0000] <hsivonen>
so that vendors of devices that can't host a self-contained browser didn't feel they have to commit to a single vendor if they want a decent browser

[03:01:00.0000] <Lachy>
unfortunately, working around browser bugs is an essential job that web developers must do for commercial sites. But for web standards, where people reading the spec are expected to use modern browsers, fixing such bugs is an unnecessary hassle

[03:01:01.0000] <Hixie>
I would say that I think a lot of this mobile "best" practice stuff could go away if Mobile Safari had serious competition. :-)

[03:01:02.0000] <Lachy>
yecril71, I have to agree with Hixie. I have no sympathy for IE users

[03:01:03.0000] <Hixie>
not just IE users

[03:02:00.0000] <Hixie>
users of any browser with pretty fundamental bugs

[03:02:01.0000] <Lachy>
yeah, all browsers have bugs. But IE is the worst.

[03:02:02.0000] <Lachy>
it's entirely possible to build a professional website for Firefox, Opera and Safari without using any hacks. But only the most basic sites work in IE without hacks

[03:03:00.0000] <hsivonen>
which reminds me that Validator.nu has a script error in IE (including 8)

[03:10:00.0000] <hsivonen>
hmm. the login thing may be a deep rathole

[03:14:00.0000] <yecril71>
http://www.2a.pl/~ne01026/test.htm

[03:14:01.0000] <hsivonen>
Hixie: regarding your email "Database section feedback": didn't Nikunj already volunteer? your email makes it look as though you are ignoring that he volunteered. (unless he volunteered only on the condition that he edits *all* the pieces he volunteered for)

[03:15:00.0000] <Hixie>
i have seen no evidence that he has volunteered other than him actually saying that he has volunteered

[03:15:01.0000] <yecril71>
(Note that the document is deliberately invalid now)

[03:15:02.0000] <hsivonen>
Hixie: I think I see what you mean.

[03:15:03.0000] <yecril71>
(It is supposed to be valid HTML5)

[03:16:00.0000] <Hixie>
could you rewrite that in JS so i can test it in other browsers?

[03:17:00.0000] <yecril71>
What for?  I do not contend it does not work in other browsers.

[03:17:01.0000] <Hixie>
/me doesn't understand what you are trying to show with that test

[03:17:02.0000] <yecril71>
That a fieldset is not a form control.

[03:17:03.0000] <Hixie>
(what's important is compatibility with all browsers, not just IE)

[03:18:00.0000] <Hixie>
fieldset is not a form control, correct

[03:18:01.0000] <yecril71>
All browsers, including IE.

[03:18:02.0000] <yecril71>
But HTML5 says it belongs to form.elements.

[03:18:03.0000] <Hixie>
the spec doesn't have a concept of "form control", though, so that seems academic

[03:18:04.0000] <yecril71>
The collection is named HTMLFormControlsCollection.

[03:18:05.0000] <yecril71>
Or something like that.

[03:19:00.0000] <yecril71>
That means it bears the concept of a form control.

[03:19:01.0000] <yecril71>
You have said you had to add this because that is what all browsers do.

[03:19:02.0000] <yecril71>
I have demonstrated it is not the case.

[03:20:00.0000] <Hixie>
the collection name is a historical artefact of little importance

[03:20:01.0000] <yecril71>
With that attitude, HTML5 is likely to become a historical artefact of little importance.

[03:20:02.0000] <Hixie>
it may be that it is not all browsers but just some browsers, then

[03:21:00.0000] <Hixie>
i expect one day it will be, yes

[03:21:01.0000] <yecril71>
I do not think that "some browsers" is a good argument to break logic and backward compatibility.

[03:22:00.0000] <Hixie>
according to http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E...%3Cform%3E%3Cfieldset%3E%3C%2Ffieldset%3E%3C%2Fform%3E%0A%3Cscript%3Ew%28document.forms[0].elements.length%29%3C%2Fscript%3E

[03:22:01.0000] <Hixie>
IE, Firefox, and Opera all put <fieldset> in form.elements

[03:22:02.0000] <yecril71>
I would rather ask those browsers to fix their implementations, because it is a bug.

[03:22:03.0000] <hsivonen>
Hixie: do I read correctly that the new http auth stuff doesn't ask browsers to change anything in their behavior?

[03:23:00.0000] <Hixie>
correct

[03:23:01.0000] <hsivonen>
ok.

[03:31:00.0000] <Hixie>
zcorpan!

[03:31:01.0000] <zcorpan>
hey Hixie

[03:32:00.0000] <Hixie>
now i wish i had said on irc what i wanted to tell you, for i have forgotten it

[03:32:01.0000] <zcorpan>
:(

[03:32:02.0000] <Hixie>
oh one was that i invented QUOTA_EXCEEDED_ERR with code 22, and wanted to ask you if you could add it along with codes 1-21 to web dom core

[03:33:00.0000] <zcorpan>
ok, will do

[03:33:01.0000] <yecril71>
All right, I give up.

[03:33:02.0000] <yecril71>
<http://msdn.microsoft.com/en-us/library/ms537449(VS.85).aspx#ctl00_rs1_WikiContent_2_Container>

[03:35:00.0000] <Lachy>
I still can't figure out why Pentasis is having such a difficult time comprehending the purpose of the time element, nor why he's suggesting such ridiculous changes.

[03:35:01.0000] <Lachy>
I guess he's just interested in some kind of theoretical purity, rather than trying to address any serious practical issues

[03:35:02.0000] <Hixie>
theoretical purity isn't a bad thing in and of itself

[03:36:00.0000] <Hixie>
if the people who extended html over the years had slightly more concern over theoretical purity, we'd be in a much better state

[03:36:01.0000] <zcorpan>
Hixie: added

[03:36:02.0000] <Hixie>
wow that was quick

[03:36:03.0000] <Hixie>
i need to ask you to do web dom core changes more often

[03:36:04.0000] <Hixie>
:-D

[03:36:05.0000] <zcorpan>
:)

[03:37:00.0000] <Hixie>
did i ask you what the eta was on a fpwd yet?

[03:37:01.0000] <zcorpan>
not sure but there's no eta yet

[03:37:02.0000] <Hixie>
k

[03:37:03.0000] <zcorpan>
i'm trying to get the spec moved to a w3c wg

[03:38:00.0000] <Hixie>
won't webapps take it?

[03:38:01.0000] <zcorpan>
haven't approached them yet

[03:38:02.0000] <Hixie>
oh one of the other things was a webkit bug i came across that was about weird (but compat-required) getElementById() behavior, iirc

[03:39:00.0000] <Hixie>
i tried cc'ing you but you didn't seem to have an account

[03:39:01.0000] <Hixie>
i forget which bug now

[03:39:02.0000] <zcorpan>
i think i'm still using the @hotmail account on b.w.o :S

[03:39:03.0000] <Lachy>
Hixie, true. But when your theory is suggesting that an element for marking up dates using ISO-8601 isn't adequate because it still can't accurately represent historical dates from centuries ago, rather than just worrying about the use cases it was designed for, then it's being taken too far

[03:39:04.0000] <zcorpan>
i guess i should change that

[03:40:00.0000] <hsivonen>
I think we should define input type=date to apply to booking of hotels & transport

[03:40:01.0000] <Hixie>
Lachy: i think he ended up actually saying the opposite -- that we should limit it further (e.g. not allow 2AD) because that was too historical and wasn't accurate either

[03:40:02.0000] <hsivonen>
and we should define <time> as a piece for microformats meant for scheduling secular civilian meetings

[03:40:03.0000] <zcorpan>
is there a use case for <input type=url multiple>?

[03:41:00.0000] <Lachy>
I've generally found the things he's said to be confusing

[03:41:01.0000] <zcorpan>
i think validator.nu has such a field doesn't it?

[03:41:02.0000] <Hixie>
zcorpan: i'm sure one could be invented

[03:41:03.0000] <Hixie>
whether it's a common enough case to worry about is another question

[03:41:04.0000] <Hixie>
nobody has asked for it yet

[03:41:05.0000] <Hixie>
afaik

[03:42:00.0000] <hsivonen>
for the v.nu use case, <input type=url multiple> would be backwards-incompatible with Opera

[03:42:01.0000] <hsivonen>
I guess <input type=email multiple> isn't nice for Opera, either

[03:42:02.0000] <zcorpan>
right

[03:43:00.0000] <yecril71>
Why are SCRIPT elements not allowed inside TABLE elements?

[03:43:01.0000] <hsivonen>
yecril71: legacy

[03:43:02.0000] <Hixie>
and keeping things simple

[03:44:00.0000] <Lachy>
sure, we could limit it more, but picking cut-off point at the unix epoch 1970-01-01 wouldn't given enough range for people to mark up, e.g. birthdates, and anywhere else would be just arbitrary

[03:44:01.0000] <yecril71>
Simple for whom?

[03:44:02.0000] <Hixie>
me

[03:44:03.0000] <Hixie>
and authors in general

[03:44:04.0000] <yecril71>
It is not simple that, once I need to produce a part of a table with a script,

[03:44:05.0000] <hsivonen>
Lachy: I think anything but 1970-01-01 and 0001-01-01 would be arbitrary

[03:44:06.0000] <Hixie>
Lachy: yeah

[03:44:07.0000] <yecril71>
I have to produce the whole table.

[03:44:08.0000] <zcorpan>
hsivonen: but <script>s aren't foster parented, right?

[03:45:00.0000] <hsivonen>
zcorpan: oh? I don't remember.

[03:45:01.0000] <Hixie>
yecril71: just use DOM manipulation

[03:46:00.0000] <yecril71>
(IE7 handles this use case cleanly)

[03:46:01.0000] <zcorpan>
hsivonen: yep. "A start tag whose tag name is one of: "style", "script""

[03:47:00.0000] <hsivonen>
ok

[03:47:01.0000] <Hixie>
<input type=hidden> is magical too

[03:47:02.0000] <Hixie>
and <script> needs to be made magical in <select>, which is going to be a pain

[03:48:00.0000] <yecril71>
That means it should be supported but the document is still nonconforming?

[03:48:01.0000] <hsivonen>
yecril71: I withdraw what I said about legacy

[03:49:00.0000] <yecril71>
Why?  TABLE elements cannot contain SCRIPT elements directly in HTML4 either.

[03:49:01.0000] <Hixie>
the restriction is one we inherited from html4 and one we will keep because allowing script in the middle of the table model encourages bad authoring practices (such as using document.write())

[03:50:00.0000] <hsivonen>
yecril71: I thought there was a parser legacy issue there. I wasn't referring to validation legacy.

[03:52:00.0000] <yecril71>
Correct me if I am wrong, but there is no way to ask the document to add more rows to the preceding table, unless that table has an ID.

[03:53:00.0000] <yecril71>
While it is possible inside.

[03:53:01.0000] <yecril71>
(supposing I place the SCRIPT right after the TABLE, that is.)

[03:54:00.0000] <Hixie>
just grab the last table from document.getElementsByTagName('table')

[03:54:01.0000] <yecril71>
Thanks.

[04:01:00.0000] <zcorpan>
Hixie: is it https://bugs.webkit.org/show_bug.cgi?id=6006 ?

[04:01:01.0000] <Hixie>
yes

[04:01:02.0000] <Hixie>
wow

[04:01:03.0000] <Hixie>
good call

[04:02:00.0000] <zcorpan>
first on a search for getelementbyid :)

[04:04:00.0000] <Hixie>
:-)

[04:06:00.0000] <zcorpan>
it seems we have lots of bugs saying that getElementById works with name=''

[04:06:01.0000] <zcorpan>
which we dropped in 9.5

[04:07:00.0000] <zcorpan>
and no bugs on it not working with name='', afaict

[04:07:01.0000] <zcorpan>
also, i think ie8 doesn't look at name='' (in ie8 mode)

[04:09:00.0000] <Hixie>
i'll trust you to spec something that matches the web and that browsers are willing to converge on

[04:09:01.0000] <yecril71>
I think if a statement needs an in-transaction callback and an after-transaction callback, that amounts to two callbacks.

[04:09:02.0000] <Hixie>
i'm just glad it's not my problem for once :-)

[04:09:03.0000] <yecril71>
So the most straightforward thing to do would be to allow two as required.

[04:10:00.0000] <yecril71>
You could also provide for a callback to figure out whether the transaction has finished

[04:10:01.0000] <yecril71>
and report that it needs to be called afterwards if not.

[04:11:00.0000] <Hixie>
/me looks at feedback from anne about %-encoding in name="" attributes and #fragids, and decides to call it a night

[06:46:00.0000] <hsivonen>
hmm. interesting. Gecko and HTML5 deal with noframes and noembed in very different ways

[06:46:01.0000] <hsivonen>
in terms of implementation

[06:46:02.0000] <hsivonen>
not necessarily from the POV of pages

[06:49:00.0000] <zcorpan>
hsivonen: how are they different?

[06:49:01.0000] <hsivonen>
Gecko keeps track of a nesting depth in noXXX elements and turns off <base> and form control handling when depth > 0

[06:50:00.0000] <hsivonen>
HTML5 treats noXXX as CDATA elements

[06:51:00.0000] <zcorpan>
hmm, my copy of firefox seems to insert a single text node in <noembed> -- not elements

[06:52:00.0000] <hsivonen>
zcorpan: it's possible that the tokenizer has changed and the depth tracking is now dead code

[06:52:01.0000] <hsivonen>
I was looking at the tree builder code

[06:52:02.0000] <zcorpan>
hsivonen: yeah, i remember dbaron saying there was similar dead code for <iframe> a while back

[06:53:00.0000] <zcorpan>
that disabled scripts or something

[06:53:01.0000] <hsivonen>
I should have that the tree builder code looks like that--not that Firefox does it :-)

[06:53:02.0000] <hsivonen>
can't really trust the looks of the Gecko parser code

[06:54:00.0000] <zcorpan>
is that code used for xhtml?

[06:54:01.0000] <hsivonen>
no

[07:00:00.0000] <Lachy>
if it's dead code, whats the point of keeping it around? Is it just that no-one has thought to remove it yet, and verify that it really is dead?

[07:05:00.0000] <hsivonen>
Lachy: most likely it's dead and forgotten and now no one wants to touch the parser more than absolutely necessary

[07:09:00.0000] <Lachy>
ok. I suppose it won't matter too much since I assume they'll be replacing the parser entirely with a new HTML5 parser soon enough

[07:09:01.0000] <hsivonen>
hopefully :-)

[07:41:00.0000] <zcorpan>
hey punctation is allowed in encoding declarations

[07:41:01.0000] <zcorpan>
you can make smileys out of encoding names

[07:42:00.0000] <zcorpan>
~u_^t^_f8

[07:45:00.0000] <Lachy>
zcorpan, I can't see how that example you gave can be seen as a smiley?

[07:48:00.0000] <zcorpan>
Lachy: dunno, come up with something better :)

[07:52:00.0000] <zcorpan>
hey you could even have multiline ascii art

[07:53:00.0000] <hsivonen>
/me can't wait to debug input with multiline ascii art charsets

[07:54:00.0000] <hsivonen>
w00t. the C++ version of the HTML5 parser *finally* links all the way

[07:54:01.0000] <Dashiva>
Is newline allowed, though?

[07:54:02.0000] <hsivonen>
any language that doesn't have C++-style linkage must give a huge productivity boost compared to C++

[07:55:00.0000] <Lachy>
zcorpan, where in the spec does it say punctation is allowed?

[07:55:01.0000] <Lachy>
is it that they're ignored for the parsing requirements, or that they're considered conforming too?

[07:56:00.0000] <zcorpan>
"The value must be a valid character encoding name, and must be the preferred name for that encoding."

[07:57:00.0000] <zcorpan>
hsivonen: validator.nu doesn't complain about punctation

[07:58:00.0000] <Lachy>
zcorpan, how do you interpret that as allowing punctuation?

[07:58:01.0000] <hsivonen>
zcorpan: thanks. I filed http://bugzilla.validator.nu/show_bug.cgi?id=337

[07:58:02.0000] <zcorpan>
Lachy: i don't :)

[07:58:03.0000] <Lachy>
wtf? You said "punctation is allowed in encoding declarations"

[07:58:04.0000] <zcorpan>
yeah, i was mistaken

[07:58:05.0000] <zcorpan>
i was just playing around in v.nu

[07:58:06.0000] <Lachy>
oh

[07:59:00.0000] <Philip`>
yecril71: The thread around http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2008-June/014984.html discusses the IE heading CSS bug, and that post suggests a simple workaround, but I guess Hixie cares more about the theoretical purity of the spec's markup than about its impact on users :-)

[09:41:00.0000] <BenMillard>
Lachy, I sometimes manage to make graphical websites which support Fx2, Fx3, O9, most recent Safari, IE6 and IE7 without hacks (re: http://krijnhoetmer.nl/irc-logs/whatwg/20081126#l-395)

[09:41:01.0000] <BenMillard>
usually, the show-stoppers aren't really CSS problems, it's fundamental breakage in the rendering engine :)

[09:42:00.0000] <Philip`>
/me can't even make non-graphical sites without finding browser bugs :-(

[09:42:01.0000] <BenMillard>
usually there are multiple ways of getting the same visual effect, like yecril71 points out

[09:43:00.0000] <BenMillard>
and when the job has a requirement that it *must* look correct in that set of browsers before your client pays you, well, you learn to compromise :)

[09:46:00.0000] <BenMillard>
yecril71, in applications on Windows, the equivalent of <fieldset> is called Frame (at least in VB6) and you add it to a form (aka window) as a control (aka a form control)

[09:46:01.0000] <BenMillard>
more specifically, it's a "container control" along with PictureBox, TabStrip and suchlike

[09:49:00.0000] <BenMillard>
krijnh, linkification stopped short by [ character: http://krijnhoetmer.nl/irc-logs/whatwg/20081126#l-424

[09:57:00.0000] <BenMillard>
Philip`, using position:relative on elements with negative margins is how I thought it was supposed to work, because I'm so used to doing that for IE :P

[09:58:00.0000] <BenMillard>
sometimes I use a negative top or left value instead of negative margin, due to margin bugs

[10:11:00.0000] <Philip`>
/me gets down to 3.6MB for an animated visualisation of the internal links in every 5th revision of the HTML 5 spec since its history began, which doesn't seem too bad

[10:11:01.0000] <Philip`>
BenMillard: Why does position:relative help in those cases?

[10:15:00.0000] <gsnedders>
/me wonders if that is danbri in one of his TPAC photos

[10:16:00.0000] <gsnedders>
No, it isn't.

[10:22:00.0000] <BenMillard>
Philip`, various strange things start working properly with position:relative in IE...often inexplicably :)

[10:22:01.0000] <BenMillard>
(like graphical bullets on list items in semi-arbitrary conditions)

[10:46:00.0000] <hsivonen>
do I read the html5lib correctly when I think it sanizes by discarding tokens before the tree builder?

[10:46:01.0000] <hsivonen>
s/html5lib/html5lib source/

[10:50:00.0000] <jgraham>
hsivonen: Yes

[10:51:00.0000] <hsivonen>
jgraham: does it discard script content?

[10:51:01.0000] <jgraham>
hsivonen: I believe so

[10:52:00.0000] <jgraham>
(but I guess there may be bugs)

[10:55:00.0000] <hsivonen>
It seems that the dominant design of HTML sanitizers is to throw stuff away between the tokenizer and the tree builder

[10:56:00.0000] <hsivonen>
Hixie: looks like different rules are called for here compared to the infoset coercion stuff

[10:56:01.0000] <hsivonen>
(not to suggest that HTML5 should prescribe the rules, but anyway)

[10:57:00.0000] <sayrer>
hsivonen, have you tested the new IE8 method?

[10:58:00.0000] <hsivonen>
sayrer: nope. What's the new IE8 method?

[10:58:01.0000] <sayrer>
hsivonen, toSafeHTML or some such

[10:58:02.0000] <sayrer>
hsivonen, yeah, that's it

[11:00:00.0000] <hsivonen>
"Update 11/20/08: changed reference to toSafeHTML to toStaticHTML" says IE blog

[11:01:00.0000] <hsivonen>
hmm. string to string method

[11:02:00.0000] <hsivonen>
sayrer: this is mail&news stuff, isn't it: http://mxr.mozilla.org/mozilla-central/source/content/base/src/mozSanitizingSerializer.cpp

[11:02:01.0000] <sayrer>
hsivonen, yeah. I decided I couldn't use that, way back when. Don't remember why.

[11:02:02.0000] <hsivonen>
sayrer: ok

[11:03:00.0000] <hsivonen>
I should test toStaticHTML to see what it actually does

[11:03:01.0000] <hsivonen>
thanks

[11:09:00.0000] <gsnedders>
http://www.flickr.com/photos/gsnedders/3061950334/

[11:09:01.0000] <gsnedders>
http://www.flickr.com/photos/gsnedders/3061106047/

[11:09:02.0000] <gsnedders>
who are those someones?

[11:09:03.0000] <hsivonen>
gsnedders: in 3061950334 Felix Sasaki

[11:09:04.0000] <hsivonen>
gsnedders: in 3061106047 Steve Zilles

[11:17:00.0000] <sayrer>
hsivonen, I may have run screaming from http://mxr.mozilla.org/mozilla-central/source/content/base/src/mozSanitizingSerializer.cpp#474 

[11:18:00.0000] <sayrer>
and also the pref parsing

[11:19:00.0000] <hsivonen>
sayrer: seems to be the wrong way round...

[11:32:00.0000] <gsnedders>
hsivonen: ah, thx

[12:11:00.0000] <hsivonen>
http://groups.google.com/group/mozilla.dev.planning/msg/f2dd45413cc68413

[12:16:00.0000] <hsivonen>
Hixie: am reading the spec correctly that it's OK to have an event loop spin between document.close() and the tokenizer emitting the EOF?

[12:17:00.0000] <gsnedders>
/me wonders if he gets a VPS how quickly he'll screw it up

[12:18:00.0000] <gsnedders>
hsivonen: that is lovely.

[12:43:00.0000] <Philip`>
gsnedders: You can always just reinstall it once you break everything

[12:43:01.0000] <gsnedders>
Philip`: :P

[12:44:00.0000] <Philip`>
/me discovered Ubuntu has "ufw", which makes firewall configuration actually sane - you say stuff like "ufw allow 80/tcp" and it does what you want, and you don't have to even know what iptables are

[12:49:00.0000] <gsnedders>
wow

[13:22:00.0000] <Hixie>
hsivonen: not only is it ok, it is required, because the tokeniser only emits stuff as part of the event loop

[13:26:00.0000] <hsivonen>
Hixie: doesn't the tokenizer emit non-EOF tokens immediately on first-level document.write()?

[13:26:01.0000] <Hixie>
yeah but that is still originally part of an event loop step

[13:26:02.0000] <hsivonen>
Hixie: but it's good that a spin is OK with document.close()

[13:27:00.0000] <hsivonen>
ok

[13:57:00.0000] <krijnh>
BenMillard: fixed, thanks

[13:59:00.0000] <Dashiva>
That was a short-lived feature (@html-auth)

[14:04:00.0000] <Philip`>
/me apologises for helping kill it

[14:07:00.0000] <Philip`>
(Actually I'm blame Julian, for originally suggesting that there could be a security issue)

[14:08:00.0000] <Dashiva>
Jonas would've picked up the slack anyhow, it seems :)

[14:08:01.0000] <sicking>
off with its head!

[14:11:00.0000] <Dashiva>
sicking: Do you have a secret new proposal that will rock our socks?

[14:11:01.0000] <sicking>
nothing secret

[14:11:02.0000] <sicking>
i've argued for something like OpenID for a while, just not very heavily

[14:12:00.0000] <sicking>
OpenID, the way it looks like now, with all its redirects and stuff, is no good though

[14:12:01.0000] <sicking>
but there a lot that can be done if we build it into the browser

[14:12:02.0000] <sicking>
basically we need something like microsofts CardSpace, but as a more open platform

[14:14:00.0000] <Dashiva>
Does cardspace avoid the "phishing enabling" of current openid?

[14:16:00.0000] <sicking>
that is my understanding

[14:16:01.0000] <sayrer>
sicking, how so?

[14:16:02.0000] <sicking>
however, i have not heard what all the complaints about neither cardspace nor openid are, so it's entirely possible that neither of them are very close to what we need

[14:17:00.0000] <sicking>
sayrer, you don't type a password, you just click on an image to choose which identity to use

[14:17:01.0000] <sayrer>
that is good

[14:17:02.0000] <sayrer>
fwiw, I did like the idea of using the 401 body for this

[14:17:03.0000] <hsivonen>
sicking: I think it would need to degrade gracefully into the current OpenID experience in browsers that don't implement the future thing

[14:17:04.0000] <Lachy>
Philip`, I find it difficult to believe that anyone would be stupid enough to introduce a XSS bug into a 401 page, especially because it's effectivly saying you need to log in before you can do anything

[14:17:05.0000] <sayrer>
only Safari is broken w.r.t. that extension point

[14:18:00.0000] <sicking>
hsivonen, it needs to degrade into something for sure

[14:18:01.0000] <hsivonen>
sicking: so in that sense, seeking to put the hook into the OpenID ID provider code might work

[14:18:02.0000] <Lachy>
and the XSS attack you outlined would need so many different bugs to occur in just the right way, it seems highly unlikely

[14:18:03.0000] <hsivonen>
but that would mean users would have to enter a URI into a field still

[14:18:04.0000] <hsivonen>
unless the field can be reliably autocompleted by the browser, too

[14:18:05.0000] <sayrer>
I had the idea that the 401 body should be the non-existant notion of svg static

[14:19:00.0000] <sayrer>
so browsers can create a difficult to simulte UI

[14:19:01.0000] <sayrer>
and show a little bit of branding next to it

[14:19:02.0000] <hsivonen>
what's svg static?

[14:20:00.0000] <sayrer>
svg with scripting, animation, fonts, etc

[14:20:01.0000] <sayrer>
er

[14:20:02.0000] <sayrer>
witout

[14:20:03.0000] <sayrer>
without

[14:20:04.0000] <hsivonen>
ah

[14:20:05.0000] <sayrer>
there is no subset that matches that

[14:20:06.0000] <sayrer>
but I'm thinking full-screen shadowed / UI box

[14:20:07.0000] <sayrer>
I guess flash or quicktime might be coerced

[14:20:08.0000] <sayrer>
to do that

[14:20:09.0000] <sayrer>
but it would be much harder

[14:21:00.0000] <Lachy>
although I'm still glad the www-authenticate feature was removed, since it seemed quite useless in practice for all but a very niche market

[14:22:00.0000] <Philip`>
Lachy: It doesn't seem implausible that someone would have e.g. login.php?return=... which gives you the login form with a <a href="$return">Go back</a> and returns 401 and WWW-Authenticate HTML because they want to tell bots to log in that way, and introduce the XSS hole that way

[14:22:01.0000] <Philip`>
Lachy: and I don't see what bugs the attack needs, other than the XSS one

[14:23:00.0000] <Philip`>
Lachy: Also, you shouldn't underestimate people's stupidity :-p

[14:25:00.0000] <Lachy>
it requires the bot to access the page via a URL which exploits the XSS attacks

[14:26:00.0000] <Philip`>
Bots follow links, and it's trivial to put a link onto most people's sites

[14:26:01.0000] <Lachy>
that may depend on the purpose of the bot, and where it was following links from

[14:26:02.0000] <Philip`>
(via blog comments, or referrer logs, or whatever)

[14:26:03.0000] <Philip`>
The purpose of the bot is to follow all the links it can find on the site :-)

[14:27:00.0000] <sayrer>
could just turn of scripts on 401 pages

[14:27:01.0000] <Philip`>
sayrer: There aren't any scripts involved here

[14:27:02.0000] <sayrer>
oh I see

[14:28:00.0000] <Lachy>
I suppose a such a link could occur on the site itself if it allowed user generated content of some kind

[14:28:01.0000] <Lachy>
and only used the 401 to prevent access to member areas

[14:30:00.0000] <Lachy>
it may not be a new problem though. If there are bots that perform this kind of log in already, by being manually configured with the form name, they would be vulnerable to the same attack

[14:30:01.0000] <Lachy>
although without the www-authenticate header advertising that, it's less likely

[14:31:00.0000] <Lachy>
but the same attack could be used against users by getting them to follow the link

[14:33:00.0000] <Philip`>
It's not a problem unique to WWW-Authenticate, but WWW-Authenticate was designed in a way that encourages behaviour that would encounter that problem

[14:51:00.0000] <Hixie>
i hven't removed it yet btw

[14:52:00.0000] <Hixie>
somewhat waiting for someone to take it over or for julian to say he doesn't want it or to propose some other alternative first

[14:55:00.0000] <Hixie>
we seem to be getting close to the point (or maybe past the point) where i need to respond to this <i>/<small> thread

[14:58:00.0000] <KrocCamen>
i/small? I use those, what’s going on with them?

[14:59:00.0000] <Hixie>
nothing especially

[14:59:01.0000] <Hixie>
but there's been a thread about them for some time now

[14:59:02.0000] <Hixie>
which is starting to get very "meta"

[15:00:00.0000] <KrocCamen>
I use i as a comment marker in code spans (so that in RSS readers it shows up as italics, but has no particular meaning in the HTML5 source). Small I use either inside a paragraph—as normal—but also as a paragraph iteself (since you can use other elements instead of P where they are better suited)

[15:07:00.0000] <KrocCamen>
How all the HTML4 elements and the new HTML5 sectioning elements are at the moment is just how I like it. My one and single miff is the removal of type on OL/UL. I need that for my site. I’ve no classes and my article text uses links to a footnote list - which are in alpha. Since the article text has to use the real letters to link to the footnote, it’s a structural thing, not merely a style.

[15:09:00.0000] <Hixie>
css should support (and will support) substituting in the footnote number in the appropriate place

[15:09:01.0000] <sicking>
Hixie, got a sec to talk about baseURIs and security contexts of |new Worker|?

[15:09:02.0000] <Hixie>
sure

[15:10:00.0000] <Hixie>
the spec says to use the origin of the worker's script, and limits the worker's script to same-origin of the creator

[15:10:01.0000] <Hixie>
which makes it somewhat simple and allows us flexibility to extend it in future in different directions

[15:10:02.0000] <Hixie>
depending on what authors want

[15:10:03.0000] <sicking>
Hixie, so basically, the way that |new Worker| works is that it establishes a new security context for the new worker.

[15:11:00.0000] <Hixie>
not sure what a "security context" is, but it does create a new scripting context, yes

[15:11:01.0000] <sicking>
well, every page has a 'principal'. i.e. a something that we do all neccessary security checks against

[15:12:00.0000] <Hixie>
i think the spec just uses the origin for that

[15:12:01.0000] <Hixie>
but that's an implementation detail

[15:12:02.0000] <Hixie>
continue

[15:12:03.0000] <sicking>
when setting up a new worker that means that we have to set up a new principal

[15:13:00.0000] <sicking>
another way to look at it is that it sets up a new browsing context i guess

[15:13:01.0000] <sicking>
this makes a lot of sense to me for shared workers, since aren't coupled with any one browsing context

[15:14:00.0000] <sicking>
but i'm less sure that makes sense for a dedicated worker, which is pretty tightly coupled with a single 'normal' browsing context

[15:16:00.0000] <Hixie>
the way the spec defines it side-steps this problem

[15:16:01.0000] <Hixie>
iirc

[15:16:02.0000] <sicking>
the end result is that you have to apply a same-origin restriction on |new Worker|

[15:17:00.0000] <sicking>
and that uris are resolved agasint the uri of the worker, rather than the uri of the opening page

[15:17:01.0000] <Hixie>
yes

[15:17:02.0000] <sicking>
neither of which i'm sure are desireable for dedicated workers

[15:18:00.0000] <Hixie>
oh we definitely want the base URI to be the base URI of the worker, otherwise the behavior of hte worker changes based on where it is called

[15:18:01.0000] <Hixie>
but we invented lcoation.resolveURL() to address that

[15:18:02.0000] <sicking>
that's what you have with <script> today though, and it doesn't seem to be a problem

[15:19:00.0000] <Hixie>
it's only not a problem because people have to write all their shared scripts using absolute URLs everywhere

[15:19:01.0000] <Hixie>
otherwise they wouldn't work

[15:20:00.0000] <Hixie>
anyway as mentioned earlier, the limit on same-origin workers is intentional. it lets us ship this, then work out what people want to do with non-same-origin dedicated workers -- have them run in their original origin, AC protected; have them run on the caller's origin, or something else

[15:22:00.0000] <sicking>
well, that's definitely what we'll have to do as a consequence of workers running with the security context of their own uri

[15:22:01.0000] <sicking>
if we didn't have do that we'd be in the same situation as <script> and .importScripts

[15:23:00.0000] <Hixie>
<script> is a mess, we don't want to copy that :-)

[15:23:01.0000] <sicking>
.importScripts copies that

[15:23:02.0000] <Hixie>
copies what?

[15:24:00.0000] <Hixie>
importScript does same-origin checks

[15:24:01.0000] <sicking>
really?

[15:24:02.0000] <sicking>
why?

[15:24:03.0000] <Hixie>
yes, you complained about it once

[15:24:04.0000] <sicking>
i thought it got changed

[15:24:05.0000] <Hixie>
doesn't look like it

[15:24:06.0000] <sicking>
ugh

[15:24:07.0000] <sicking>
why do same-origin checks there?

[15:24:08.0000] <Hixie>
did i respond to your complaint?

[15:25:00.0000] <sicking>
yes, don't remember what you said though

[15:25:01.0000] <Hixie>
me either

[15:26:00.0000] <sicking>
so why have same-origin restrictions on importScripts?

[15:26:01.0000] <Hixie>
/me looks in the archives

[15:27:00.0000] <Hixie>
my reply was:

[15:27:01.0000] <Hixie>
I don't recall the precise reason, but I seem to recall concern over 

[15:27:02.0000] <Hixie>
specific attack vectors are what caused us to restrict this.

[15:27:03.0000] <Hixie>
i guess we can change it

[15:28:00.0000] <sicking>
it seems like behaving exactly like <script> won't cause any new attack vectors

[15:28:01.0000] <Hixie>
well like i said, <script> is hardly what we want to be copying

[15:29:00.0000] <Hixie>
are we not concerned about people faking a browsing context environment and then calling someone's cookie-protected JSON?

[15:30:00.0000] <sicking>
sure, the same way that it can happen for <script>

[15:30:01.0000] <Hixie>
don't we want to stop it for <script> too?

[15:30:02.0000] <Hixie>
in due course?

[15:30:03.0000] <sicking>
yeah, but that's never going to happen

[15:30:04.0000] <Hixie>
seems bad to introduce new ways to do things we know are bad

[15:31:00.0000] <sicking>
if we are going to drive it through for <script>, the extra amount of work involved in driving it through for .importScripts is probably smaller than can be measured

[15:31:01.0000] <Hixie>
the well i checkedin the change

[15:32:00.0000] <sicking>
cool

[15:32:01.0000] <sicking>
so i'm still not convinced on |new Worker|

[15:32:02.0000] <sicking>
but i know that me and aaron has different mental models there

[15:33:00.0000] <sicking>
as in, is |new Worker| simply an out-of-thread <script>? Or is it more like an <iframe>

[15:33:01.0000] <Hixie>
well right now the spec takes the conservative approach that is compatible with both mental models, allowing us to pick either one later, based on author feedback

[15:33:02.0000] <Hixie>
which i think is probably a good approach for now

[15:33:03.0000] <sicking>
well, once we set the baseURI we can't change that, so i suspect whatever we choose now we'll stick with forever

[15:34:00.0000] <Hixie>
i don't see how it is dependent on baseURI

[15:34:01.0000] <sicking>
i think the baseURI will go hand in hand with the security context

[15:35:00.0000] <Hixie>
well in that case i definitely would want to do the model we hve now

[15:35:01.0000] <Hixie>
that <script> has the document's baseURI is something we fixed with CSS and we should definitely fix it here too

[15:35:02.0000] <Hixie>
relative URLs shouldn't change meaning based on who is using them

[15:35:03.0000] <Hixie>
that's just wacky

[15:35:04.0000] <Hixie>
based on who is referring to the document that contains them, i should say

[15:36:00.0000] <sicking>
depends on what you are doing. Any time you pass a uri across baseURI borders you have pain

[15:36:01.0000] <sicking>
since you can't use relative URIs any more

[15:37:00.0000] <sicking>
so that means that you have pain any time you pass a URI through postMessage

[15:38:00.0000] <Hixie>
that's why we added location.resolveURL()

[15:39:00.0000] <sicking>
right, it doesn't change the fact that you have pain through

[15:40:00.0000] <sicking>
resolving relative URIs is something developers aren't used to

[15:40:01.0000] <sicking>
(since it has been largely impossible before)

[15:45:00.0000] <Hixie>
yeah, you're trading one pain for another

[15:45:01.0000] <Hixie>
but one of them is predictable :-)

[15:45:02.0000] <sicking>
the only time you have pain today is when you are fetching a resource relative to the script file

[15:45:03.0000] <sicking>
(today == in windows)

[15:46:00.0000] <sicking>
but not when fetching resources relative to the document

[15:46:01.0000] <Hixie>
right, and the only pain we're adding is when you're passing a relative url around

[15:46:02.0000] <sicking>
not sure which is most common, but i'd imagine most data would live close to the document

[15:47:00.0000] <sicking>
well, most uris are relative i'd think

[15:47:01.0000] <Hixie>
i don't think most workers are going to involve passing urls around frankly

[15:48:00.0000] <sicking>
any time you want to use XHR

[15:48:01.0000] <sicking>
granted, that's going to be an extra common case in our impl since that is the only API we have avaialble (we won't have localStorage support in FF3.1)

[15:48:02.0000] <Hixie>
nah, most of the time the script will be hard-coded to fetch particular files

[15:49:00.0000] <sicking>
oh?

[15:49:01.0000] <Hixie>
i'd expect so

[15:49:02.0000] <Hixie>
why would you put the logic in the main document and just offload the loading? it doesn't gain you anything

[15:49:03.0000] <Hixie>
you could just as easily do the loading in the main document

[15:49:04.0000] <sicking>
loading+processing

[15:50:00.0000] <sicking>
the whole point with dedicated workers is that the processing is heavy

[15:50:01.0000] <sicking>
so unless you have the data available inline in the main document you're probably going to want to fetch it using XHR

[15:51:00.0000] <Hixie>
right but when do you do laoding and processing without knowing what you'll be loading?

[15:51:01.0000] <sicking>
which of course you could do on the main thread and then pass it off to the worker

[15:53:00.0000] <sicking>
a parsing/processing library wouldn't know what to parse/process until given the data

[15:54:00.0000] <sicking>
the strongest argument that i can think of for using the securitycontext/baseURI of the script file is that that is what we'll have to do for shared workers

[15:54:01.0000] <Hixie>
most of the time you'll be doing something like "fetch the contacts list and keep me updated as to progress" or stuff like that, as far as i can tell

[15:55:00.0000] <sicking>
in a shared worker I agree

[15:55:01.0000] <sicking>
but why would you do that in a dedicated worker?

[15:55:02.0000] <sicking>
rather than just doing it on the main thread

[15:56:00.0000] <Hixie>
so that you can use sync xhr, stuff like that

[15:56:01.0000] <sicking>
"stuff like that"?

[15:56:02.0000] <Hixie>
not yielding

[15:56:03.0000] <Hixie>
workers are going to be mostly used just because they're hugely convenient instead of having to keep worrying about other scripts

[15:57:00.0000] <sicking>
but if all you are doing is pulling down some light-weight file, you're going to have to jump through as many hoops to offload to a worker thread, as you would have to jump through using async XHR on the main thread

[15:58:00.0000] <Hixie>
for the contacts case, i would imagine the worker would be refetching the file regularly, doing some pretty complex processing, and just posting change notifications over

[15:58:01.0000] <Hixie>
that's a lot of work


2008-11-27
[16:01:00.0000] <sicking>
so the question is, is the script going to have the URI to pull data from built in? Or is it going to receive it from the document

[16:02:00.0000] <Hixie>
i don't see why it wouldn't

[16:02:01.0000] <sicking>
if you want to pull data from several sources then you wouldn't build it in

[16:03:00.0000] <sicking>
or rather

[16:04:00.0000] <sicking>
do you host a script centrally that can be used to process data from several locations, or do you host one script per uri that serves data

[16:04:01.0000] <Hixie>
what are the cases where you'll be doing the same processing on a whole bunch of files whose urls you wouldn't hard code?

[16:05:00.0000] <sicking>
if you have several uris all serving data in the same format

[16:05:01.0000] <sicking>
to use your example:

[16:05:02.0000] <Hixie>
yes, what case would that be?

[16:07:00.0000] <sicking>
you have several URIs that serve JSON objects, one serving address book, one serving folder list, one serving inbox. And you want to be notified about changes in either of them

[16:07:01.0000] <sicking>
or

[16:07:02.0000] <sicking>
you have a library that can parse word documents and converts them to JSON objects being passed to the main thread

[16:07:03.0000] <Hixie>
wouldn't you have a worker for the address book, one for the folder list, one for the inbox, etc? they'd all be different, surely?

[16:08:00.0000] <Hixie>
where are the word documents going to come from?

[16:08:01.0000] <Hixie>
anyway, i agree that it's possible for urls to be a problem

[16:08:02.0000] <Hixie>
i just don't see it being a common one

[16:08:03.0000] <Hixie>
i don't think any of the examples i wrote ended up passing a url over the channel

[16:09:00.0000] <sicking>
as soon as your code is generic enough that it can process data from several uris you'll run into this i'd think

[16:09:01.0000] <sicking>
and having multiple uris serve the same data format seems not that weird to me

[16:10:00.0000] <sicking>
i think the question really is, do you have one worker per data source, or do you have generic libraries for multiple data sources

[16:11:00.0000] <Hixie>
i think the former will be more common, personally

[16:11:01.0000] <sicking>
I don't know. Is that the case today?

[16:12:00.0000] <sicking>
generally in the coding I do I separate data from code

[16:12:01.0000] <sicking>
but i know the web often is less clean

[16:14:00.0000] <Hixie>
basically at this point from my point of view here's where i stand:

[16:15:00.0000] <Hixie>
i have feedback from credible people arguing it should be both ways

[16:15:01.0000] <Hixie>
(you and aaron in particular)

[16:15:02.0000] <Hixie>
making the baseURI be the worker's baseURI has three advantages:

[16:15:03.0000] <Hixie>
- it's more resilient to reuse if there are relative urls in the source

[16:16:00.0000] <Hixie>
- it's more compatible with shared workers

[16:16:01.0000] <Hixie>
- it's the "right" think according to web architecture principles

[16:16:02.0000] <Hixie>
s/think/thing/

[16:17:00.0000] <Hixie>
so for me there's a weak bias towards doing it the way we have now

[16:17:01.0000] <Hixie>
i agree that it's not an especially strong argument

[16:17:02.0000] <Hixie>
but i can't really justify doing it the other way around, as that is a (mildly) weaker position

[16:19:00.0000] <sicking>
i think the middle one is the strongest argument there

[16:19:01.0000] <sicking>
the first one can be countered by that it's less resilient if the URIs are passed in from postMessage

[16:20:00.0000] <sicking>
and the third one only applies if the URI originates in the source, which we don't actually know since all we get is a string

[16:20:01.0000] <sicking>
the shared worker argument i do buy though

[16:22:00.0000] <sicking>
the only counter argument i have is that if we allow the baseURI (and security context) be that of the document, we can allow |new Worker| to load scripts cross site

[16:23:00.0000] <sicking>
i.e. we could allow people to write generic crypto/parsing/processing/etc libraries that can be hosted on their site and allow cross-site linking

[16:23:01.0000] <sicking>
however you can do that mostly using importScripts, you just need to host a thin wrapper on your own site

[16:23:02.0000] <Hixie>
we could allow cross-site workers even without doing that, just have the workers run in the origin of the third-party host, and use AC to protect random scripts

[16:23:03.0000] <Hixie>
or yeah, use importScripts() from a data: URL

[16:24:00.0000] <Hixie>
new Worker("data:text/javascript,importScripts('http://example.com/shared.js')");

[16:25:00.0000] <sicking>
ugh

[16:25:01.0000] <sicking>
hmm.. we won't allow data: with the current patches...

[16:25:02.0000] <sicking>
that can be fixed of course

[16:25:03.0000] <sicking>
just using AC is the wrong thing

[16:26:00.0000] <sicking>
AC means that you can read the resources, not that you can execute script in its security context

[16:26:01.0000] <Hixie>
ok actually data: won't work as currently specced

[16:26:02.0000] <Hixie>
AC2 then :-)

[16:26:03.0000] <Hixie>
we'll need to change AC a bit for <video> v2 anyway

[16:29:00.0000] <sicking>
if we want to allow cross-site workers we'll probably need something very different i think

[16:31:00.0000] <sicking>
/me is also very unhappy about the result of <video>

[16:38:00.0000] <Hixie>
yeah me too

[16:38:01.0000] <Hixie>
allowing cross-site video is dumb imho

[16:59:00.0000] <Hixie>
Lachy: fwiw, restarting NetAuthAgent on the connecting machine fixed my screen sharing

[17:22:00.0000] <Lachy>
Hixie, ok

[17:23:00.0000] <BenMillard>
krijnh, after noodling with your IRC Logs front page for a while, I've put some ideas here: http://projectcerbera.com/!dev/irc-logs/front

[17:23:01.0000] <BenMillard>
(files it uses are listed here: http://projectcerbera.com/!dev/irc-logs/)

[17:23:02.0000] <Lachy>
Hixie, re the <dialog> discussion, I think it would be a bad idea to keep both <dialog> and aria-role="dialog" when they mean different things

[17:24:00.0000] <Lachy>
and given that aria is being implemented, the best resolution would be to revisit that <dialog> naming debate :-(

[17:24:01.0000] <Lachy>
(or one of them should be dropped)

[17:28:00.0000] <BenMillard>
krijnh, I think Table + Link is my favourite listing for the front page (if you reply in this channel when I'm not here, I'll still see it in the logs)

[17:34:00.0000] <Hixie>
Lachy: yeah, something will likely have to change

[18:33:00.0000] <Lachy>
I updated the decriptions of the element categories in the authoring guide http://dev.w3.org/html5/html-author/#categories

[18:42:00.0000] <Hixie>
can anyone find me a select widget on mac os x or windows default installs that is initially blank?

[18:42:01.0000] <Hixie>
but where nothing is not a valid value?

[18:42:02.0000] <Hixie>
trying to find precedent for the idea of <select required> with a non-valid initial value

[18:44:00.0000] <Lachy>
I can't recall any. I think that practice may have only become widely used in web pages

[18:49:00.0000] <sayrer>
Hixie, pretty sure the OS X installation process has such things

[18:50:00.0000] <sayrer>
maybe that's only text boxes, though

[23:17:00.0000] <BenMillard>
jgraham, here's the first section of that Wikipedia timeline in the Table Inspector: http://xrl.us/oyipv

[23:22:00.0000] <BenMillard>
jgraham, I retrofitted <th> for header cells since that's what I remember it using first time I reviewed it. The actual table now uses <td> for header cells: http://en.wikipedia.org/wiki/Template:Ferrari_road_car_timeline_1960-2009

[23:23:00.0000] <BenMillard>
jgraham, note that the decade column headers get highlit correctly but the yearly column headers don't; in contrast even the most complicated row relationships are highlight perfectly. (Was discussed here: http://lists.w3.org/Archives/Public/public-html/2008Oct/thread.html#msg121)

[23:42:00.0000] <krijnh>
BenMillard: come back! ;p

[00:14:00.0000] <yecril71>
Philip`!  The resistance against making the headers visible in IE7 is not about theoretical purity.

[00:14:01.0000] <yecril71>
It is not theoretically pure to borrow the background from the following element.

[00:15:00.0000] <yecril71>
I guess it is more about "pushing it to the limit".

[00:17:00.0000] <yecril71>
BenMillard!  Thanks for the info about controls in VB6.

[00:18:00.0000] <yecril71>
I still insist "control" is a misnomer because it cannot be used to control anything.

[00:18:01.0000] <yecril71>
Think about a dashboard in an a�roplane.

[00:19:00.0000] <yecril71>
There are controls and there are panes.

[00:19:01.0000] <yecril71>
Panes are not controls.

[00:25:00.0000] <yecril71>
The title attribute has a problem that it cannot be declared as a rule.

[00:26:00.0000] <yecril71>
You have to laboriously define it on every element where it applies.

[00:26:01.0000] <yecril71>
Or use a script to do so.

[00:27:00.0000] <yecril71>
I think CSS should take care of that sooner or later.

[00:29:00.0000] <yecril71>
It is crazy to suggest that the closing tag could be optional if the opening tag has some attribute.

[00:29:01.0000] <yecril71>
That would make the parser very hard to maintain.

[01:23:00.0000] <hsivonen>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cform%20id%3Da%3Efoo%3C%2Fform%3E%3Cscript%3Edocument.getElementsByTagName(%27form%27)[0].innerHTML%20%3D%20%22%3Cform%20id%3Db%3Ebar%3C%2Fform%3E%22%3B%3C%2Fscript%3E

[01:27:00.0000] <yecril71>
As of IE7, unknown run-time error.

[01:28:00.0000] <yecril71>
And the FORM is in the HEAD and contains the BODY.

[01:30:00.0000] <hsivonen>
the result in IE8 is interesting

[01:31:00.0000] <hsivonen>
adding a <body> tag helps to see what I actually meant to test

[01:32:00.0000] <hsivonen>
so http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Cbody%3E%3Cform%20id%3Da%3Efoo%3C%2Fform%3E%3Cscript%3Edocument.getElementsByTagName('form')%5B0%5D.innerHTML%20%3D%20%22%3Cform%20id%3Db%3Ebar%3C%2Fform%3E%22%3B%3C%2Fscript%3E

[01:32:01.0000] <yecril71>
Unknown run-time error.

[01:33:00.0000] <hsivonen>
yeah, that happens in IE8, too

[01:36:00.0000] <yecril71>
What happens if you appendChild instead?

[01:37:00.0000] <yecril71>
The text 

[01:37:01.0000] <yecril71>
corresponds to E_FAIL.

[01:37:02.0000] <hsivonen>
I haven't tried

[01:37:03.0000] <yecril71>
This is how mshtml responds to someone trying to torture it.

[01:45:00.0000] <yecril71>
I can see no reason why the command elements should be hidden from the viewer and accessible from the keyboard only.

[01:45:01.0000] <yecril71>
This is extremely user-unfriendly.

[01:58:00.0000] <hsivonen>
does someone already have a mutation event console similar to Hixie's DOM viewer?

[02:07:00.0000] <Lachy>
hsivonen, would that be a console that logs all DOM3 mutation events like DOMNodeInserted, DOMAttrModified, etc?

[02:09:00.0000] <hsivonen>
Lachy: yes

[02:27:00.0000] <Lachy>
I'm writing a blog entry about input type=search and placeholder. See draft: http://blog.whatwg.org/?p=371&preview=true

[02:27:01.0000] <Lachy>
Are there any other advantages it has only type=text, which I haven't mentioned?

[02:28:00.0000] <Lachy>
(only blog admins will be able to see that)

[02:43:00.0000] <hsivonen>
/me notes that Mark Birbeck's slides from a Google Tech Talk on RDFa don't show namespace declaration or CURIE prefixes

[02:43:01.0000] <hsivonen>
hmm. those are the same slides as the 20/20 slides from XTech

[03:14:00.0000] <zcorpan__>
hsivonen: http://livedom.validator.nu/?%3Cs%20o%3D%26%20t doesn't give what i expected

[03:17:00.0000] <hsivonen>
zcorpan__: what did you expect?

[03:17:01.0000] <hsivonen>
ooh. the t appears in two places

[03:17:02.0000] <hsivonen>
that's ... interesting

[03:18:00.0000] <zcorpan>
yeah :)

[03:18:01.0000] <zcorpan>
html5lib gives what i expected

[03:19:00.0000] <hsivonen>
zcorpan: thanks

[03:19:01.0000] <hsivonen>
bug filed http://bugzilla.validator.nu/show_bug.cgi?id=338

[03:25:00.0000] <tthorsen>
speaking of html5lib - I noticed that it handles the following markup differently from the html5 spec: <option><select>text

[03:26:00.0000] <zcorpan>
tthorsen: <option> handling was changed in the spec to work with <datalist>

[03:26:01.0000] <tthorsen>
the html5 spec thinks <option>'s in "in body" is ok, whereas html5lib thinks it's an error

[03:27:00.0000] <tthorsen>
sure. In the implementation I'm working on, I'm doing what the spec says.

[03:27:01.0000] <tthorsen>
does someone maintain html5lib?

[03:27:02.0000] <zcorpan>
tthorsen: hsivonen's html parser is more in line with the spec than html5lib, i think

[03:29:00.0000] <tthorsen>
zcorpan: is that the validator.nu page?

[03:30:00.0000] <zcorpan>
tthorsen: yeah

[03:30:01.0000] <zcorpan>
http://about.validator.nu/htmlparser/

[03:30:02.0000] <Philip`>
tthorsen: html5lib isn't maintained to match the spec at all times - usually it just slowly diverges as the spec changes, and then someone nice comes along and fixes it all

[03:30:03.0000] <hsivonen>
also, the trunk of the Validator.nu parser and livedom.validator.nu are out of sync (which sucks but I've been busy)

[03:31:00.0000] <Philip`>
which might take many months

[03:35:00.0000] <tthorsen>
I see. The livedom.validator.nu does <option><select> correctly, but validator.nu complains that the <option> is an error

[03:37:00.0000] <zcorpan>
seems about right

[03:37:01.0000] <tthorsen>
hsivonen: livedom.validator.nu does not do <option><optgroup><select> correctly, though. I believe it should close the open <option> when it receives the <optgroup> start tag

[03:37:02.0000] <tthorsen>
I think the result should be <option></option><optgroup><select>

[03:43:00.0000] <hsivonen>
let's see if my GWT build still runs...

[03:50:00.0000] <Lachy>
http://lists.w3.org/Archives/Public/www-archive/2008Nov/0054.html - I wonder how we could get the HTML WG to work more with the WHATWG and avoid such hostility from some people, without forcing everyone to subscribe to both lists or merging them into one (neither of which are reasonable options)

[03:53:00.0000] <Philip`>
We could cross-post everything

[03:54:00.0000] <Philip`>
or subscribe public-html onto the whatwg list

[04:11:00.0000] <zcorpan>
"we could ask the HTML WG to change the spelling of the element to DIALOUGUE" -- that's why it shouldn't be spelt 'dialogue'

[04:14:00.0000] <zcorpan>
i think we should spell it "dl" :)

[04:20:00.0000] <Philip`>
<collquy> would be much better

[04:20:01.0000] <Philip`>
Uh

[04:20:02.0000] <Philip`>
<colloquy>

[04:22:00.0000] <hsivonen>
/me agrees with spelling it dl

[04:23:00.0000] <tthorsen>
Seems both dialogue and dialog is used in the html5 spec. Same goes for color vs colour. Is that what this is about?

[04:27:00.0000] <hsivonen>
the syntax of computer languages should use en-US, so dialogue shouldn't even be on the table

[04:28:00.0000] <krijnh>
/me votes for <dl class="notdefinitionlistbutthatotherone">

[04:30:00.0000] <tthorsen>
I like dialog better than dl, simply because dialog is a lot more greppable. Searching for the string "dl" in the spec or in an html file does not sound like fun.

[04:31:00.0000] <hsivonen>
in an HTML file, you can grep for "<dl"

[04:31:01.0000] <tthorsen>
I don't care very strongly about this, though...

[04:33:00.0000] <tthorsen>
hsivonen: true. Is getElementsByTagName('dl') a legal thing to do?

[04:50:00.0000] <Lachy>
JohnResig, yt?

[05:00:00.0000] <Lachy>
hsivonen, are you sure dialogue isn't used in US english at all?

[05:01:00.0000] <hsivonen>
Lachy: hmm. I guess dialogue is allowed in en-US for the conversation sense

[05:02:00.0000] <hsivonen>
Lachy: but it's a really bad idea to rely on locale-specific distinctions in spelling for disambiguation

[05:02:01.0000] <Lachy>
I generally spell it dialogue when referring to conversation and dialog when referring to a computer dialog box

[05:03:00.0000] <Lachy>
yeah, that's true

[05:04:00.0000] <hsivonen>
/me tries to lobby for a UTF-8-only policy for Ogg text tracks

[05:05:00.0000] <Lachy>
hsivonen, on which mailing list?

[05:05:01.0000] <Lachy>
/me agrees that's a good idea

[05:05:02.0000] <MikeSmith>
Lachy: yeah, I think US English does uses both dialog and dialogue

[05:05:03.0000] <hsivonen>
Lachy: accessibility at xiph

[05:06:00.0000] <Lachy>
http://lists.xiph.org/pipermail/accessibility/2008-November/000175.html

[05:07:00.0000] <Lachy>
hsivonen, are you sure it's true that all character sets can be mapped to Unicode? Someone told me at TPAC that there were some known issues with mapping Shift-JIS to unicode for some characters, IIRC

[05:09:00.0000] <MikeSmith>
I think there are some obscure person-name characters in Shift-JIS that are not in Unicode

[05:09:01.0000] <MikeSmith>
variants for people's names

[05:09:02.0000] <Lachy>
that same person also tried to tell me that there were plans to develop a new Unicode-encoding optimised more for asian languages, as an alternative to UTF-8 and -16

[05:09:03.0000] <hsivonen>
Lachy: IIRC, there has been some wobbliness in the mapping from Shift_JIS to Unicode, but Real Software(tm) uses UTF-8 or UTF-16 in RAM anyway, so there's no way to do Shift_JIS without coercing it to Unicode somehow.

[05:09:04.0000] <Philip`>
MikeSmith: Those people should just change their names, to save us all the bother

[05:09:05.0000] <MikeSmith>
heh

[05:09:06.0000] <Lachy>
heh

[05:10:00.0000] <Lachy>
all the asian people I know seem to have an ASCII representation of their names anyway, so they should just use that

[05:10:01.0000] <hsivonen>
well, then there's ARMSCII, which, IIRC, contains one character that you'd need to map to PUA

[05:10:02.0000] <hsivonen>
just like you need to map the Apple logo from MacRoman onto PUA

[05:11:00.0000] <MikeSmith>
banks and government in Japan have to support whatever family-name characters people use, even if very rare. this is one reason for anti-Unicode sentiment in Japan

[05:11:01.0000] <Lachy>
ok, that makes sense

[05:11:02.0000] <hsivonen>
I'd be seriously unhappy if my parents had tried to be creative beyound the limits of Unicode

[05:11:03.0000] <Lachy>
so why doesn't Unicode just add the missing characters?

[05:12:00.0000] <hsivonen>
Lachy: you can make up Kanji

[05:12:01.0000] <Lachy>
wtf?

[05:12:02.0000] <hsivonen>
Lachy: so any the whole concept of a coded character set is culturally insensitive

[05:12:03.0000] <MikeSmith>
hsivonen: it's not their personal names, but the "spelling" of the family name. i.e. names that have been in use and written that way for many generations

[05:13:00.0000] <Lachy>
hsivonen, without some sort of coded character set, how is it possible to represent Kanji in any encoding?

[05:13:01.0000] <hsivonen>
MikeSmith: so how does the government store the names digitally?

[05:13:02.0000] <Lachy>
they could do it as bitmap images

[05:13:03.0000] <hsivonen>
Lachy: you'd have to give bit representations to strokes

[05:13:04.0000] <MikeSmith>
hsivonen: I dunno the details. I should read up on it more and find out

[05:14:00.0000] <hsivonen>
Lachy: as opposed to giving bits to a block of Kanji

[05:14:01.0000] <MikeSmith>
I'm sure myakura knows better than I do

[05:14:02.0000] <Lachy>
ok. So in a sense, each stroke would be somewhat like a combining char

[05:14:03.0000] <Philip`>
Unicode could include characters like COMBINING BLACK PIXEL AT 2,7 and you could build bitmaps out of that

[05:16:00.0000] <hsivonen>
/me is quite happy to use a foreign an alphabet that was not invented in Finland and whose rendering has been adjusted to technology many times

[05:17:00.0000] <myakura>
/me doesn't really know the details re the name-storing :(

[05:18:00.0000] <Dashiva>
You have Mojikyo, which stores glyphs rather than code points (as I understand it)

[05:19:00.0000] <hsivonen>
Dashiva: as vector graphics?

[05:20:00.0000] <Dashiva>
Don't know what type of font it is

[05:20:01.0000] <hsivonen>
how does one do a vanity search on Google for a name that's not "text" as far as computers are concerned?

[05:28:00.0000] <hsivonen>
/me wonders if by "spelling" MikeSmith meant the kind of glyph variation that Unicode has deemed as being a stylistic font thing and not part of the essence of the abstract character

[05:30:00.0000] <Dashiva>
I found this page helpful: http://www.jbrowse.com/text/unij.html

[05:32:00.0000] <Dashiva>
Last section of "personal names" seems relevant

[05:33:00.0000] <hsivonen>
For the last 5 years, I've been wondering what the character reportoire and the length limits for legal names in Finland are, but I haven't been able to find out

[05:34:00.0000] <hsivonen>
however, I gather that in Finland, the legal name doesn't encode case and you are just assumed to capitalize the first letter of each token

[05:34:01.0000] <hsivonen>
the number of tokens is limited by law

[05:36:00.0000] <hsivonen>
also, I've so far discovered that a person who gets a Finnish citizenship and whose previous official name was in Russian gets the Latin name algorithmically without discretion

[05:43:00.0000] <Lachy>
"In 1969, the JIS C 6220 standard was created, [...] The 'C' in the name refers to electronics; the standard was later renamed to JIS X 0201, where 'X' refers to information technology." - I wonder how anyone came up with using a C to stand for "electronics" and an X for IT?!

[05:44:00.0000] <Lachy>
unless C really refers to computers, not electronics

[05:45:00.0000] <Philip`>
Lachy: Maybe electronics was the third category they chose to deal with

[05:47:00.0000] <MikeSmith>
btw, the problem characters in Japanese are called "gaiji"

[05:53:00.0000] <Dashiva>
Oh, so beer break isn't good enough for him anymore

[05:53:01.0000] <Lachy>
that article gives me the impression that the problems were created by the Japanese own inability to create a single coherent character set, and that their anger towards Unicode is misguided, since Unicode was forced to inherit the problems of pre-existing Japanese encodings

[05:54:00.0000] <Dashiva>
That doesn't cover han unification 

[05:56:00.0000] <hsivonen>
Dashiva: as far as Han Unification goes, Japan has the most market power of the countries affected, so why is using Japanese-oriented fonts an issue?

[05:57:00.0000] <hsivonen>
an issue in Japan that is

[05:59:00.0000] <hsivonen>
I mean, it doesn't bother my digital typography that some retro newspapers in Germany like to use fraktur fonts for our shared character reportoire

[05:59:01.0000] <Dashiva>
hsivonen: And how many years until China overtakes Japan?

[06:01:00.0000] <hsivonen>
Dashiva: is the crux of the matter really a pro-active fear of Chinese typography spilling over as the PRC's economic status changes?

[06:01:01.0000] <Dashiva>
It also introduces opportunities for hilarity (or not) where a 3rd party who doesn't speak the language uses the wrong font on an otherwise okay document

[06:04:00.0000] <hsivonen>
I guess my views on these issues are affected by my native language always having used imported fonts and the choice of my name being motivated by compatibility rather than by unusualness

[06:22:00.0000] <annevk3>
how do I make 8.04 update to 8.10?

[06:24:00.0000] <Philip`>
annevk3: Have you tried http://www.ubuntu.com/getubuntu/upgrading ?

[06:28:00.0000] <takkaria>
update-manager -d I think

[06:30:00.0000] <annevk3>
thanks, didn't know 8.04 changed my settings

[06:30:01.0000] <Lachy>
Hixie, gsnedders, there seems to be a bug with anolis.  Section "4.10.4.1.2 Text state and Search state" is listed in the TOC as "4.10.4.1.2 Text state and"

[06:30:02.0000] <takkaria>
/me is finding 8.10 buggy as hell, though, so doesnt' recommend upgrading

[06:30:03.0000] <annevk3>
ok, cancelled

[06:33:00.0000] <mpt>
takkaria, what in particular?

[06:35:00.0000] <Philip`>
/me finds it much easier to let old servers succumb to entropy than to upgrade them

[06:36:00.0000] <Philip`>
(I have a Solaris box which I don't think has been updated in about six years, since I don't even know how to update software on it)

[06:36:01.0000] <Philip`>
(and if I did update anything, I don't have any way to test that it hasn't broken lots of things that run on the server :-( )

[06:37:00.0000] <Dashiva>
And that's why we still have so many apache 1.3 installations too, I bet :)

[06:39:00.0000] <Philip`>
"Apache/1.3.28"

[06:41:00.0000] <zcorpan__>
hsivonen: you could make the <input> attribute list more useful by scraping the type states in the spec and removing attributes from the list that don't apply to the state

[06:41:01.0000] <zcorpan>
hsivonen: in the bookkeeping div

[06:43:00.0000] <zcorpan>
hsivonen: first paragraph of each such div, links whose href starts with #attr-

[06:44:00.0000] <zcorpan>
hsivonen: and then you'd say, "Element-specific attributes for element input in the Image state:"

[06:44:01.0000] <zcorpan>
/me files a bug

[06:50:00.0000] <zcorpan>
http://bugzilla.validator.nu/show_bug.cgi?id=339

[06:50:01.0000] <takkaria>
mpt: wireless works intermittently, NetworkManager plays up a fair bit, and since the upgrade, I tend to get "No protocol specified" messages to the console when starting X apps

[06:51:00.0000] <mpt>
takkaria, wireless problems are highly driver-dependent, so problems you have will be unrelated to any annevk3 has unless you use the same wireless card

[06:56:00.0000] <Philip`>
mpt: But driver-dependent problems still demonstrate that upgrades can break things that used to work, so if you have something that currently works then it's probably safer to not upgrade

[07:00:00.0000] <mpt>
That conclusion may be correct, but the logic is not. :-)

[07:07:00.0000] <Philip`>
mpt: The logic is that takkaria's evidence proves 8.10 is not strictly better than 8.04 in all cases and that in some cases it is worse, and annevk3 doesn't know in advance which of the cases will affect him, so there is a non-zero risk that upgrading will make things worse, so it is perfectly logical to recommend against upgrading given the evidence :-)

[07:08:00.0000] <mpt>
But if that was true no-one with a working system would ever upgrade

[07:10:00.0000] <annevk3>
s/would/should/ :)

[07:11:00.0000] <annevk3>
takkaria, do you have a lenovo/thinkpad?

[07:12:00.0000] <takkaria>
no, I'm using an asus

[07:14:00.0000] <wilhelm>
I haven't upgraded my Thinkpad yet either.

[07:16:00.0000] <Philip`>
mpt: People would still upgrade their perfectly-working systems to an entirely unknown new version, because people are irrational :-)

[07:17:00.0000] <mpt>
We have this live CD thingy, so you can try out the new version and make sure it works.

[07:17:01.0000] <Philip`>
/me likes Gentoo because upgrades are continuous and so they tend to break one thing at a time

[07:18:00.0000] <Philip`>
(I suppose it's partly a downside that they tend to break one thing, but it's an upside that they tend to not break multiple independent things)

[07:19:00.0000] <Philip`>
(Anyway, mostly I just like the colours in Gentoo's package manager)

[07:20:00.0000] <wilhelm>
Do anyone here know about any statistics on how people spend their time online?

[07:21:00.0000] <wilhelm>
(How time is distributed between various types of sites.)

[07:22:00.0000] <zcorpan>
wilhelm: do you actually just want to know how much porn people watch?

[07:22:01.0000] <mpt>
wilhelm, <http://www.pewinternet.org/reports.asp> might have something in that vein

[07:22:02.0000] <wilhelm>
zcorpan: Yes. “All sites must work”. (c;

[07:23:00.0000] <wilhelm>
mpt: Thanks.

[07:23:01.0000] <erlehmann>
Porn is the driving force of media technology - so unless we have creative commons licensed porn in <video> elements, standards aren't there

[07:23:02.0000] <mpt>
particularly <http://www.pewinternet.org/PPF/r/213/report_display.asp>

[07:24:00.0000] <annevk3>
I met someone today at a university in Germany who has studied user behavior on the Web: http://www.l3s.de/~herder/thesis/index.php

[07:25:00.0000] <wilhelm>
Thanks.

[07:27:00.0000] <Philip`>
erlehmann: If that was the case, how come Youtube is the most popular video sharing site?

[07:29:00.0000] <Philip`>
Woah, Alexa has a downloadable list of its top million sites

[07:29:01.0000] <erlehmann>
Philip`: "youtube" isn't a technology ?

[07:30:00.0000] <Philip`>
/me downloads the list

[07:30:01.0000] <Philip`>
That could be quite useful for people surveying popular sites

[07:32:00.0000] <Dashiva>
I wonder how much of the list is significant, with regard to how many people use alexa

[07:32:01.0000] <Philip`>
erlehmann: It's an application of a particular media technology, and seems to indicate that web-based Flash video is being driven by family-friendly items like music videos and illegal TV clips and cats

[07:33:00.0000] <Philip`>
Dashiva: What do you mean by "significant"?

[07:34:00.0000] <Dashiva>
Philip`: Statistically

[07:34:01.0000] <Philip`>
(Alexa says "Generally, traffic rankings of 100,000 and above should be regarded as not reliable")

[07:34:02.0000] <Dashiva>
Ah, so they cover that

[07:35:00.0000] <Philip`>
Dashiva: What do you mean by "statistically significant"? :-)

[07:36:00.0000] <Philip`>
Nobody would assume they mean "there is a 95% statistical chance that tailedfox.com is ranked 12345", because that'd be silly

[07:37:00.0000] <Philip`>
or that the site ranked 12345 is more popular than the site ranked 12346

[07:37:01.0000] <Philip`>
so I guess the question is what you want to measure and to determine the significance of

[07:38:00.0000] <Dashiva>
But you could say something like "there's a 95% chance that tailedfox.com is within ranks [6172, 24690]"

[07:39:00.0000] <Philip`>
That would be useful to know

[07:39:01.0000] <annevk3>
http://twitter.com/banderson623/statuses/1022729083

[07:40:00.0000] <Philip`>
/me appreciates how the data is in CSV format, like "95012,whatwg.org", instead of dmoz's RDF

[07:40:01.0000] <Philip`>
1844,w3.org

[07:40:02.0000] <Philip`>
1845,biblegateway.com

[07:40:03.0000] <Philip`>
The W3C is more popular than the bible!

[07:40:04.0000] <Philip`>
but less popular than a Zimbabwean bank

[07:41:00.0000] <Philip`>
Oh, sorry, I'm being ignorant

[07:41:01.0000] <Philip`>
The W3C is less popular than a South African bank

[07:44:00.0000] <jgraham>
annevk3: A HTML5 book would be great

[07:44:01.0000] <zcorpan>
hey jgraham

[07:44:02.0000] <jgraham>
hey

[07:44:03.0000] <zcorpan>
you're starting on monday right?

[07:44:04.0000] <jgraham>
Right

[07:45:00.0000] <jgraham>
/me just finished his thesis viva :)

[07:45:01.0000] <wilhelm>
Welcome aboard. (c;

[07:45:02.0000] <zcorpan>
you'll probably sit next to me and i'll be your contact person or whatever it's called

[07:45:03.0000] <jgraham>
awesome

[07:45:04.0000] <jgraham>
wilhelm: Thanks :)

[07:46:00.0000] <zcorpan>
i'm supposed to give you things to do but i have no idea -- maybe you can read my email :)

[07:46:01.0000] <Philip`>
Is Opera engorging on more WHATWG people?

[07:47:00.0000] <annevk3>
we'd happily accept you :)

[07:48:00.0000] <Philip`>
I already have plenty to work on for the next two years :-)

[07:48:01.0000] <jgraham>
/me often wonders how Philip` has any time to do his PhD

[07:49:00.0000] <gsnedders>
/me still needs to apply for work next summer

[07:49:01.0000] <Philip`>
jgraham: I thought the idea was that you could be lazy and not do any work for the first year

[07:49:02.0000] <jgraham>
Philip`: Don't ask me, I'm not really sure that I achieved anything in four years, but my examiners seemed to like it

[07:50:00.0000] <gsnedders>
w00t!

[07:50:01.0000] <gsnedders>
Unconditional for Edinburgh

[07:50:02.0000] <Philip`>
The government keeps giving me money regardless of what I spend my time doing

[07:50:03.0000] <jgraham>
gsnedders: Cool

[07:51:00.0000] <gsnedders>
There again, the course I'm applying for has the normal Scottish entrance requirements (BBBB at Higher, which historically was the requirements for pretty much every course in Scotland), and has very few applicants 

[07:51:01.0000] <gsnedders>
Normally all applicants got offers, and I exceed the entrance requirements, so there was little question.

[07:51:02.0000] <jgraham>
gsnedders: Is this computational physics or something?

[07:52:00.0000] <gsnedders>
jgraham: yeah

[07:53:00.0000] <gsnedders>
http://www.sra.ed.ac.uk/admissions/Statistics/2007-08_entry/Programme2007/degree_programme_public_version/37.htm

[07:54:00.0000] <Philip`>
Isn't the LHC going to solve physics, so you wouldn't have anything to work on after graduating?

[07:54:01.0000] <gsnedders>
Philip`: No, it's going to disprove physics.

[07:54:02.0000] <jgraham>
Philip`: If only

[07:54:03.0000] <Philip`>
I suppose you could sabotage the LHC

[07:54:04.0000] <Philip`>
or maybe you already did

[07:55:00.0000] <gsnedders>
(1 to 5 applicants for the BSc course, 9 applicants for the MPhys course; same number of offers, 1 to 5 people start each course)

[07:58:00.0000] <takkaria>
/me is applying for a summer internship at Opera

[07:59:00.0000] <Philip`>
You're not doing much to dispel the WHATWG browser conspiracy notions :-(

[07:59:01.0000] <annevk3>
takkaria, cool, ping one of us if your application got stuck somehow

[07:59:02.0000] <wilhelm>
takkaria: Cool. For what kind of position? Developer or QA?

[08:00:00.0000] <Philip`>
/me wonders if Opera pays lots of money compared to other places he could look for jobs :-)

[08:00:01.0000] <jgraham>
/me wonders if he is supposed to unsubscribe from public-html or something

[08:00:02.0000] <takkaria>
annevk3: I've not written the covering letter yet, but will be putting it all through next week sometime

[08:00:03.0000] <annevk3>
kk

[08:01:00.0000] <gsnedders>
/me ought to apply for a summer internship at Opera too

[08:01:01.0000] <annevk3>
jgraham, you should e-mail chaals⊙oc and arrange it with him

[08:01:02.0000] <gsnedders>
even though technically he won't be enrolled in a uni course till mid-August

[08:01:03.0000] <annevk3>
jgraham, just wait for him to get back to you

[08:01:04.0000] <takkaria>
wilhelm: I'd be happy doing QA, though development work out be fun

[08:01:05.0000] <gsnedders>
/me doesn't have the 1337 hacking skills to do development

[08:02:00.0000] <wilhelm>
gsnedders: Doesn't matter. I didn't pass any of the formal requirements.

[08:02:01.0000] <gsnedders>
wilhelm: Including the English bit?

[08:02:02.0000] <gsnedders>
/me still needs to finish writing CV

[08:02:03.0000] <hsivonen>
jgraham: congrats on the new job. Where will you be located?

[08:02:04.0000] <takkaria>
jgraham: what was your thesis on?

[08:03:00.0000] <wilhelm>
gsnedders: No. But I have no education, so HR discarded my application. “Unqualified”.

[08:03:01.0000] <jgraham>
hsivonen: Thanks. I'm in Linkoping with zcorpan 

[08:03:02.0000] <gsnedders>
wilhelm: Then how did you get the place?

[08:03:03.0000] <wilhelm>
The hiring procedures have, uhm, changed since then. (c:

[08:03:04.0000] <jgraham>
takkaria: "Heating and Transport Processes in the Intracluster Medium" is what it says on the cover

[08:04:00.0000] <Philip`>
jgraham: Sounds like fun

[08:04:01.0000] <gsnedders>
/me expects that has helped takkaria less than he was expecting

[08:04:02.0000] <takkaria>
actually, I was expecting something entirely opaque, since all the thesis titles I've ever heard are pretty opaque to me :)

[08:05:00.0000] <wilhelm>
gsnedders: My application was recommended by an employee, and my manager followed up on in despite HR's decision. Now I've been here for four years.

[08:06:00.0000] <gsnedders>
hmm. <http://www.opera.com/company/jobs/list/?dept=summer_interns&location=all> and <http://www.opera.com/company/jobs/internship/> are different

[08:06:01.0000] <gsnedders>
ah, the latter one seems to be for summer '08 on the face of it

[08:07:00.0000] <wilhelm>
Indeed. They might have messed something up during the recent upgrade.

[08:07:01.0000] <gsnedders>
http://www.opera.com/company/jobs/opening/211/ is what I want?

[08:08:00.0000] <hsivonen>
/me is a bit concerned that googling for mutation event test cases doesn't find a test suite from the usual suspects on the first page of results

[08:08:01.0000] <wilhelm>
gsnedders: Yeah, either should work fine. They all end up in the same system.

[08:10:00.0000] <gsnedders>
/me knows he will be enrolled in _some_ uni course next year

[08:10:01.0000] <gsnedders>
I have an unconditional offer for Edinburgh, as I said, so if nothing else I should be going there :P

[08:11:00.0000] <gsnedders>
Does the transcript of grades need to be a formal copy of them?

[08:16:00.0000] <jgraham>
gsnedders: Dunno

[08:18:00.0000] <gsnedders>
wilhelm: have any clue? know how I could find out?

[08:19:00.0000] <wilhelm>
I think so.

[08:19:01.0000] <gsnedders>
/me wonders how he can get a copy

[08:20:00.0000] <gsnedders>
What's an example of something which ends up with a different parent, which makes SAX impossible?

[08:24:00.0000] <Philip`>
gsnedders: You mean like <table>foo ?

[08:24:01.0000] <gsnedders>
Philip`: Ah, just stuff like that does? Nothing more complex?

[08:24:02.0000] <Philip`>
You could make it more complex if you want :-)

[08:24:03.0000] <takkaria>
hmm, I wonder my certificates are for my transcripts

[08:26:00.0000] <Philip`>
gsnedders: It doesn't make SAX impossible - it just makes it impossible to do it without buffering an unbounded number of elements before emitting anything

[08:27:00.0000] <Philip`>
gsnedders: (Also, the ability to write <html foo=bar> anywhere in the document (which will add that attribute to the root <html> element) makes it impossible to output any SAX events at all, until you've reached EOF)

[08:28:00.0000] <zcorpan>
unless you abort

[08:30:00.0000] <Philip`>
True, though I prefer the pro-life stance towards HTML parsers

[08:32:00.0000] <annevk3>
or unless you also keep a tree around and let SAX patch it

[08:33:00.0000] <Philip`>
I don't think SAX has any events for tree-patching, so either it's not SAX or it's not streaming

[08:34:00.0000] <Philip`>
(depending on which side of the SAX interface you do the tree stuff on)

[08:34:01.0000] <BenMillard>
krijnh, I'm back. :)

[08:35:00.0000] <annevk3>
Philip`, I see, fair enough

[08:38:00.0000] <krijn>
BenMillard: hi :)

[08:38:01.0000] <krijn>
Your stylesheet links to http://ben.local/!dev/irc-logs/screen.css

[08:38:02.0000] <BenMillard>
krijn, you saw my ideas then?

[08:38:03.0000] <krijn>
Yeah, neat

[08:38:04.0000] <BenMillard>
d'oh, let me fix that

[08:39:00.0000] <BenMillard>
(it's because I have a <base>)

[08:39:01.0000] <krijn>
Yeah, saw that

[08:41:00.0000] <BenMillard>
stylesheet <link> fixed

[08:42:00.0000] <krijn>
It needs some rounded corners :)

[08:42:01.0000] <BenMillard>
hah, maybe!

[08:43:00.0000] <BenMillard>
keeping it non-graphical is more faithful to the service you currently provide, although you can take as much or as little from this idea as you like

[08:43:01.0000] <BenMillard>
I tried applying the WHATWG spec colours to it

[08:44:00.0000] <krijn>
Table + Link is your favourite?

[08:44:01.0000] <krijn>
Which one is that? :)

[08:45:00.0000] <krijn>
Ah, whatwg-table-link

[08:45:01.0000] <BenMillard>
you can hover over the channel name at the top of each colour to get a title tooltip

[08:46:00.0000] <BenMillard>
I chose 8 days for the table so you can see "this time last week"

[08:46:01.0000] <krijn>
Easy for the Last week in HTML author :)

[08:47:00.0000] <BenMillard>
yeah, it seemed like a reasonable use-case :P

[08:47:01.0000] <krijn>
We should ask Hixie for approval first, of course :)

[08:48:00.0000] <krijn>
/me tries to get mentioned on Last Week

[08:48:01.0000] <krijn>
Anyways

[08:48:02.0000] <BenMillard>
heh, there's no mad rush...we could let people review it and suggest further tweaks and so forth

[08:48:03.0000] <krijn>
I'll update it tomorrow

[08:49:00.0000] <BenMillard>
wow!

[08:49:01.0000] <BenMillard>
another change, I moved the search box to the top-right as that's common on big sites

[08:49:02.0000] <Philip`>
I don't like the table having left-aligned headings for right-aligned columns

[08:49:03.0000] <krijn>
Yeh

[08:49:04.0000] <krijn>
Philip`: try a user stylesheet :+

[08:50:00.0000] <BenMillard>
I'll change it, see what peeps think

[08:50:01.0000] <krijn>
Also, I'm getting faster internet in a month or something, so the logs will be a bit snappier

[08:50:02.0000] <jgraham>
BenMillard: BTW, I will change the table inspector UI at some point

[08:51:00.0000] <BenMillard>
jgraham, cool!

[08:51:01.0000] <BenMillard>
krijn, that's a point: by having a single link to the archives instead of the long list, maybe it would be a bit faster to generate and the markup would be lighter weight

[08:51:02.0000] <BenMillard>
I know the table for 8 days is less markup than the current list for 5 days, due to the repetitive title attributes being replaced by <th>

[08:51:03.0000] <jgraham>
If you have an idea for a design that demphasises the legacy algorithms that would be cool

[08:52:00.0000] <jgraham>
BenMillard: ^^^

[08:52:01.0000] <BenMillard>
jgraham, we could reverse the order of that list? that's about all I can think of :)

[08:52:02.0000] <krijn>
BenMillard: hiding the referrers also speeds it upt

[08:52:03.0000] <krijn>
*up

[08:52:04.0000] <Philip`>
krijn: Maybe you could host (or mirror or something) the files on somewhere like html5.org, if internet speed is an issue?

[08:52:05.0000] <krijn>
It's no issue for me

[08:53:00.0000] <krijn>
I hardly have time to read the logs myself :(

[08:53:01.0000] <BenMillard>
krijn, yeah I couldn't think why the referrers would be useful to anyone so I just scrapped them :P

[08:53:02.0000] <Philip`>
I like the referrers, they're useful for spying on people

[08:54:00.0000] <BenMillard>
perhaps a single link to a page which lists them would be a good comprimise? that way, they won't slow down the front page but will still be available for weirdos like Philip`?

[08:55:00.0000] <krijn>
They do slow down the homepage anyway

[08:55:01.0000] <krijn>
I think..

[08:55:02.0000] <zcorpan>
like http://krijnhoetmer.nl/irc-logs/?gimme-a-hobby ?

[08:56:00.0000] <krijn>
Lightning fast hooray \o/

[08:57:00.0000] <zcorpan>
you should have an ajaxy button next to the tagline

[08:57:01.0000] <zcorpan>
to get the next tagline

[08:57:02.0000] <annevk3>
"While We All Agree Flash Is Still A Superior Technology ;)" :p

[08:57:03.0000] <krijn>
Perhaps it's not that smart to read 2500 filenames and sort them on each request :)

[08:57:04.0000] <Philip`>
Is there a list of all of them?

[08:57:05.0000] <krijn>
Nah

[08:58:00.0000] <krijn>
Lachy did some grepping

[08:58:01.0000] <jgraham>
BenMillard: Maybe a "Show legacy algorithms" checkbox that is deselected by default?

[08:59:00.0000] <jgraham>
/me is maybe overcopmplicating things

[08:59:01.0000] <BenMillard>
jgraham, sounds like deprecation vs. obsoletion to me :)

[08:59:02.0000] <BenMillard>
if they are good enough to keep in the project, I say you might as well keep listing them normally

[08:59:03.0000] <BenMillard>
(but it's your tool, you can do what you like with it)

[09:01:00.0000] <zcorpan>
jgraham: you could use the preferred algorithm by default and hide all the others under a "more options"

[09:01:01.0000] <BenMillard>
jgraham & zcorpan, here's the UI with the preferred algorithms listed first: http://projectcerbera.com/!dev/table-inspector/new

[09:02:00.0000] <jgraham>
zcorpan: The problem is that youw want HTML5 (Nov 2008) and Smartheaders to be easilly accessible whereas HTML 5 (Sept 2007) is just a distraction

[09:02:01.0000] <BenMillard>
I guess a cheap trick would be to set size="2" so only HTML5 and Smart Headers are shown without scrolling?

[09:02:02.0000] <jgraham>
BenMillard: Cheap tricks are, at least, cheap

[09:02:03.0000] <BenMillard>
OK, take a look now

[09:03:00.0000] <BenMillard>
this seems quite a neat way to do it, actually

[09:03:01.0000] <zcorpan>
why do you want to hide the others?

[09:04:00.0000] <BenMillard>
the older ones aren't really contenders in the what-gets-into-HTML5 race

[09:04:01.0000] <jgraham>
zcorpan: How useful is it to know what the old broken version of the HTML5 spec did?

[09:04:02.0000] <zcorpan>
jgraham: so remove it from the UI altogether

[09:04:03.0000] <jgraham>
zcorpan: Maybe that is the best idea

[09:05:00.0000] <zcorpan>
size=2 just gives a worse overview of which algorithms are available

[09:05:01.0000] <BenMillard>
I thought the older algorithms may be of interest to accessibility professionals, particularly HTML4 and "experimental", to compare them to the modern ones

[09:05:02.0000] <BenMillard>
but as it turns out, next to nobody has done testing with the inspector

[09:06:00.0000] <BenMillard>
I'll jgraham decide what goes into the real interface

[09:06:01.0000] <BenMillard>
s/I'll/I'll let/

[09:06:02.0000] <zcorpan>
BenMillard: accessibility professionals don't need to test things

[09:07:00.0000] <BenMillard>
zcorpan, it would be nice if they do so before giving feedback about how HTML5 is ruining accessibility :)

[09:07:01.0000] <zcorpan>
BenMillard: i agree

[09:09:00.0000] <BenMillard>
Philip`, table headers are all centred now: http://projectcerbera.com/!dev/irc-logs/front

[09:10:00.0000] <BenMillard>
Krijn, that demo also shows what I meant by "a link to a page of referrers" in the footer

[09:10:01.0000] <Philip`>
Why can't I use SVG stroke="rgba(...)" in Opera (9.6) or Safari (3.0)?

[09:10:02.0000] <BenMillard>
zcorpan, do you have an opinion on that IRC Logs demo? ^^^

[09:12:00.0000] <BenMillard>
I think month view is overkill for a front page, but could work in yearly and monthly archive pages

[09:12:01.0000] <zcorpan>
BenMillard: the calendars don't contain the same information

[09:12:02.0000] <zcorpan>
i think the left most works fine, except instead with an archives link

[09:13:00.0000] <zcorpan>
but i don't really care :)

[09:14:00.0000] <BenMillard>
zcorpan, yeah the calendars don't have number of lines or number of important lines...which is a bit lame

[09:14:01.0000] <zcorpan>
the logs sometimes forget who i am, which is probably what is most annoying for me

[09:14:02.0000] <BenMillard>
yeah, I get that if I don't use them for like a week or so

[09:15:00.0000] <zcorpan>
krijn: is it because of the expiration date of the cookies? 

[09:24:00.0000] <Philip`>
http://philip.html5.org/misc/spec-links-anim.svgz

[09:25:00.0000] <BenMillard>
jgraham, attempted WHATWG colouring and made <dl> order match <select> order: http://projectcerbera.com/!dev/table-inspector/new

[09:26:00.0000] <jgraham>
BenMillard: Dunno about the green :)

[09:27:00.0000] <jgraham>
Philip`: That is cool. Am I supposed to think anyhting more complex than that?

[09:28:00.0000] <Philip`>
jgraham: No - it doesn't have any actual point

[09:28:01.0000] <Philip`>
although it does provide the shocking insight that the spec has got more complex over time

[09:32:00.0000] <Philip`>
(Also, Opera seems slow at SVG)

[09:44:00.0000] <gsnedders>
/me thinks he ought to send off accommodation/meal request form for Cambridge interview

[10:02:00.0000] <gsnedders>
Damn you browsers apart from Chrome and IE8!

[10:04:00.0000] <Philip`>
(?)

[10:06:00.0000] <gsnedders>
one tab bringing down the whole browser

[10:53:00.0000] <csarven>
So, I'm trying to put together a "sidebar" or "supplemental_information" area and I was going to put aside and section elements together. Would something like this be accurate: <aside><section></section> <section></section> ...</aside>

[11:08:00.0000] <krijn>
BenMillard: "a link to a page of referrers" - thanks, now I know what a link looks like :)

[11:19:00.0000] <zcorpan>
csarven: looks reasonable

[11:36:00.0000] <zcorpan>
http://www.sitepoint.com/forums/showthread.php?t=585619

[11:40:00.0000] <hsivonen>
Philip`: the spec ref animation is very cool. thanks

[11:41:00.0000] <hsivonen>
what use case were mutation events introduced for in the first place?

[11:42:00.0000] <hsivonen>
notifying various internal parts of the browser makes sense, but those parts can promise not to mutate the DOM themselves

[11:42:01.0000] <hsivonen>
however, letting scripts mutate the DOM on mutation events seems like asking for trouble

[11:43:00.0000] <gavin>
I think the main use case was "browser developers don't have enough headaches"

[11:45:00.0000] <hsivonen>
perhaps there wasn't enough headache back then

[11:45:01.0000] <Philip`>
Maybe you want to implement SVG in pure JS, and you need some way to detect when the DOM has changed so that you can update all your internal data structures and draw everything

[11:45:02.0000] <Philip`>
(though you don't need synchronous mutation events for that)

[11:46:00.0000] <Philip`>
(so there are probably better solutions in that case)

[11:50:00.0000] <hsivonen>
can XBL execute script in the context of the document it binds to as a response of a document mutation?

[14:32:00.0000] <zcorpan_>
should DOMException be in module dom?

[14:32:01.0000] <annevk3>
hsivonen, regarding the parsing of noXXX elements, have you considered that the tree builder code might be relevant in case the feature is disabled, e.g. when plugins are disabled <noembed> is parsed differently, the same might go for frames if Gecko still offers that feature

[14:33:00.0000] <hsivonen>
annevk3: if plug-ins are disabled, shouldn't the form controls in noembed work?

[14:34:00.0000] <hsivonen>
I wonder if disabling plug-ins actually changes noembed parsing

[14:35:00.0000] <zcorpan_>
it does in opera, or at least did, not sure if we changed that

[14:35:01.0000] <annevk3>
oh, apologies, I guess I didn't read the logs carefully enough to consider what Gecko is actually doing in those scenarios

[14:40:00.0000] <heycam>
zcorpan_, it should be (if only so that the java package is in org.w3c.dom)

[14:40:01.0000] <heycam>
er, s/package/class/

[14:41:00.0000] <heycam>
zcorpan_, will you be making NodeLists callable?

[14:43:00.0000] <zcorpan_>
heycam: callable?

[14:45:00.0000] <annevk3>
zcorpan_, list(x) syntax rather than list[x]

[14:45:01.0000] <heycam>
yes that

[14:45:02.0000] <heycam>
i wasn't sure whether that's something people agreed on needing to specify

[14:46:00.0000] <heycam>
i added [Callable] to web idl that could be used to do that, if needed

[14:50:00.0000] <zcorpan_>
do browsers support that on NodeLists?

[14:52:00.0000] <annevk3>
I thought some did, at least

[14:53:00.0000] <Lachy>
re the talk about design principles in http://lists.w3.org/Archives/Public/public-html/2008Nov/0576.html - is anyone actually planning to continue editing that document?

[14:53:01.0000] <Lachy>
it hasn't been touched for a year now, since the FPWD was published

[15:02:00.0000] <zcorpan_>
so ie, webkit and opera supports it, but not firefox

[15:02:01.0000] <zcorpan_>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Cscript%3Ew(document.getElementsByTagName('html')(0))%3C%2Fscript%3E

[15:08:00.0000] <zcorpan_>
heycam: added, thanks

[15:11:00.0000] <zcorpan_>
ie does it for NamedNodeMap as well but webkit and opera not

[15:30:00.0000] <Philip`>
Is list('length') the same as list['length'] and list.length?

[15:30:01.0000] <Philip`>
Uh, maybe s/list/map/ or whatever this applies to

[15:30:02.0000] <heycam>
list['length'] is (in most impls)

[15:30:03.0000] <heycam>
list('length') i haven't tested

[15:31:00.0000] <Dashiva>
I figured () would be item lookup, and length isn't an item

[15:34:00.0000] <zcorpan_>
what Dashiva said

[15:34:01.0000] <Philip`>
Hmm, IE6 says document.all.length is some integer and document.all('length') is null, unless there's an element with name=length in which case they're both equal to that element

[15:35:00.0000] <heycam>
some collection class testing: http://mcc.id.au/temp/2008/webidl-tests/htmlcollection-behaviour.html

[15:35:01.0000] <zcorpan_>
is document.all a NodeList?

[15:35:02.0000] <heycam>
so that's for HTMLCollection, not NodeList

[15:36:00.0000] <heycam>
and the tests themselves: http://mcc.id.au/temp/2008/webidl-tests/htmlcollection.html

[15:36:01.0000] <zcorpan_>
heycam: nice

[15:37:00.0000] <zcorpan_>
/me would like a column for what's correct spec-wise :)

[15:37:01.0000] <heycam>
heh

[15:38:00.0000] <heycam>
i used that to decide the behaviour for [IndexGetter] etc.

[15:38:01.0000] <zcorpan_>
"Index property is on the collection object" -- clearly opera is the odd one out there

[15:38:02.0000] <heycam>
i think what's required by the spec is the majority for each row

[15:39:00.0000] <zcorpan_>
ok

[15:40:00.0000] <Hixie>
i can add a <!-- comment --> of some sort in place of the missing paragraphs if that would help

[15:40:01.0000] <zcorpan_>
heycam: so then it's biased towards webkit?

[15:40:02.0000] <heycam>
well the behaviour in chrome and webkit is different sometimes

[15:41:00.0000] <heycam>
so i'm not sure how much of a bias there is

[15:41:01.0000] <heycam>
but probably there is some

[15:41:02.0000] <zcorpan_>
Hixie: maybe but the real problem was probably me trying to use my brain instead of going to bed

[15:42:00.0000] <Hixie>
heh

[15:43:00.0000] <zcorpan_>
which reminds me

[15:44:00.0000] <zcorpan_>
nn

[15:44:01.0000] <Hixie>
nn


2008-11-28
[18:24:00.0000] <Lachy>
/me updates HTML authoring guide. Now that I'm using anolis to generate it, I've begun converting the document itself to use HTML5

[19:33:00.0000] <Lachy>
http://www.themaninblue.com/experiment/JS-909/

[01:33:00.0000] <Hish>
Hixie: ping

[01:38:00.0000] <Hixie>
Hish: vaguely here

[01:41:00.0000] <Hish>
short question regarding representing a spreadsheet in html5: currently I display a cell like this: <td><div>(html content)</div></td>. Regarding formula cells I consider doing something like this: <td><output>(html content)</output></td>. Does this make sense?

[01:42:00.0000] <Hixie>
why the <div>?

[01:42:01.0000] <Hixie>
<output> makes sense there if you are serialising the spreadsheet, sure

[01:43:00.0000] <Hixie>
you could even use <input> everywhere else and use onforminput="" to do form value recalculation

[01:43:01.0000] <Hixie>
in html5

[01:44:00.0000] <Hish>
the div because I want to be able to rotate cell content using CSS SVG transform.

[01:44:01.0000] <Hish>
for both, inout and output I need to display rich text html content. so value attributes don't work here

[01:45:00.0000] <annevk3>
<td contenteditable>

[01:46:00.0000] <Hixie>
k

[01:46:01.0000] <Hish>
annevk3: sure. that's what I use instead of input or textarea elements.

[01:47:00.0000] <hsivonen>
Hish: <output> makes sense if its value is computed from from controls and this is declared for future AT or is a plain string set using .value. otherwise, there's no point in using it

[01:49:00.0000] <Hixie>
or if you want to style calculated fields specially

[01:49:01.0000] <Hish>
hsivonen: the nice part about output is the "for" tag which allows me to track dependencies very easily. The only missing part is a definition for ranges. I try to avoid listing all cells ID rather than just list a start and end offset within the "for" tag

[01:49:02.0000] <Hixie>
but that's about it

[01:52:00.0000] <Hish>
ok. seems that I'll try the output approach. thx for your help.

[01:59:00.0000] <hsivonen>
I now have a build of Minefield with an HTML5 parser that works on some pages but fails to start layout if there's document.write

[02:19:00.0000] <hsivonen>
is the initial containing block for HTML documents 'html' or 'body'?

[02:23:00.0000] <Lachy_>
/me wonders if Thanksgiving is the reason for the sudden drop in traffic on public-html. If so, I'm thankful for that :-)

[02:24:00.0000] <krijn>
zcorpan: would setting the cookie again on each request solve that?

[02:24:01.0000] <hsivonen>
if one calls document.open() on a document hosted by an iframe, should the open() call zap the children of the document or should the first document.write() zap them?

[02:27:00.0000] <hsivonen>
hmm. the spec says open() zaps the children

[02:32:00.0000] <hsivonen>
Gecko's document.open impl doesn't conform

[02:36:00.0000] <ap>
hsivonen: https://bugzilla.mozilla.org/show_bug.cgi?id=325352 ?

[02:38:00.0000] <hsivonen>
ap: yes. thanks.

[02:40:00.0000] <hsivonen>
the comments in the source say that the old root element is put back there in order to avoid confusing scrollbars

[02:49:00.0000] <krijn>
krijn: Test

[03:01:00.0000] <zcorpan>
krijn: dunno

[03:04:00.0000] <zcorpan>
krijn: do you use whois for the logs? or a hard-coded list of names?

[03:04:01.0000] <krijn>
For the nicknames you mean?

[03:04:02.0000] <krijn>
A hard-coded list

[03:04:03.0000] <zcorpan>
what's up with the spacing?

[03:04:04.0000] <krijn>
I've only got 71 nicknames listed

[03:05:00.0000] <krijn>
Spacing?

[03:05:01.0000] <zcorpan>
each line seems to have 1em bottom padding or so

[03:05:02.0000] <krijn>
Not anymore

[03:05:03.0000] <krijn>
When I take of the </li>'s

[03:05:04.0000] <zcorpan>
ah

[03:06:00.0000] <krijn>
Probably because of white-space: pre-wrap

[03:09:00.0000] <zcorpan>
krijn: what regexp do you use for linkification?

[03:10:00.0000] <krijn>
You don't want to know :)

[03:19:00.0000] <zcorpan>
actually i do, that's why i asked :)

[03:26:00.0000] <pergj>
Is there any spec that defines the allowed range of an integer length? CSS 2.1 does not appear to do that.

[03:26:01.0000] <pergj>
I guess a signed 32 bit value should be sufficient?

[03:27:00.0000] <jmb>
I believe that's implementation defined

[03:28:00.0000] <pergj>
problem is that some pages set position for instance to -200000 to be absolutely sure that something does not appear on the screen.

[03:28:01.0000] <pergj>
So if you have only 16 bits.........

[03:28:02.0000] <jmb>
uhuh

[03:28:03.0000] <jmb>
so clamp that to INT_MIN

[03:29:00.0000] <pergj>
would work most of the time. But what if a script does fancy arithmetics to get it back again. 

[03:29:01.0000] <krijn>
zcorpan: test

[03:29:02.0000] <krijn>
http://krijnhoetmer.nl/irc-logs/ - new feature! :)

[03:29:03.0000] <jmb>
pergj: good question. badness, I guess :)

[03:30:00.0000] <jmb>
I'd be interested to know if anyone actually does that, though

[03:31:00.0000] <krijn>
BenMillard: hope this is handy for you

[03:32:00.0000] <zcorpan>
krijn: nice

[03:37:00.0000] <krijn>
Almost time for Google Ads now :)

[04:21:00.0000] <MikeSmith>
/me wonders of tthorsen is around and if he perhaps knows anything about CE-HTML or Web4CE

[04:24:00.0000] <tthorsen>
I'm around, but I hadn't heard of CE-HTML until now

[04:26:00.0000] <MikeSmith>
tthorsen: OK. was just asking because it relates to embedded browsers for certain kinds of devices

[04:27:00.0000] <MikeSmith>
wondering what browsers actually support it

[04:27:01.0000] <MikeSmith>
the spec's unfortunately no freely available

[04:27:02.0000] <MikeSmith>
have to pay 210 USD for the privilege of reading it

[04:28:00.0000] <tthorsen>
what kind of devices? If it's a real subset of normal internet standards, then I guess all browsers support it

[04:28:01.0000] <MikeSmith>
it's not a subset -- it adds new elements and new browser/scripting APIs

[04:28:02.0000] <MikeSmith>
as far as what devices, I think stuff like set-top boxes

[04:28:03.0000] <MikeSmith>
and other home electronics

[04:29:00.0000] <MikeSmith>
not so much mobile, as far as I can tell

[04:29:01.0000] <Philip`>
It sounded like it was a non-web thing, and wasn't expected to interoperate with any other content or other browsers

[04:30:00.0000] <MikeSmith>
well, some of the new markup and APIs it introduces seems to possibly overlap with parts of HTML5

[04:30:01.0000] <jmb>
MikeSmith: certainly, ANT and Oregan's markets are largely STBs, so that sounds reasonable

[04:30:02.0000] <MikeSmith>
e.g., I think video API

[04:30:03.0000] <MikeSmith>
jmb: OK

[04:30:04.0000] <MikeSmith>
I don't know much about ANT

[04:30:05.0000] <MikeSmith>
I don't know much about the set-top box market

[04:31:00.0000] <tthorsen>
so, for 210 USD I'm allowed to implement a browser for which I have to create all the content myself?

[04:31:01.0000] <MikeSmith>
heh

[04:31:02.0000] <MikeSmith>
pretty much, i guess

[04:31:03.0000] <tthorsen>
I can write up such a spec for half the price -- just let me know :-)

[04:32:00.0000] <Lachy_>
tthorsen, wow, you must be really cheap if you can write a spec for just $105 worth of work. That's barely 2 or 3 days worth on even a low salary

[04:33:00.0000] <MikeSmith>
I wasn't even aware that the Consumer Electronics Association was a standards-development org

[04:33:01.0000] <MikeSmith>
but it apparently they produce a truckload of them

[04:33:02.0000] <MikeSmith>
http://www.ce.org/Standards/2502.asp

[04:33:03.0000] <MikeSmith>
mostly hardware stuff

[04:34:00.0000] <MikeSmith>
but I guess they have decided to venture forth into producing standards for Web technologies

[04:34:01.0000] <Lachy_>
it does seem a little silly to make an HTML profile and charge money to give access to the spec, when vendors can access the full HTML spec freely, as well as content authors

[04:35:00.0000] <MikeSmith>
Lachy_: well, it's not a profile. it seems to introduce a number of new elements and browser APIs

[04:35:01.0000] <Lachy>
what documentation have you found that says it introduces new features?

[04:36:00.0000] <Philip`>
MikeSmith: They probably aren't producing standards for web technologies - they're standards for their particular markets, which happen to reuse web technologies because that gives them a solid base (of specs and browsers and tools and knowledge etc) to build on

[04:37:00.0000] <MikeSmith>
Lachy: peruseth thou that PDFs linked to from http://www.ce.org/Standards/browseByCommittee_2757.asp

[04:37:01.0000] <Lachy>
Philip`, that sounds like the kind of argument that led to the development of nonsense like WAP and the HTML mobile profile

[04:37:02.0000] <Philip`>
Lachy: http://www.ce.org/Standards/CEA-2014-A_Preview.pdf - the XHTML profile is 10 pages out of ~167

[04:37:03.0000] <Lachy>
*XHTML mobile profile

[04:37:04.0000] <tthorsen>
If they're introducing new features, the success of this spec does depend on whether someone will bother to implement these features in a browser.

[04:39:00.0000] <MikeSmith>
Philip`: I guess that some of what they're adding might be generalized, or generally useful

[04:39:01.0000] <MikeSmith>
not saying that is, just that it might be

[04:39:02.0000] <MikeSmith>
and would be disappointing if they were reinventing wheels that are better built in other places

[04:40:00.0000] <tthorsen>
Could be they have some good ideas, but if they aren't merged into html5 I bet they will all be forgotten in a few years time.

[04:40:01.0000] <Philip`>
MikeSmith: I guess it's hard to tell if we're too cheap to buy a copy of the spec and they're not willing to tell us what their features are

[04:40:02.0000] <Philip`>
s/ if/, if/

[04:43:00.0000] <Lachy>
apparently Opera is a member of CEA http://www.ce.org/Membership/Directory/default.aspx?alpha1=o

[04:43:01.0000] <MikeSmith>
I wonder if I buy a copy of the spec and write a comparison or overview of the markup and API extensions, would that count as a violation of the license...

[04:43:02.0000] <tthorsen>
I don't think this is a matter of us being cheap -- I think it's a matter of them being crazy for charging money for stuff they should be begging implementors to implement.

[04:43:03.0000] <MikeSmith>
Lachy: cool, I didn't know Opera made loudspeakers also

[04:44:00.0000] <Lachy>
what?

[04:44:01.0000] <MikeSmith>
just a joke

[04:44:02.0000] <MikeSmith>
"Opera Loudspeakers/Unison Research" listed there too

[04:45:00.0000] <Lachy>
oh

[04:45:01.0000] <MikeSmith>
notable that ACCESS isn't a member

[04:48:00.0000] <tthorsen>
/me wonders what the advantages to being a member are

[04:49:00.0000] <tthorsen>
other than being able to brag about it, of course

[04:49:01.0000] <MikeSmith>
tthorsen: you get a 20% discount off the price of anything they publish :)~

[04:49:02.0000] <tthorsen>
awesome

[04:53:00.0000] <pergj>
so, apparently, only one Browser vendor is into this market

[04:54:00.0000] <Lachy>
pergj, Microsoft is also a member

[04:55:00.0000] <Lachy>
and Apple

[04:55:01.0000] <Lachy>
so that's 3 browser vendors

[04:57:00.0000] <Lachy>
MikeSmith, you could contact CEA and ask them if there is any overlap between their work on CE-HTML and HTML5, and whether it's worth getting them to work with us on HTML5

[04:58:00.0000] <MikeSmith>
Lachy: yeah, I would like to get in touch with some of the people who have actually participated in the work, or even those who edited the spec. I'm working on it

[04:58:01.0000] <Lachy>
ok

[05:06:00.0000] <hsivonen>
what's ANT's product like?

[05:06:01.0000] <jmb>
never seen it

[05:10:00.0000] <MikeSmith>
hmm, didn't know that ANT actually made their own browser

[05:10:01.0000] <MikeSmith>
http://www.antlimited.com/ant_galio_browser.asp?menu=153

[05:10:02.0000] <MikeSmith>
interesting bullet point: "Remote Event Support (HTML 5)"

[05:13:00.0000] <MikeSmith>
http://oregan.net/tv_web_browser_and_ui.php also maybe worth perusing

[05:14:00.0000] <MikeSmith>
..or maybe not

[05:15:00.0000] <MikeSmith>
not a lot of details there

[07:43:00.0000] <mookid>
err...

[07:43:01.0000] <mookid>
I missed the RTF drama cos I ignored the thread

[07:43:02.0000] <mookid>
epic.

[07:51:00.0000] <Philip`>
RTF?

[07:53:00.0000] <mookid>
fielding

[07:53:01.0000] <mookid>
it's interesting because I honestly had no idea that was going on

[07:53:02.0000] <mookid>
infact I can't even find the thread in my mailbox for some reason

[07:54:00.0000] <Philip`>
Ah

[07:54:01.0000] <mookid>
but it's interesting that he gets the same amount of pissed off I do at idiots who seem to have this jihad against architecture

[07:54:02.0000] <Philip`>
The one in public-html?

[07:54:03.0000] <mookid>
oh it's a different one to the one I'm on?

[07:54:04.0000] <mookid>
Am I on some beardy one then?

[07:54:05.0000] <mookid>
shiiiiiit I fked that up

[07:55:00.0000] <mookid>
maybe I'll forward all of that thread to the right list then..

[07:55:01.0000] <mookid>
:D

[07:55:02.0000] <mookid>
I'm not suprised most of whatwg didn't have a clue what I was getting at 

[07:55:03.0000] <mookid>
there's even guys on rest-discuss who dont get it

[07:55:04.0000] <takkaria>
hmm, mozilla's bugzilla instance looks different from a few weeks ago

[07:55:05.0000] <Philip`>
public-html is the W3C's HTML WG list

[07:55:06.0000] <mookid>
ahhh I see ok

[07:56:00.0000] <mookid>
I might post to that list then

[07:56:01.0000] <mookid>
:D

[07:56:02.0000] <Philip`>
It's largely the same group of people as the WHATWG

[07:56:03.0000] <mookid>
yeah i know but I can mop the floor with some of you dummies mroe publicy

[07:56:04.0000] <Philip`>
and if you didn't convince the relevant parties on the WHATWG, you probably won't convince the same people with the same arguments on a different list

[07:57:00.0000] <Philip`>
s/on the WHATWG/on the WHATWG list/

[07:57:01.0000] <mookid>
apart from the fact that most of the people on that list have demonstrated a complete lack of understanding

[07:57:02.0000] <mookid>
evidenced by the fact that they make counter-points which I've already covered in previous posts

[07:57:03.0000] <MikeSmith>
mookid: you're not allowed to use the word "architecture" here, nor "namespaces". unless you put cursewords in front of them. or otherwise heap abuse on them in some way. and you're allowed to use "extensibility" but not if you put "decentralized" or "distributed" in front of it

[07:57:04.0000] <mookid>
it's this rediculous thing where small minded people who've got used to a certain way of looking at things

[07:57:05.0000] <mookid>
are incapable of changing their perspective

[07:57:06.0000] <mookid>
so they attack you as 'wrong'

[07:58:00.0000] <mookid>
and 'misunderstanding' 'how it works in practice'

[07:58:01.0000] <Philip`>
mookid: That may be because people don't read all the previous posts, which is unhelpful but probably inevitable given the volume of mail that is often generated

[07:58:02.0000] <MikeSmith>
/me plays "Oh god, please don't let me be misunderstood" for mookid 

[07:58:03.0000] <MikeSmith>
mookid: are you a Morrisey fan?

[07:59:00.0000] <Dashiva>
So did anything change, or is it still appeal to architecture/authority?

[07:59:01.0000] <mookid>
are you deliberately coming across as a simpleton or is that unintentional?

[07:59:02.0000] <mookid>
I guess working on some irrelevant numpty spec is your best way to look important to non-techies

[07:59:03.0000] <mookid>
no offence.

[08:00:00.0000] <mookid>
offense

[08:00:01.0000] <MikeSmith>
what's numpty?

[08:00:02.0000] <mookid>
mentally challenged

[08:00:03.0000] <mookid>
but hey - who cares about this silly billy 'architecture' stuff - We've got VIDEO TAGS

[08:00:04.0000] <MikeSmith>
I'm mentally "differently abled"

[08:00:05.0000] <Dashiva>
Well, if HTML5 is irrelevant, surely you can just ignore us

[08:01:00.0000] <mookid>
you 'architects' can get lost with your solutions to problems that are already solved - we're gonna work on VIDEO TAGS

[08:01:01.0000] <mookid>
bwahahahaha.

[08:01:02.0000] <takkaria>
at some point I think you stepped from being someone who has a point and is trying to get other people to believe it into being a rude and offensive troll

[08:01:03.0000] <mookid>
it's hard to take you people seriously when the best you can do is attack my position as a developer

[08:02:00.0000] <mookid>
as 'cloud 9' or whatever retarded nonsense you use to describe it

[08:02:01.0000] <mookid>
it's cloud 9 to you because its above you, most likely

[08:02:02.0000] <mookid>
which would explain why you cant see any value in it

[08:02:03.0000] <mookid>
because you dont understand it

[08:02:04.0000] <mookid>
but I'm sure beigna code monkey is really mind bending Neo-from-the-matrix style stuff

[08:03:00.0000] <MikeSmith>
mookid: you clearly seem to have something against retarded people. and nonsense

[08:03:01.0000] <mookid>
not retarded people

[08:03:02.0000] <mookid>
retarded nonsense

[08:03:03.0000] <MikeSmith>
ah

[08:03:04.0000] <MikeSmith>
that clears things up

[08:03:05.0000] <mookid>
look it up

[08:04:00.0000] <MikeSmith>
you got anything against plain nonsense, or is it just retarded nonsense in particular?

[08:04:01.0000] <mookid>
both - you tick both boxes so you're a full house

[08:05:00.0000] <MikeSmith>
/me tries to pat himself on the back but finds that he lacks the proper motor skills

[08:05:01.0000] <mookid>
It's nonsensical politics like this that is dragged up because your fear caused by lack of understanding drags everything down

[08:05:02.0000] <mookid>
every time I had a conversation on here

[08:05:03.0000] <mookid>
broke it down

[08:05:04.0000] <mookid>
line by line

[08:06:00.0000] <mookid>
someone goes afk

[08:06:01.0000] <mookid>
never comes back

[08:06:02.0000] <mookid>
apart from Philip` who was honest enough to find it worth considering

[08:07:00.0000] <MikeSmith>
mookid: sorry, I got to go afk for a while

[08:07:01.0000] <MikeSmith>
bbiab

[08:07:02.0000] <mookid>
when people look back through these logs in 40 years time

[08:07:03.0000] <mookid>
you'll be the bad guy with his head up his arse

[08:07:04.0000] <mookid>
:)

[08:08:00.0000] <MikeSmith>
I feel sorry for anybody who looks back through these logs in 40 years time

[08:08:01.0000] <takkaria>
lolwut

[08:08:02.0000] <Dashiva>
They should just read the lastweek archives instead

[08:08:03.0000] <MikeSmith>
and I'm already the bad guy with his head up his arse. ask anybody. we don't need to wait 40 years for people to agree about that.

[08:09:00.0000] <Dashiva>
MikeSmith: I don't agree!

[08:09:01.0000] <Dashiva>
You're the guy who takes beer breaks

[08:09:02.0000] <Philip`>
Dashiva: They're not just beer any more

[08:09:03.0000] <mookid>
Crack aswell no doubt

[08:10:00.0000] <Dashiva>
It's interesting to see that @accept got the exact same response on rest-discuss

[08:12:00.0000] <mookid>
they'll look through these logs to try and work out how anyone could possibly be so incompetant as to not get Hypertext working with the protocol

[08:12:01.0000] <Philip`>
Dashiva: Was there less sarcasm on there?

[08:12:02.0000] <mookid>
that takes high levels of stupidity it really does

[08:12:03.0000] <mookid>
'oh well HTML is used in emails'

[08:12:04.0000] <mookid>
'and in help documents'

[08:14:00.0000] <mookid>
just the most ridiculous exchange I've had - and the stubborness of that kind of small mindedness astonishes me

[08:14:01.0000] <Dashiva>
Philip`: Less sarcasm, more technical beatdown

[08:15:00.0000] <mookid>
Dashiva: actually, it didn't - it got that response from abuot 4 people who clearly dont understand the implications or what I'm getting at

[08:15:01.0000] <mookid>
one of the first responses basically said

[08:15:02.0000] <mookid>
"your looking for the right solution but yuo need to take into account all of HTTP conneg"

[08:15:03.0000] <mookid>
"WHATWG are a lost cause, but your intentions are noble"

[08:16:00.0000] <mookid>
then the idiots chimed in with their regurgitated scripture they think they understand

[08:16:01.0000] <mookid>
if you notice.. no one with any calibre has told me to stand down

[08:16:02.0000] <Dashiva>
There we go with the authority again

[08:17:00.0000] <mookid>
well presumably smoeone with some weight would step in and put an end to it if I didnt have a point

[08:17:01.0000] <mookid>
here we go with the little man's lament: everyone's equal

[08:17:02.0000] <mookid>
the world can be fair

[08:17:03.0000] <mookid>
no it can't.

[08:17:04.0000] <mookid>
some people are good at some stuff and bad at other stuff

[08:17:05.0000] <mookid>
that's life

[08:18:00.0000] <mookid>
HTML guys are good at doing dirty work no one else does

[08:18:01.0000] <Dashiva>
And you're good at alienating everyone who doesn't agree with you

[08:18:02.0000] <mookid>
well what exactly do you expect

[08:18:03.0000] <mookid>
I've had no constructive exchange with any of you other than Philip` 

[08:19:00.0000] <mookid>
just arguing in circles ignoring previous posts

[08:19:01.0000] <Philip`>
/me doesn't remember being constructive

[08:19:02.0000] <mookid>
liar.

[08:19:03.0000] <mookid>
asking good questions is constructive

[08:19:04.0000] <mookid>
:P

[08:20:00.0000] <mookid>
the problem here is that URI's have been totally raped for such a long time

[08:20:01.0000] <mookid>
that all the UA's and stuff arent geared up to use HTTP properly

[08:20:02.0000] <mookid>
and so you guys wont budge

[08:20:03.0000] <mookid>
because you cant see any other alternative

[08:20:04.0000] <mookid>
even though, realistically, the markup is going to have to be the thing that moves first

[08:21:00.0000] <mookid>
and I'm providing a way to make that move without disturbing the old URI-based way of doing it

[08:21:01.0000] <mookid>
you *still* resist it on the basis that "the costs to the internet will be too high - we much protect netizens"

[08:22:00.0000] <mookid>
when I talked with Hix about it it basically came down to that

[08:22:01.0000] <mookid>
"I dont see it that way, and if we allow developers to do things differently it could make some stuff not work very well. It is my job to protect the internet"

[08:23:00.0000] <mookid>
and whilst that is incredibly amusing

[08:23:01.0000] <mookid>
it's also epically stupid

[08:23:02.0000] <mookid>
it's a markup language.

[08:23:03.0000] <mookid>
cmon.

[08:23:04.0000] <mookid>
I know it's nice for everyone to feel important and it helps the work get done but seriously

[08:23:05.0000] <Dashiva>
When one of your use cases is "I want to host images with .htm extensions", then yes, protection is in order

[08:23:06.0000] <mookid>
let's no kid ourselves here

[08:23:07.0000] <mookid>
er

[08:24:00.0000] <mookid>
what?

[08:24:01.0000] <mookid>
that's not an extension anyway that's just a URI pattern

[08:24:02.0000] <mookid>
and I'm saying the opposite

[08:24:03.0000] <mookid>
when you put a '.' in a URI

[08:25:00.0000] <mookid>
that doesnt actually mean anything

[08:25:01.0000] <Dashiva>
People are used to working with files. They are used to extensions having meaning. So while your abstract view may ignore it, .htm does suggest it's a HTML document

[08:25:02.0000] <mookid>
URIs are not file paths

[08:25:03.0000] <mookid>
I dont care what people are used to

[08:25:04.0000] <mookid>
I care what the web can be

[08:25:05.0000] <mookid>
and how the web can (and will eventually) work

[08:25:06.0000] <mookid>
ok ok

[08:25:07.0000] <mookid>
just read this please

[08:25:08.0000] <mookid>
MY PROPOSAL

[08:26:00.0000] <mookid>
DOES NOT STOP YOU

[08:26:01.0000] <mookid>
FROM DOING THINGS

[08:26:02.0000] <mookid>
THE SAME WAY

[08:26:03.0000] <mookid>
IT JUST ALLOWS ME

[08:26:04.0000] <mookid>
TO DO IT MAY WAY

[08:26:05.0000] <mookid>
ASWELL

[08:26:06.0000] <mookid>
get it?

[08:26:07.0000] <Dashiva>
And your way is redundant, while being confusing in addition

[08:26:08.0000] <mookid>
it's not redundant and it's not confusing

[08:27:00.0000] <mookid>
URI based conneg is confusing if the developers an idiot

[08:27:01.0000] <mookid>
there's nothign sotpping me from setting up /index.html and serving an executable

[08:27:02.0000] <Dashiva>
There is one thing: Common sense

[08:27:03.0000] <mookid>
what the hell would yuo know abuot that?

[08:27:04.0000] <Philip`>
Common sense? On the internet? You can't be serious

[08:28:00.0000] <mookid>
Common sense would tell me that people need to learn what URI's actually mean

[08:28:01.0000] <mookid>
and how HTTP works

[08:28:02.0000] <mookid>
and that all applications talk HTTP

[08:28:03.0000] <mookid>
and that you dont need to go throgh a browser to use HTTP

[08:28:04.0000] <mookid>
but you dont understand that

[08:28:05.0000] <mookid>
so you think I'm wrong

[08:28:06.0000] <Dashiva>
Right, so you push a huge burden on everyone in the world, just so you can have your attribute

[08:28:07.0000] <mookid>
FOR MY APP

[08:28:08.0000] <mookid>
AS A DEVELOPER

[08:28:09.0000] <mookid>
THAT IS MY CHOICE

[08:28:10.0000] <mookid>
ITS MY APPLICATION

[08:29:00.0000] <mookid>
if I want to use HTTP properly

[08:29:01.0000] <mookid>
why should I not be able to just because idiots like you dont

[08:29:02.0000] <Dashiva>
hsivonen, you could teach mookid about externalities?

[08:29:03.0000] <mookid>
I know more abuot economics and market forces than you do

[08:30:00.0000] <Dashiva>
Then you should apply that knowledge

[08:30:01.0000] <mookid>
I have.

[08:30:02.0000] <mookid>
let the market decide

[08:30:03.0000] <mookid>
allow both

[08:30:04.0000] <mookid>
see which applications survive

[08:30:05.0000] <Philip`>
Currently the market is deciding to not implement the feature you want

[08:30:06.0000] <Dashiva>
And the marked decides it doesn't want to waste resources implementing something that already works

[08:30:07.0000] <mookid>
because the idiots in control of the markup dont allow it

[08:30:08.0000] <mookid>
not because it's not possible

[08:30:09.0000] <mookid>
that's a governance problem

[08:30:10.0000] <Dashiva>
Go make your own browser then

[08:31:00.0000] <mookid>
government inefficiencies taking effect

[08:31:01.0000] <mookid>
WHATWG is the bush admiinistration of teh web

[08:31:02.0000] <mookid>
"dont worry world you are free (to do as we tell you)"

[08:31:03.0000] <Philip`>
Browser developers do loads of things that the "[people] in control of the markup" don't allow, if they think it's a worthwhile thing to add to their browser

[08:31:04.0000] <mookid>
here's have some freedom

[08:31:05.0000] <hsivonen>
Dashiva: I'm not going to teach Econ 101 right now, but I can suggest Wikipedia

[08:32:00.0000] <mookid>
Philip`: that's not really a fair argument since there's very little incentive to attempt that if it isn't backed by standards

[08:32:01.0000] <Dashiva>
eh?

[08:32:02.0000] <mookid>
particularly something like this which is 'already achievable' but just in an inferior way

[08:32:03.0000] <mookid>
so you have to explain to peopel who dont understand you

[08:32:04.0000] <mookid>
why it's really great

[08:32:05.0000] <mookid>
and that's hard

[08:32:06.0000] <Dashiva>
"very little incentive to attempt that if it isn't backed by standards" <-- Is this referring to the real world?

[08:32:07.0000] <mookid>
where as if everyone just accepted HTTP conneg existed

[08:33:00.0000] <mookid>
and it would be great to allow browsers to actually make decent use of it

[08:33:01.0000] <mookid>
we could let developers decide

[08:33:02.0000] <Philip`>
/me discovers that Trac now has blame support, and is happy

[08:33:03.0000] <mookid>
it's such a simple mechanism to implement anyway

[08:33:04.0000] <mookid>
browsers already have accept headers

[08:33:05.0000] <mookid>
one optional attribute..

[08:33:06.0000] <Dashiva>
Browsers negotiate on their own behalf, not yours

[08:33:07.0000] <mookid>
your an idiot

[08:34:00.0000] <Dashiva>
If you want to force a request to accept only X, you make a separate URL that only provides X

[08:34:01.0000] <Dashiva>
Problem solved

[08:34:02.0000] <mookid>
jesus fucking christ

[08:34:03.0000] <mookid>
give me strength

[08:34:04.0000] <hsivonen>
mookid: are you Mike from http://mykanjo.co.uk/ ?

[08:34:05.0000] <mookid>
no

[08:36:00.0000] <hsivonen>
mookid: oh. what's your From address in the email thread you referred to?

[08:36:01.0000] <mookid>
wow yuo caught me

[08:36:02.0000] <mookid>
well done inspector

[08:36:03.0000] <mookid>
why ask the question then

[08:37:00.0000] <hsivonen>
I just wanted to make sure I connect the right email and IRC names

[08:37:01.0000] <mookid>
why?

[08:38:00.0000] <mookid>
Dashiva: if you think what you just said is a valid point then you really dont have a clue what I'm saying

[08:38:01.0000] <hsivonen>
mookid: connecting the names helps me follow discussion

[08:38:02.0000] <Dashiva>
It's not what you're saying, it's the consequences of what you're proposing

[08:39:00.0000] <mookid>
hsivonen: you dont discuss anything you just leap in if you think yuo can be pedantic about something

[08:39:01.0000] <mookid>
Dashiva: and what would those be exactly?

[08:39:02.0000] <mookid>
that there would be 2 different, co-existing solutions to a problem

[08:40:00.0000] <Dashiva>
That there would be added work for anyone implementing HTML5

[08:40:01.0000] <mookid>
no sht

[08:40:02.0000] <mookid>
It's more important than bloody video tags

[08:41:00.0000] <Dashiva>
Video provides something that's not already supported. (Flash is not native video, before you try to say it.)

[08:41:01.0000] <mookid>
but.. I can have flash in a page

[08:41:02.0000] <mookid>
video works already

[08:41:03.0000] <mookid>
problem solved

[08:41:04.0000] <mookid>
right?

[08:41:05.0000] <mookid>
:)

[08:42:00.0000] <Dashiva>
No, that's not native video

[08:42:01.0000] <mookid>
so?

[08:42:02.0000] <mookid>
it still works

[08:42:03.0000] <mookid>
youtube's not done too badly

[08:42:04.0000] <Dashiva>
We don't want the web to be dependent on proprietary technology, y'know

[08:42:05.0000] <mookid>
?

[08:42:06.0000] <mookid>
so I can't write my own youtube?

[08:43:00.0000] <mookid>
I think I probably can and you're talking out of where your poo poo comes from

[08:43:01.0000] <Philip`>
mookid: Flash doesn't work on all platforms, and you presumably have to pay Adobe if you want them to port it to yours, which is a problem

[08:44:00.0000] <mookid>
errr..

[08:44:01.0000] <mookid>
I'm not even going to bother carrying this on it's so obvious how conflicting these 2 approaches are

[08:44:02.0000] <Philip`>
(where "platforms" includes iPhones and Wiis and that kind of thing, which want to be able to access the web)

[08:44:03.0000] <mookid>
we dont need video tags.

[08:44:04.0000] <mookid>
video works

[08:45:00.0000] <mookid>
you're doing aload of work for something that no one really needs

[08:45:01.0000] <mookid>
according to your logic at least

[08:45:02.0000] <mookid>
because it already works

[08:45:03.0000] <mookid>
there already are methods for providing video

[08:47:00.0000] <mookid>
they work perfectly well

[08:47:01.0000] <mookid>
no improvement for end users

[08:47:02.0000] <mookid>
not substantial ones anyway

[08:47:03.0000] <Philip`>
Those existing methods have significant practical problems (cost, non-portability, dependence on a particular company, etc), which are sufficient to make at least three major browser developers quite eager to implement it

[08:48:00.0000] <mookid>
ooooooh

[08:48:01.0000] <mookid>
the browser develoeprs

[08:48:02.0000] <mookid>
gee

[08:48:03.0000] <mookid>
they've done a real good job so far

[08:48:04.0000] <mookid>
what with CSS and everything

[08:49:00.0000] <mookid>
I'm sure that won't turn out to be another 5 year long mess of inconsistency and mud slinging

[08:49:01.0000] <Philip`>
It doesn't matter how good a job they've done - what matters is that they decide what they're going to implement (and so they decide what features users will have access to, and what problems authors will have to put up with, etc)

[08:49:02.0000] <mookid>
video works.

[08:49:03.0000] <gsnedders>
mookid: You realize browsers are going to implement such things regardless of whether there's a spec or not?

[08:49:04.0000] <mookid>
it's works fine

[08:49:05.0000] <mookid>
right.. but browsers have INVESTED interest

[08:49:06.0000] <mookid>
in making everyone thing

[08:49:07.0000] <Philip`>
/me disagrees that video works fine, since he can't actually see any of it in his main browser, because the Flash plugin is broken and keeps crashing

[08:49:08.0000] <mookid>
think

[08:49:09.0000] <gsnedders>
mookid: There will be even more inconsistency if we rely on everything reverse-engineering everything else.

[08:50:00.0000] <mookid>
that you ned to go through a browser

[08:50:01.0000] <mookid>
to suse HTTP

[08:50:02.0000] <mookid>
use^

[08:50:03.0000] <gsnedders>
mookid: The way to get the most consistency is to write a spec.

[08:50:04.0000] <mookid>
why the hell should I have to click a link that opens up a link to a pdf and then opens that after being downloaded from the browser in acrobat?

[08:51:00.0000] <mookid>
that's so stupid it's unreal

[08:51:01.0000] <mookid>
why not just.. open it in acrobat..

[08:51:02.0000] <mookid>
DUN DUN DUUUUUH!

[08:51:03.0000] <gsnedders>
mookid: Security? PDF is scriptable.

[08:51:04.0000] <mookid>
security

[08:52:00.0000] <mookid>
seriously?

[08:52:01.0000] <mookid>
what do you know about security

[08:52:02.0000] <mookid>
explain the difference in security between those two procedures

[08:52:03.0000] <mookid>
infact

[08:52:04.0000] <gsnedders>
mookid: One relies upon specific user interaction, the other does not

[08:52:05.0000] <mookid>
er.. right.

[08:52:06.0000] <gsnedders>
mookid: (to actually open it as well as naviagting to it)

[08:53:00.0000] <gsnedders>
mookid: it is quite easily possible to fool users into navigating to things

[08:53:01.0000] <mookid>
why the hell should I use a browser to get to a pdf dcument though?

[08:53:02.0000] <gsnedders>
mookid: What makes you?

[08:53:03.0000] <mookid>
er..

[08:53:04.0000] <mookid>
well the use case I'm being given is

[08:53:05.0000] <gsnedders>
mookid: Nothing stops you from using another HTML parser to get the URL of the PDF and using some other application to download it

[08:53:06.0000] <mookid>
someone sends me an email

[08:53:07.0000] <mookid>
to a URI

[08:53:08.0000] <mookid>
unless the pdf is explicitly in the URI

[08:54:00.0000] <mookid>
if I click it it will only get me the html

[08:54:01.0000] <gsnedders>
mookid: So you want your email client to make an HTTP request to see what sort of file it is?

[08:54:02.0000] <mookid>
see what content types are available yeah

[08:54:03.0000] <mookid>
content negotiation

[08:54:04.0000] <mookid>
it's genius.

[08:55:00.0000] <mookid>
resources + representations

[08:55:01.0000] <gsnedders>
mookid: Sure, but the resources should be identical. What difference does it make?

[08:55:02.0000] <gsnedders>
(in terms of content)

[08:55:03.0000] <mookid>
well nothing other than just representation

[08:55:04.0000] <mookid>
that's how it should be

[08:55:05.0000] <mookid>
that's why they're at the same URI

[08:55:06.0000] <gsnedders>
So why does it matter?

[08:55:07.0000] <gsnedders>
I don't get this.

[08:55:08.0000] <mookid>
what?

[08:55:09.0000] <gsnedders>
Why does it matter you get HTML and not the PDF?

[08:56:00.0000] <mookid>
if the resources are the *same* they should be indentified by the same resource identifier

[08:56:01.0000] <mookid>
not two seperate ones

[08:56:02.0000] <mookid>
because the only difference is the representation

[08:56:03.0000] <gsnedders>
/me still doesn't get the problem here

[08:56:04.0000] <mookid>
which HTTP provides a way to negotiate

[08:56:05.0000] <gsnedders>
mookid: Yeah, sure. Where's the disagreement?

[08:56:06.0000] <mookid>
well if I have /report serving html and pdf

[08:57:00.0000] <mookid>
and I want to link to it in my HTML application

[08:57:01.0000] <mookid>
a browser will only ever be able to get hold of the HTML

[08:57:02.0000] <mookid>
I cant make a link to the pdf

[08:57:03.0000] <mookid>
there's no way of overriding the accept header

[08:57:04.0000] <mookid>
so HTTP content negotiaition doesnt work

[08:57:05.0000] <gsnedders>
Linking to the PDF is avoiding the entire point of the content-negotiation. The entire point of the content negotiation is to get a copy that the current HTTP client understands.

[08:58:00.0000] <gsnedders>
How do you know the HTTP client will understand what you put in the accept header?

[08:58:01.0000] <mookid>
no

[08:58:02.0000] <mookid>
it's not just UA

[08:58:03.0000] <mookid>
read the RFC

[08:58:04.0000] <mookid>
it's UA and user preference

[08:58:05.0000] <gsnedders>
I'm well aware of what the RFC says.

[08:58:06.0000] <gsnedders>
This isn't UA preference.

[08:58:07.0000] <mookid>
well clearly you arent

[08:58:08.0000] <gsnedders>
This is the server preference.

[08:58:09.0000] <mookid>
no it's USER preference

[08:58:10.0000] <gsnedders>
You want server preference.

[08:58:11.0000] <gsnedders>
This isn't the user making a choice.

[08:59:00.0000] <mookid>
that's the user preference.. they dont have to click that link do they?

[08:59:01.0000] <gsnedders>
No, but how do they get the version they would be sent with their normal accept header then?

[08:59:02.0000] <gsnedders>
They have already specified their preference in their normal accept header.

[09:00:00.0000] <gsnedders>
Why do you want to provide a UI to change it within the link itself?

[09:00:01.0000] <mookid>
well that link should only be a link to one representation

[09:00:02.0000] <mookid>
just that one link

[09:00:03.0000] <mookid>
for that one request

[09:00:04.0000] <mookid>
I could provide the standard link and the pdf link next to each other

[09:00:05.0000] <gsnedders>
What you want is a means for a user to choose a different file format preference for one request. That shouldn't be done _in the link_.

[09:00:06.0000] <mookid>
one with no @accept set and the with pdf type specified

[09:01:00.0000] <mookid>
it IS done in the link

[09:01:01.0000] <mookid>
even if it's in the URI

[09:01:02.0000] <mookid>
it's still IN THE LINK

[09:01:03.0000] <mookid>
it's just SHOEHORNED server control

[09:01:04.0000] <mookid>
so you can feel ok abuot it

[09:01:05.0000] <mookid>
that criticism doesnt make any sense at all

[09:02:00.0000] <mookid>
your alternative is to have a link to report.pdf.. what is that if it's not server specified and 'in the link' ?

[09:02:01.0000] <gsnedders>
If the user wants to override their ready made preference about preferred formats, they should do it within their HTTP client. We don't need to extend RFC26616's section on HTTP urls to do this at all

[09:02:02.0000] <mookid>
That's not an extension it's completely within the bounds of it

[09:03:00.0000] <gsnedders>
Where do you want the override of the accept header to be specified?

[09:03:01.0000] <mookid>
in the markup of the link..

[09:03:02.0000] <Philip`>
/me didn't know they had written that many RFCs yet

[09:03:03.0000] <gsnedders>
mookid: Then you've only solved your problem for HTML, and not for anything like PDF.

[09:04:00.0000] <mookid>
because BROWSERS

[09:04:01.0000] <mookid>
HTML BROWSERS

[09:04:02.0000] <mookid>
are a special case

[09:04:03.0000] <mookid>
I've already explained this

[09:04:04.0000] <gsnedders>
WHY!?

[09:04:05.0000] <mookid>
THEY ACCEPT */*

[09:04:06.0000] <gsnedders>
So?

[09:04:07.0000] <mookid>
NAME ANOTHER UA THAT IT WOULD MAKE SENSE TO DO THAT

[09:04:08.0000] <Dashiva>
wget

[09:04:09.0000] <gsnedders>
curl, wget, GET?

[09:04:10.0000] <mookid>
er...

[09:04:11.0000] <mookid>
they all have configurable accept headers

[09:04:12.0000] <mookid>
good point though.

[09:04:13.0000] <gsnedders>
so do some browsers.

[09:04:14.0000] <mookid>
no they dont

[09:05:00.0000] <mookid>
not in the context of an application

[09:05:01.0000] <mookid>
it needs to be dynamic

[09:05:02.0000] <gsnedders>
mookid: Then that is an issue for their UI, not for HTML.

[09:05:03.0000] <mookid>
because I could have 3 links all next to each other to /report one for deafult, one specifically for xml and one for pdf

[09:05:04.0000] <mookid>
what?

[09:05:05.0000] <mookid>
Hyper text

[09:05:06.0000] <mookid>
markup

[09:06:00.0000] <mookid>
if yuo're marking up the hyper text

[09:06:01.0000] <mookid>
then surely it's a good idea to have hyper text actually able to hyper link properly

[09:06:02.0000] <mookid>
and content negotatiate over the transfer protocol

[09:06:03.0000] <gsnedders>
mookid: If the user wants to override their ready made choice about what format they prefer, they should make it within their UA. It should not be done from a link.

[09:07:00.0000] <gsnedders>
mookid: It should content negotiate and give it what the user prefers, and the user has already made their preference.

[09:07:01.0000] <mookid>
and they can.. but if a developer wants to be specific

[09:07:02.0000] <mookid>
the developer should eb able to be specific

[09:07:03.0000] <mookid>
for a given link

[09:07:04.0000] <mookid>
it's his/her link in his/her application

[09:07:05.0000] <Dashiva>
And all of this just because you don't want to make a URL

[09:07:06.0000] <mookid>
well the implication of doing things this way are far reaching

[09:08:00.0000] <mookid>
but I dont expect you to even begin to understand any of that

[09:08:01.0000] <gsnedders>
mookid: Then you have no content negotiation. You are requesting a specific format of a specific file. The accept header isn't for that.

[09:08:02.0000] <mookid>
what?

[09:08:03.0000] <mookid>
that's still content negotiation

[09:08:04.0000] <mookid>
do you know what content negotiation means?

[09:08:05.0000] <Dashiva>
content ultimatum

[09:08:06.0000] <gsnedders>
mookid: IT'S NOT THE USER'S CHOICE ANYMORE.

[09:08:07.0000] <mookid>
WHAT@:?

[09:08:08.0000] <mookid>
HOW IS IT THEIR CHOICE

[09:08:09.0000] <mookid>
IF YOU GIVE THEM FUCKING LINK

[09:08:10.0000] <krijn>
-_-

[09:08:11.0000] <mookid>
TO REPORT.PDF

[09:08:12.0000] <mookid>
YOU CRETIN

[09:08:13.0000] <mookid>
THINK ABOUT WHAT YOU'RE SAYING

[09:09:00.0000] <mookid>
jesus

[09:09:01.0000] <gsnedders>
If you're going to start calling me a cretin, I'm leaving. There's no use in arguing this with you.

[09:09:02.0000] <mookid>
ohhhh

[09:09:03.0000] <mookid>
good move

[09:09:04.0000] <mookid>
probably for the best this is going badly for you

[09:09:05.0000] <Dashiva>
/me chuckles

[09:09:06.0000] <Dashiva>
Yeah gsnedders, you totally got schooled! :P

[09:10:00.0000] <Dashiva>
Now that I know I can win all debates by calling people cretins, nothing will stand in my way

[09:10:01.0000] <mookid>
well seriously I mean.. if you exspect me to just listen to you say that and not get irritated by yuor blatant lack of consideration for what you're actually saying

[09:10:02.0000] <mookid>
well go on

[09:11:00.0000] <mookid>
ignoring that I said cretin

[09:11:01.0000] <mookid>
want to respond?

[09:11:02.0000] <mookid>
no. becuase you cant

[09:11:03.0000] <gsnedders>
mookid: I'm not the one making ad hominem attacks.

[09:11:04.0000] <mookid>
well go on

[09:11:05.0000] <mookid>
be the bigger man

[09:11:06.0000] <mookid>
ignore it

[09:11:07.0000] <mookid>
and respond to what I actually sai

[09:11:08.0000] <mookid>
said

[09:11:09.0000] <mookid>
how is a link to /report.pdf giving the user a choice?

[09:11:10.0000] <gsnedders>
mookid: No, there is no point in arguing with people who revert to such attacks. You're ignoring everything I say.

[09:12:00.0000] <mookid>
no I'm processing it, understanding it, and then pointing out how you're apparently superior alternative has exactly the same 'problem'

[09:12:01.0000] <mookid>
so

[09:12:02.0000] <mookid>
how is a link to /report.pdf giving the user a choice?

[09:12:03.0000] <krijn>
"You can either open this, or not"

[09:12:04.0000] <mookid>
^

[09:12:05.0000] <Dashiva>
Because report.pdf also implies report.html

[09:13:00.0000] <mookid>
er

[09:13:01.0000] <mookid>
no it doesnt

[09:13:02.0000] <Dashiva>
Sure it does, otherwise you wouldn't need @accept on the link

[09:13:03.0000] <mookid>
er

[09:13:04.0000] <mookid>
lol 

[09:14:00.0000] <mookid>
dont you think my way of just having /report is a *tiny bit* more clear on that?

[09:14:01.0000] <Dashiva>
No, it's less clear

[09:14:02.0000] <Dashiva>
Users can relate to .pdf and .html

[09:14:03.0000] <mookid>
how are they supposed to know they are the same resource/

[09:14:04.0000] <Dashiva>
They don't need to

[09:14:05.0000] <mookid>
what happens if one part of my appication they are

[09:14:06.0000] <mookid>
what?

[09:14:07.0000] <mookid>
but this guy is saying

[09:15:00.0000] <Dashiva>
It's your job as a developer to not confuse the users

[09:15:01.0000] <mookid>
'this link with the accept header specified is not giving the user the choice'

[09:15:02.0000] <mookid>
I'm saying

[09:15:03.0000] <mookid>
neither is a link to  /report.pdf

[09:15:04.0000] <mookid>
because it's not

[09:15:05.0000] <gsnedders>
I am saying that control over the accept header lies with the UA and the user. Nothing on the server, or from the server, should have any control. That is all.

[09:15:06.0000] <mookid>
why?

[09:15:07.0000] <mookid>
when browsers accept all types anyway

[09:16:00.0000] <mookid>
and when the only alternative is to do it in the URI

[09:16:01.0000] <mookid>
in exactly the same 'controlling' 'choiceles' way

[09:16:02.0000] <krijn>
BenMillard: I took some of your ideas, not yet all

[09:16:03.0000] <gsnedders>
It isn't over ruling content negotiation doing it that way though, mookid 

[09:16:04.0000] <mookid>
what?

[09:17:00.0000] <mookid>
putting .pdf in the URI if it's actually a representation

[09:17:01.0000] <mookid>
and not a seperate resource

[09:17:02.0000] <mookid>
*IS* content negotiation

[09:17:03.0000] <gsnedders>
if you request /report.pdf where only one resource exists you expect to get that resource.

[09:17:04.0000] <mookid>
it's URI-BASED CONTENT NEGOTIATION

[09:17:05.0000] <krijn>
And it's great!

[09:17:06.0000] <krijn>
Better than sliced bread

[09:17:07.0000] <mookid>
:)

[09:17:08.0000] <gsnedders>
if you request /report where both HTML and PDF exist, and you prefer HTML, you should get HTML.

[09:18:00.0000] <mookid>
yeah..

[09:18:01.0000] <gsnedders>
if you want to make a request for a PDF version of /report, you shouldn't be using the /report URI.

[09:18:02.0000] <mookid>
what?

[09:18:03.0000] <Dashiva>
gsnedders: He's had that explained many times, it doesn't get through

[09:19:00.0000] <mookid>
it doesnt make sense

[09:19:01.0000] <gsnedders>
mookid: the /report URI gives what I prefer. I prefer HTML. Give me that when I request it.

[09:19:02.0000] <gsnedders>
mookid: Therefore, you can't use that URI if you want to give me a PDF.

[09:19:03.0000] <mookid>
NO

[09:19:04.0000] <mookid>
YOUR RIGHT

[09:19:05.0000] <mookid>
THATS WHAT THIS PROPOSAL IS FOR

[09:20:00.0000] <krijn>
lol

[09:20:01.0000] <mookid>
if it already worked

[09:20:02.0000] <mookid>
why would I be proposing something?

[09:20:03.0000] <mookid>
were you dorpped on your head as a child or something?

[09:20:04.0000] <Dashiva>
(It does work, if you aren't silly enough to insist on putting everything into a single URL)

[09:20:05.0000] <krijn>
Why is URI based conneg bad, but HTML based conneg good?

[09:20:06.0000] <mookid>
silly enough?

[09:20:07.0000] <mookid>
it's not HTML based conneg

[09:20:08.0000] <mookid>
it's HTTP based conneg

[09:20:09.0000] <gsnedders>
mookid: You don't want to use the same URI. I've already said I prefer HTML. I don't want something in an HTML page to overrule what I've said I prefer.

[09:20:10.0000] <krijn>
Ow, I thought you were still talking about accept=""

[09:21:00.0000] <krijn>
Or something

[09:21:01.0000] <gsnedders>
mookid: No, you're having HTML over rule HTTP conneg

[09:21:02.0000] <mookid>
I am but that's not the html doing it that's the html telling the browser to use HTTP

[09:21:03.0000] <mookid>
it's http conneg not html conneg

[09:21:04.0000] <mookid>
NO IM NOT

[09:21:05.0000] <gsnedders>
mookid: You're changing HTTP conneg via HTML.

[09:21:06.0000] <mookid>
I'm having html make use of HTTP

[09:21:07.0000] <krijn>
And how would you hand over that PDF version via IM? Or mail?

[09:22:00.0000] <krijn>
Exactly, using the URI

[09:22:01.0000] <mookid>
I'm changing the client request via html so it's appropriate for the request inidcated by the link - what th hell is wrong with that?

[09:22:02.0000] <Philip`>
krijn: We are still talking about accept="", I think :-)

[09:22:03.0000] <krijn>
Ow

[09:22:04.0000] <gsnedders>
mookid: You already can. With a different URI. What's wrong with that?

[09:22:05.0000] <mookid>
oh for fucks sake

[09:22:06.0000] <gsnedders>
mookid: You still won't solve the problem for IM or email?

[09:22:07.0000] <mookid>
I KNOW THAT

[09:22:08.0000] <mookid>
you think I dont know that?

[09:23:00.0000] <gsnedders>
mookid: How do you link directly to the PDF from IM?

[09:23:01.0000] <mookid>
you tell someone to open the URI in pdf viewer

[09:23:02.0000] <mookid>
why should a URI have to go through a brower?

[09:23:03.0000] <mookid>
browser

[09:23:04.0000] <Dashiva>
Actually, wouldn't your proposal entrench this practice and actively hurt people using IM and email

[09:23:05.0000] <gsnedders>
mookid: Then why can't you tell someone from your HTML page, "hey, open this URI in a pdf viewer"?

[09:23:06.0000] <gsnedders>
mookid: (if you want the PDF version and not the HTML one)

[09:24:00.0000] <mookid>
no because if the IM clients talk HTTP properly they can get a list of the content types and give the user a right click menu that lists all the UAs on their system they can open the URI with

[09:24:01.0000] <mookid>
there is

[09:24:02.0000] <mookid>
NO GOOD REASON

[09:24:03.0000] <mookid>
EVERY HTTP REQUEST

[09:24:04.0000] <mookid>
NEEDS TO GO

[09:24:05.0000] <mookid>
THROUGH AN HTML BROWSER

[09:24:06.0000] <gsnedders>
Who's arguing it does?

[09:24:07.0000] <Philip`>
How can an IM client get a list of available content types when it's given a URI?

[09:25:00.0000] <mookid>
options

[09:25:01.0000] <Dashiva>
And now we're updating all the IM clients in the world too

[09:26:00.0000] <Dashiva>
What else needs fixing?

[09:26:01.0000] <krijn>
I'm convinced. This is just brilliant.

[09:26:02.0000] <krijn>
Now let's go shopping.

[09:26:03.0000] <mookid>
if you want to continue doing it your URI-based way you can

[09:26:04.0000] <mookid>
and if I want to do it my way I can

[09:26:05.0000] <krijn>
Ow, darn

[09:26:06.0000] <mookid>
at the moment there's only your way

[09:26:07.0000] <mookid>
dont you think it would be better to start allowing applications to work that way?

[09:27:00.0000] <krijn>
No

[09:27:01.0000] <mookid>
and start using URIs as they were intended

[09:27:02.0000] <mookid>
ok well you're an idiot then

[09:27:03.0000] <gsnedders>
mookid: What makes you think they were intended that way?

[09:27:04.0000] <krijn>
(That's a fact btw)

[09:27:05.0000] <mookid>
UNIFORM

[09:27:06.0000] <mookid>
RESOURCE

[09:27:07.0000] <mookid>
IDENTIFIER

[09:27:08.0000] <mookid>
resources are nothing to do with representations

[09:27:09.0000] <mookid>
YEs

[09:27:10.0000] <mookid>
you can use them to do that

[09:27:11.0000] <mookid>
that's not how they were intended

[09:28:00.0000] <mookid>
that's like saying a news paper is a weapon

[09:28:01.0000] <Philip`>
mookid: Should browsers do the same kind of right click menu to let you see all the content types of a URI?

[09:28:02.0000] <krijn>
mookid: shouldn't you be telling browser vendors to do something about */*

[09:28:03.0000] <gsnedders>
mookid: Why should IM clients use OPTIONS and HTML UAs not?

[09:28:04.0000] <mookid>
Philip`: yes - but there are situations in an HTML applications where you would want to be specific

[09:28:05.0000] <mookid>
but yeah I take your point that would minimize it

[09:29:00.0000] <gsnedders>
/me was beat by Philip`, again :\

[09:29:01.0000] <mookid>
krijn: no - that's the correct default they just GET pretty much any content

[09:30:00.0000] <mookid>
so if I give you example.com/somethinginteresting - your browser should just go there and be able to download it or view it or something

[09:30:01.0000] <gsnedders>
mookid: Why do you want to link to a representation of a resource using a URI?

[09:30:02.0000] <BenMillard>
Krijn, fast changes :)

[09:30:03.0000] <krijn>
:)

[09:30:04.0000] <mookid>
gsnedders: I dont

[09:30:05.0000] <gsnedders>
mookid: That seems backwards to me.

[09:30:06.0000] <mookid>
erm

[09:30:07.0000] <mookid>
LOL

[09:30:08.0000] <gsnedders>
mookid: That's exactly what you're trying to do.

[09:30:09.0000] <Dashiva>
mookid: If browsers can get all representations with options, couldn't the user just right-click and choose pdf himself?

[09:30:10.0000] <BenMillard>
Krijn, my name was already entered and the lines directed at me count is working, but the message telling me to enter my name is still present

[09:30:11.0000] <mookid>
no it's exactly the OPPOSITE

[09:30:12.0000] <mookid>
of what I'm doing

[09:31:00.0000] <mookid>
the representation of /report is pdf

[09:31:01.0000] <Philip`>
krijn: It's strange how sans-serif fonts make sites look much more modern :-)

[09:31:02.0000] <mookid>
putting the representation in the URI

[09:31:03.0000] <mookid>
is creating /report.pdf

[09:31:04.0000] <krijn>
BenMillard: doh, good point

[09:31:05.0000] <BenMillard>
Philip`, Verdana is teh pwn for digital screens.

[09:31:06.0000] <mookid>
Dashiva: yes but some applications have links that need to be to a specific representation

[09:32:00.0000] <gsnedders>
mookid: No, you want to get a specific representation by a URI.

[09:32:01.0000] <mookid>
negotiate

[09:32:02.0000] <mookid>
is the word you're looking for

[09:32:03.0000] <mookid>
:)

[09:32:04.0000] <gsnedders>
mookid: I've said what I want. I want HTML.

[09:32:05.0000] <mookid>
?

[09:32:06.0000] <gsnedders>
mookid: Stop trying to force me to do something different.

[09:32:07.0000] <mookid>
idiot.

[09:32:08.0000] <mookid>
I'm not

[09:32:09.0000] <Philip`>
krijn: \ and " in the nickname field gets backslash-escaped

[09:32:10.0000] <mookid>
you can do what you are doing now

[09:32:11.0000] <mookid>
I just can't do it my way

[09:32:12.0000] <gsnedders>
mookid: Can you stop insulting people?

[09:33:00.0000] <Philip`>
krijn: and \" in the CSS fields gets unescaped

[09:33:01.0000] <gsnedders>
mookid: The answer would appear to be "No".

[09:33:02.0000] <mookid>
well it's frustrating.. you fail to give me a reason why I shouldn't be allowed to do it my way

[09:33:03.0000] <krijn>
Philip`: why would you want those in a nickname

[09:33:04.0000] <gsnedders>
mookid: I've given you plenty of reasons.

[09:33:05.0000] <mookid>
no you havent.

[09:33:06.0000] <Philip`>
krijn: I don't want those in a nickname, I just want to find bugs in your code ;-)

[09:33:07.0000] <krijn>
There are a lot

[09:33:08.0000] <mookid>
Philip`: has he given me any decent reasons?

[09:33:09.0000] <mookid>
cos I missed them

[09:34:00.0000] <BenMillard>
Krijn, I like how much narrower each channel is :)

[09:34:01.0000] <krijn>
Me too

[09:34:02.0000] <Philip`>
mookid: Probably, but I don't know since I was trying to not pay any attention

[09:34:03.0000] <BenMillard>
Krijn, if you allow any sensible separator in the "Nicknames" field, you could remove the note saying how to separate them

[09:34:04.0000] <mookid>
haha cop out :P

[09:35:00.0000] <BenMillard>
;, | etc

[09:35:01.0000] <krijn>
It's in the title as well

[09:35:02.0000] <krijn>
Agreed

[09:35:03.0000] <Philip`>
I still want to know why live.com sends requests with random search keywords in the Referer

[09:36:00.0000] <Philip`>
(I've seen the same on my own sites, and it does appear to be live.com's crawlers sending those Referers)

[09:36:01.0000] <krijn>
Who googled for 'krijn html5'? :\

[09:36:02.0000] <BenMillard>
Krijn, I also like how you use empty <td> instead of <td>0

[09:36:03.0000] <krijn>
I don't want anything to do with this sinking ship!

[09:36:04.0000] <krijn>
(Trying to make Last Week again here)

[09:37:00.0000] <mookid>
this is a prime example of why your group is percieved with such contempt:

[09:37:01.0000] <mookid>
17:33 < gsnedders> mookid: I've said what I want. I want HTML.

[09:38:00.0000] <gsnedders>
mookid: Well, I've specified in my UA what format I want it to get if possible when it makes an HTTP request.

[09:39:00.0000] <gsnedders>
mookid: I don't want any page to change my preference without my consent when my preferred choice is available.

[09:39:01.0000] <mookid>
your browsers preferences are */*

[09:39:02.0000] <mookid>
when it comes down to it

[09:39:03.0000] <mookid>
so..

[09:40:00.0000] <gsnedders>
mookid: No, my preferences are given through quality values

[09:40:01.0000] <gsnedders>
mookid: The fact that I will accept anything is a different question. I've given a preference.

[09:40:02.0000] <BenMillard>
Henny Swan (Opera) on ActiveX in Google Chrome: http://www.iheni.com/dont-go-spoiling-the-party-google-chrome-to-support-activex-in-korea/

[09:40:03.0000] <mookid>
so your browser doesn't have a */* in its default Accept header

[09:40:04.0000] <mookid>
?

[09:40:05.0000] <gsnedders>
mookid: It does.

[09:40:06.0000] <mookid>
well then

[09:40:07.0000] <gsnedders>
mookid: But it prefers other MIME types.

[09:40:08.0000] <mookid>
uyeah...

[09:40:09.0000] <gsnedders>
mookid: I've sent a preference. The fact that I accept anything is irrelevant.

[09:41:00.0000] <gsnedders>
mookid: quality values exist for a reason.

[09:41:01.0000] <mookid>
what if a link in my application is to a sepcific representation (content type) of a resource

[09:41:02.0000] <mookid>
(URI)

[09:41:03.0000] <gsnedders>
mookid: Why do you want to override my preference?

[09:42:00.0000] <mookid>
well seen as though your preference includes everything

[09:42:01.0000] <mookid>
I think it's safe to override it in the markup

[09:42:02.0000] <gsnedders>
No, my preference doesn't.

[09:42:03.0000] <mookid>
yeah

[09:42:04.0000] <annevk3>
krijn, "Your nicknames" doesn't specify a format for entering multiple

[09:42:05.0000] <mookid>
YOUR PREFERENCE IS IRRELEVANT FOR THAT LINK

[09:42:06.0000] <gsnedders>
LOOK AT QUALITY VALUES. I HAVE A PREFERENCE.

[09:42:07.0000] <mookid>
ITS PART OF THE APPLICATION

[09:42:08.0000] <gsnedders>
mookid: No, it isn't.

[09:42:09.0000] <mookid>
HOW IS THAT ANY DIFFERENC

[09:42:10.0000] <mookid>
IF ITS IN THE URI

[09:42:11.0000] <krijn>
annevk3: check title=""

[09:42:12.0000] <mookid>
IF ITS IN THE URI ITS EXACTLY THE SAME 'PROBLEM'

[09:42:13.0000] <krijn>
That's what HTML5 says I should use :)

[09:42:14.0000] <gsnedders>
mookid: If I request a URI and it has the format I prefer, I REALLY WANT THAT FORMAT.

[09:43:00.0000] <mookid>
no you dont

[09:43:01.0000] <gsnedders>
mookid: If I request a URI and it doesn't have the format I prefer in it, THEN GIVE ME SOMETHING.

[09:43:02.0000] <BenMillard>
Krijn, if people use | in their names, I guess that can't be a seperator

[09:43:03.0000] <mookid>
you look at Accept header as fixed

[09:43:04.0000] <mookid>
it's not

[09:43:05.0000] <gsnedders>
mookid: Yes, I do want what I prefer. Honestly.

[09:43:06.0000] <mookid>
it's just a header

[09:43:07.0000] <krijn>
Nope, only ,

[09:43:08.0000] <krijn>
Which is okay enough

[09:43:09.0000] <BenMillard>
krijn, ; as well?

[09:43:10.0000] <gsnedders>
mookid: It's not fixed. But I've said what I prefer already. Stop trying to make me change what I prefer.

[09:43:11.0000] <BenMillard>
krijn, space should be supported as a seperator, I think

[09:44:00.0000] <krijn>
BenMillard: there are literally 5 people using this ;)

[09:44:01.0000] <gsnedders>
mookid: Any changes should be made by an option in my UA, not by a link.

[09:44:02.0000] <BenMillard>
krijn, oh lol

[09:44:03.0000] <mookid>
gsnedders: I'm not changing what you prefer I'm telling you the appropriate preference for that request

[09:45:00.0000] <gsnedders>
mookid: You're telling me what you think the appropriate preference is for that request, so that I may reconsider my preference.

[09:45:01.0000] <gsnedders>
mookid: By having it as an attribute, it is likely that I will be forced you use what you think is the appropriate preference.

[09:45:02.0000] <krijn>
/me finds out about a new function.. http://krijnhoetmer.nl/irc-logs/whatwg/20081128#l-949

[09:45:03.0000] <krijn>
*feature

[09:46:00.0000] <BenMillard>
krijn, you'll get better alignment down each column by using td { align: right; } and th { text-align: center; }

[09:46:01.0000] <gsnedders>
mookid: Get some browser to implement what you want. Then we might look at it again.

[09:47:00.0000] <Philip`>
HTML needs a <shout> element, which is like <strong> but makes you think the author is getting too emotional

[09:47:01.0000] <krijn>
text-transform: uppercase;

[09:48:00.0000] <KrocCamen>
krijn: beat me to it :P

[09:48:01.0000] <Philip`>
krijn: But it's semantic, not stylistic

[09:48:02.0000] <KrocCamen>
or caps-lock: cruise-control;

[09:48:03.0000] <krijn>
Philip`: for the default stylesheet :)

[09:48:04.0000] <krijn>
voice: crybaby;

[09:48:05.0000] <Philip`>
krijn: I don't see the new feature there :-/

[09:48:06.0000] <BenMillard>
Krijn, table rows are 16px tall but list items are 14px tall. Making them both 14px tall would be more consistent.

[09:49:00.0000] <annevk3>
 http://mykanjo.co.uk/ toally fails in Opera

[09:49:01.0000] <hsivonen>
annevk3: it's all Flash

[09:49:02.0000] <annevk3>
that explains things

[09:50:00.0000] <Philip`>
krijn: Oh, yes I do, my browser had just cached a five hour old version which didn't have the last nine hundred lines of productive discussion

[09:50:01.0000] <krijn>
productive should be in <strong> there

[09:50:02.0000] <krijn>
annevk3: you're really missing something :)

[09:50:03.0000] <Philip`>
I think that's a use case for <sarcasm>, actually

[09:50:04.0000] <annevk3>
krijn, nice

[09:51:00.0000] <annevk3>
(the new irc-logs)

[09:53:00.0000] <BenMillard>
annevk3, glad they are proving popular (krijn took a subset of my ideas and figured out lots of other changes himself): http://projectcerbera.com/!dev/irc-logs/front

[09:53:01.0000] <Philip`>
gsnedders: Your problem could be satisfied if @accept accepted a q argument, and the 0-1 limit was removed, so people would write <a href=... accept=application/pdf;q=100.0> and you could change your UA preferences to have text/html;q=101.0 and then you'd get what you want

[09:53:02.0000] <gsnedders>
Philip`: Where are we meeting up, BTW?

[09:53:03.0000] <Philip`>
and if the author really wants you to still get the PDF, they can set q=1000.0, and if you really want the HTML then you can set q=1001.0

[09:54:00.0000] <Philip`>
gsnedders: I have no idea, and I can't even remember when it's meant to be

[09:54:01.0000] <krijn>
Also truncated my referrer database a lot, which made it faster

[09:54:02.0000] <jmb>
Philip`: and thus have an arms race to UINT_MAX? :)

[09:54:03.0000] <Philip`>
jmb: Of course not, that'd be silly - use DBL_MAX

[09:54:04.0000] <jmb>
aha. my mistake

[09:56:00.0000] <gsnedders>
Philip`: 20081208T190000Z/PT5M

[09:57:00.0000] <Philip`>
gsnedders: Could you say that in English?

[09:57:01.0000] <krijn>
svl: wow, you too! ;)

[09:57:02.0000] <gsnedders>
Philip`: :)

[09:57:03.0000] <Philip`>
I'm too lazy to separate digits mentally :-p

[09:57:04.0000] <gsnedders>
Philip`: 2008-12-08T19:00:00Z/PT5M

[09:57:05.0000] <gsnedders>
:P

[09:57:06.0000] <Philip`>
gsnedders: Ah, good, so it's not today

[09:58:00.0000] <gsnedders>
Philip`: No, it's not.

[09:58:01.0000] <svl>
well, y'know, have to use it if I'm going to try to find ways to break it. ;P

[09:58:02.0000] <BenMillard>
Krijn, when an entire column would be empty, perhaps you could avoid generating that column? for example, #wai-aria and #webapps have no lines marked as important, so that column has nothing in it

[09:58:03.0000] <BenMillard>
oh, same for #css

[09:59:00.0000] <gsnedders>
What is some good styling for lines to/from you?

[09:59:01.0000] <gsnedders>
I wants suggestions!

[09:59:02.0000] <svl>
krijn: why isn't the number under ? linked to the first line in which you were addressed?

[09:59:03.0000] <krijn>
That's on my todo

[09:59:04.0000] <svl>
idem for ! I guess...

[09:59:05.0000] <BenMillard>
gsnedders, from me: background: #efe;

[09:59:06.0000] <krijn>
Yeh

[10:00:00.0000] <BenMillard>
gsnedders, at me (using CSS injection sneakyness): background: #eeeae0; } li.flagged { background: #bcc; } li:target { background: #bdd; } ol { font: 0.625em verdana, sans-serif;

[10:00:01.0000] <krijn>
Oh, so sneaky ;)

[10:00:02.0000] <Philip`>
gsnedders: Usually at that kind of time I will be located somewhere on the route between the CL and just-above-King's and around Midsummer Common, so anywhere on that route would be easiest for me :-)

[10:00:03.0000] <krijn>
Another proof of user stylesheets failing

[10:00:04.0000] <gsnedders>
Philip`: I'm staying in Memorial Court, FWIW

[10:01:00.0000] <BenMillard>
krijn, I have a similar user stylesheet for HTML5 (Verdana text, Rockwell headings, smaller text sizes, no vertical gaps between items in lists, etc)

[10:01:01.0000] <gsnedders>
/me waits for Philip` to ask where that is

[10:01:02.0000] <gsnedders>
BenMillard: That's uselessly light :)

[10:01:03.0000] <Philip`>
gsnedders: I asked Google where it was, and got a handy map :-p

[10:02:00.0000] <BenMillard>
gsnedders, on a laptop screen that may well be true. on this CRT it passes the "squint test"

[10:03:00.0000] <BenMillard>
gsnedders, maybe you'd prefer some WHATWG colours

[10:03:01.0000] <Philip`>
gsnedders: Looks like it's a place I've walked past pretty much every day for the past three years

[10:03:02.0000] <Philip`>
/me hadn't quite noticed it during that period

[10:03:03.0000] <gsnedders>
Philip`: :)

[10:03:04.0000] <krijn>
lol @ referrers

[10:03:05.0000] <BenMillard>
gsnedders, from me/to me using the bright green left border colour from the green boxes in HTML5: background: #9f9;

[10:04:00.0000] <gsnedders>
krijn: Wow, it says "Welcome back, Geoffrey." :D

[10:04:01.0000] <krijn>
It's magic!

[10:04:02.0000] <krijn>
Omg!

[10:04:03.0000] <gsnedders>
krijn: Database of IRC nicks to names?

[10:04:04.0000] <krijn>
Yeah

[10:04:05.0000] <BenMillard>
Krijn, vertical gaps have gotten even bigger. :(

[10:05:00.0000] <BenMillard>
krijn, numbers have been centered instead of right-aligned

[10:05:01.0000] <Philip`>
Hmph, your referers are boringly XSS-proof

[10:05:02.0000] <annevk3>
ah, an actual database

[10:05:03.0000] <annevk3>
I thought it just uppercased my first nick

[10:05:04.0000] <krijn>
Nah

[10:06:00.0000] <krijn>
Philip`: I know a tiny bit, thank you

[10:06:01.0000] <Philip`>
krijn: "It's now: Fri, 28 Nov 2008 19:14:15 +0100 (CET)" - oh no it's not

[10:06:02.0000] <BenMillard>
krijn, the addition of green for headings and heading-like text is nice, imho

[10:07:00.0000] <BenMillard>
Philip`, that seems a few minutes fast to me, too

[10:07:01.0000] <BenMillard>
but dude, why Arial? :(

[10:07:02.0000] <krijn>
Hrhr

[10:08:00.0000] <Philip`>
Hmm, I think broke view-source on the logs page in Opera

[10:09:00.0000] <BenMillard>
krijn, title="Number of lines flagged as important" seems unnecessarily verbose compared to title="Important lines"

[10:15:00.0000] <BenMillard>
krijn, th abbr { display: block; } makes it easier to put the mouse over shortened table headers

[10:21:00.0000] <krijn>
Fixed & fixed

[10:22:00.0000] <BenMillard>
Krijn, cool :)

[10:23:00.0000] <BenMillard>
what about vertical gaps? table rows are now 18px high (which seems far too much) when list items are 14px high

[10:23:01.0000] <krijn>
(And nicks are case sensitive btw :)

[10:23:02.0000] <BenMillard>
krijn, bah

[10:23:03.0000] <krijn>
Yeah

[10:24:00.0000] <krijn>
Fixed

[10:34:00.0000] <krijn>
Enough for today

[10:34:01.0000] <BenMillard>
krijn, thanks for all the updates! it's more usable and prettier than before 

[10:35:00.0000] <krijn>
Np, I hope all 4 users like it ;)

[10:39:00.0000] <BenMillard>
krijn, do you have AWstats or Analog or some stats package running on the server?

[10:57:00.0000] <annevk3>
"Mr. Obama, however, seems intent on pulling the office at least partly into the 21st century on that score; aides said he hopes to have a laptop computer on his desk in the Oval Office, making him the first American president to do so." o_O

[10:57:01.0000] <annevk3>
http://www.nytimes.com/2008/11/16/us/politics/16blackberry.html

[11:02:00.0000] <Philip`>
He'll just end up wasting all his time on IRC

[11:03:00.0000] <gavin>
it would be interesting to know more about how they will manage IT security

[11:03:01.0000] <gavin>
I bet some people aren't happy with the idea

[11:15:00.0000] <Philip`>
I would suppose the government already has some experience with needing IT security

[11:15:01.0000] <Philip`>
It's not exactly a unique new challenge

[11:17:00.0000] <gavin>
computers in the oval office is, I think

[11:18:00.0000] <gavin>
though presumably there are other computers in the white house that need to be just as secure, I guess

[11:18:01.0000] <Lachy>
what would be wrong with having a computer in the oval office? I assumed that would have been a fairly common thing for the president to have in there

[11:19:00.0000] <gavin>
apparently not!

[11:19:01.0000] <Dashiva>
The President has some rather extreme documentation requirements

[11:19:02.0000] <Dashiva>
And security on top of that

[11:20:00.0000] <Lachy>
so? Surely the president uses a computer at least some of the time. What difference would it make where it's located?

[11:21:00.0000] <Dashiva>
He's got vice presidents for that :)

[11:22:00.0000] <Philip`>
How does he upload his videos to Youtube if he doesn't have computer access?

[11:26:00.0000] <gavin>
Lachy: the oval office is one of a handful of places where highly secretive discussions are held

[11:26:01.0000] <gavin>
bringing any type of electronic device into such places is not something that's taken lightly

[11:27:00.0000] <gavin>
let alone a networked computer

[11:29:00.0000] <Lachy>
Philip`, I guess one of his staff does it for him


2008-11-29
[17:07:00.0000] <Philip`>
  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND

[17:07:01.0000] <Philip`>
22726 philip    20   0 2132m 711m  21m S    1 35.3 553:42.56 opera

[17:08:00.0000] <Philip`>
(This is with a total of one tab open, and that one is empty)

[17:08:01.0000] <Philip`>
I think Opera might have a memory leak :-(

[17:08:02.0000] <Philip`>
Steps to reproduce: 1) Browse the web for two weeks.

[17:14:00.0000] <BenMillard>
Al Gilman has made a survey about how to organise WAI-ARIA integration, which I've just filled in: http://lists.w3.org/Archives/Public/public-html/2008Nov/0580.html

[19:42:00.0000] <Hixie>
hsivonen: initial containing block is neither html nor body, it's a box that contains html

[00:41:00.0000] <gsnedders>
Philip`: That's a bit vague. Can you be more specific? What sites are you meant to frequently visit throughout the two week period, for example?

[00:41:01.0000] <gsnedders>
:P

[00:47:00.0000] <annevk3>
anyone any idea where the text in "Typical default display properties" in http://www.w3.org/html/wg/markup-spec/#audio comes from?

[00:48:00.0000] <annevk3>
maybe Mike just scraped a UA style sheet?

[01:00:00.0000] <annevk3>
ah, it's from WebKit

[01:05:00.0000] <roc>
yeah, that was weird

[01:15:00.0000] <annevk3>
Hixie, in the IRC logs you commented somewhere on the usage of <keygen>, did you read http://lists.whatwg.org/pipermail/whatwg-whatwg.org/attachments/20080714/07ea5534/attachment.txt ?

[01:18:00.0000] <krijn>
BenMillard: yes, about that stats package

[01:26:00.0000] <Hixie>
annevk3: right... where's the private key used?

[01:27:00.0000] <Philip`>
/me assumed it was used to identify the client in SSL connections

[01:28:00.0000] <gsnedders>
time to return to SETTLERS2

[01:28:01.0000] <gsnedders>
That and Civ2 are two awesome procrastination tools.

[01:42:00.0000] <hsivonen>
Hixie: oh. If the Document object itself cooresponds to the initial containing block, that makes the assertions I saw more interesting

[01:42:01.0000] <hsivonen>
s/oor/orr/

[01:43:00.0000] <Hixie>
i don't know if i'd phrase it such that the document itself is the icb, but i guess if you compare the dom to the render tree you would find a document in the dom in a position analogous to where the icb is

[01:45:00.0000] <hsivonen>
I somehow managed to get the frame constructor into a state where it asserted about null icb

[03:24:00.0000] <gsnedders>
/me stops after two hours of procrastination

[03:25:00.0000] <Hixie>
/me rambles on www-tag

[03:25:01.0000] <Hixie>
i wonder what to do about select.value

[03:25:02.0000] <Hixie>
we've had several requests

[03:25:03.0000] <Hixie>
its back compat story sucks

[03:25:04.0000] <Hixie>
er, <select value=""> i should say

[03:26:00.0000] <Hixie>
oh well

[03:26:01.0000] <Hixie>
nn

[03:41:00.0000] <Lachy>
Hixie, what is <select value=""> supposed to do?

[03:46:00.0000] <gsnedders>
hsivonen: Why can't you set the parser to XML when served as text/html on validator.nu?

[03:47:00.0000] <hsivonen>
gsnedders: you can

[03:47:01.0000] <gsnedders>
hsivonen: http://validator.nu/?=&doc=http%3A%2F%2Flabs.utopian.net%2Fblogs%2Fsimplish-demo%2F&schema=http%3A%2F%2Fs.validator.nu%2Fxhtml10%2Fxhtml-transitional.rnc+http%3A%2F%2Fs.validator.nu%2Fxhtml10%2Fxhtml.sch++http%3A%2F%2Fc.validator.nu%2Fall-html4%2F&parser=xml

[03:48:00.0000] <hsivonen>
gsnedders: you need to check the box labeled "RFC 3023 Sucks"

[03:48:01.0000] <hsivonen>
gsnedders: http://validator.nu/?doc=http%3A%2F%2Flabs.utopian.net%2Fblogs%2Fsimplish-demo%2F&schema=http%3A%2F%2Fs.validator.nu%2Fxhtml10%2Fxhtml-transitional.rnc+http%3A%2F%2Fs.validator.nu%2Fxhtml10%2Fxhtml.sch++http%3A%2F%2Fc.validator.nu%2Fall-html4%2F&parser=xml&laxtype=yes

[03:48:02.0000] <gsnedders>
hsivonen: Oh, that's counter-inuative

[03:49:00.0000] <hsivonen>
gsnedders: you aren't supposed to do that. you are supposed to fix your server config :-)

[03:49:01.0000] <gsnedders>
hsivonen: :)

[03:49:02.0000] <gsnedders>
hsivonen: But I said I want an XML parser dangnamit!

[03:50:00.0000] <hsivonen>
gsnedders: in that case, the Accept header changes, but the server still sent text/html

[03:50:01.0000] <gsnedders>
hsivonen: ah

[04:03:00.0000] <gsnedders>
/me +1s Hixie's email to www-tag

[04:03:01.0000] <hsivonen>
indeed

[05:34:00.0000] <hsivonen>
http://www.hixie.ch/tests/adhoc/html/parsing/encoding/055.html is interesting

[06:33:00.0000] <takkaria>
yay for ascii art

[06:45:00.0000] <gsnedders>
jgraham, annevk3: can you update your copies of Anolis on your web services?

[06:46:00.0000] <gsnedders>
after that, Hixie can regen the spec to fix the bug Lachy reported yesterday

[08:52:00.0000] <gsnedders>
Lachy_: does <http://hg.gsnedders.com/anolis/raw-file/b1648aaa8c8f/README.html#id-generation> seem right?

[10:58:00.0000] <Philip`>
/me gets a new phone, which appears to have NetFront 3.3, and finds that it seemingly attempts to render the SVG on http://canvex.lazyilluminati.com/ but only manages to draw a white rectangle with a black outline of the text

[11:00:00.0000] <jgraham>
gsnedders: I will do that at some convenient moment

[11:00:01.0000] <jgraham>
Which may or may not be this evening

[11:07:00.0000] <gsnedders>
jgraham: :)

[11:57:00.0000] <Hixie>
Lachy: <select value=""> would be a simple replacement for <select selected>

[12:01:00.0000] <gsnedders>
Hixie: I've fixed the bug in Anolis causing the bug on 4.10.4.1.2 in the TOC; waiting on jgraham and annevk3 to update their web services :P

[12:02:00.0000] <Hixie>
cool

[12:02:01.0000] <Hixie>
hey he went away already

[12:08:00.0000] <Lachy>
Hixie, I assume you mean <option selected>. Do any browsers support <select value=""> already?

[12:08:01.0000] <Lachy>
if not, what's the point of it?

[12:08:02.0000] <Hixie>
er yes, <option selected>

[12:08:03.0000] <Hixie>
he point would be making scripts that output <select>s far easier

[12:09:00.0000] <Hixie>
it's not supported anywhere

[12:09:01.0000] <gsnedders>
Hixie: Yeah, I don't hang around :)

[12:09:02.0000] <Lachy>
ok, so you could do <select value="b"><option>a<option>b<option>c</select>

[12:10:00.0000] <Lachy>
and if value="x" doesn't match one of the options, then would it fallback to selecting the first option, as if it wasn't specified?

[12:15:00.0000] <Philip`>
http://www.ipiao.com.cn/ does <select size="1" name="to"><script>City_Default="PEK"</script><SCRIPT language=javascript src="/js/flightcity.js"></SCRIPT></select> which is the kind of thing that could be much simpler with <select value="PEK">

[12:17:00.0000] <Lachy>
I suppose mapping HTMLSelectElement.value to the new value attribute would be mostly backwards compatible.

[12:17:01.0000] <Lachy>
however, there would be an issue when there's a <option selected> and a conflicting value=""

[12:27:00.0000] <Hixie>
that's easy to sort out

[12:29:00.0000] <deltab>
wouldn't it be defaultValue?

[12:29:01.0000] <Hixie>
probably

[14:28:00.0000] <Philip`>
Hixie: <code title="attr-script-language"></code> in recent checkin

[14:32:00.0000] <Hixie>
oops

[14:32:01.0000] <Hixie>
thanks fixed

[14:41:00.0000] <gsnedders>
Lachy: can you look over http://hg.gsnedders.com/anolis/raw-file/b1648aaa8c8f/README.html#id-generation?

[14:55:00.0000] <Philip`>
/me thinks people should write spec algorithms without unnecessarily relying on goto, since real programmers discovered that was a bad idea decades ago :-p

[14:56:00.0000] <Hixie>
yeah i've been trying to move away from goto

[14:56:01.0000] <BenMillard>
jgraham, how come the cell "Indicates explicit relationship between this document and other resources. As such there are many good uses!" gets none of the 3 header cells from the 1st level applied to it here? Bug in the code or shortcoming of the algorithm? http://james.html5.org/tables/table_inspector.py?input_type=type_uri&uri=http%3A%2F%2Fkeryx.se%2Fresources%2Fhtml-elements.xhtml&source=&algorithm=smartheaders

[14:56:02.0000] <Hixie>
Philip`: unfortunately english doesn't have well-defined structured flow constructs

[14:57:00.0000] <Philip`>
Inventing some syntax for 'while' and nested 'if' seems like it should be easier to read than these "11. Jump back to step 8." things

[14:58:00.0000] <BenMillard>
it has nested lists and says "for each foo, repeat these steps" or similar in some places, which seems good

[14:58:01.0000] <Hixie>
i've definitely stopped using "jump back to step N"

[14:59:00.0000] <BenMillard>
Roy Fielding gave feedback on the structure of algorithms in HTML5, IIRC

[14:59:01.0000] <BenMillard>
something about multiple exit points?

[14:59:02.0000] <Hixie>
uri?

[14:59:03.0000] <BenMillard>
it was some time ago...I'll try to find it

[14:59:04.0000] <Philip`>
Multiple entry points are far nastier than multiple exit points

[15:01:00.0000] <BenMillard>
Philip`, maybe that was it...still looking

[15:01:01.0000] <Dashiva>
I suggested Hixie should use labelled loops a while back, since plain English for 'break' and 'continue' are also confusing with nested loops :)

[15:01:02.0000] <Philip`>
(Most people seem perfectly happy to sprinkle 'return' all through their code, since it's easy to see the exits when you're reading through the code, whereas unrestrained gotos can jump anywhere and you can't see them)

[15:01:03.0000] <BenMillard>
aha

[15:02:00.0000] <Philip`>
(INTERCAL solves that last problem by using COME FROM instead of GOTO, so you don't have to worry about invisible entry points in your code)

[15:02:01.0000] <Dashiva>
I was just bout to mention intercal :)

[15:02:02.0000] <BenMillard>
search for "entry" here: http://lists.w3.org/Archives/Public/public-html/2007Aug/0980.html

[15:02:03.0000] <Hixie>
BenMillard: regarding that table, looks like a bug in the tool; as far as i can tell the heading cells should apply per spec

[15:03:00.0000] <Hixie>
Dashiva: basically i do labelled loops now

[15:03:01.0000] <BenMillard>
Hixie, that link is to the e-mail from Roy, August 2007

[15:03:02.0000] <BenMillard>
Hixie, yeah that may well be the case

[15:04:00.0000] <Lachy>
gsnedders, the algorithm is significantly better than it was. Most problems with it before have been fixed

[15:05:00.0000] <BenMillard>
Hixie & jgraham, in the HTML5 mode the 3 1st-level of header cells are applied. However, the "Information" header in the 2nd level is not applied: http://lists.w3.org/Archives/Public/public-html/2007Aug/0980.html

[15:05:01.0000] <Lachy>
But I have a problem with the way the variable i is incremented

[15:06:00.0000] <BenMillard>
Hixie, my reading of the HTML5 algorithm is that both levels should be applied to the data cell, but the header cells don't apply to each other?

[15:06:01.0000] <Hixie>
BenMillard: i don't really understand that e-mail

[15:06:02.0000] <Hixie>
BenMillard: yeah, i think that's right

[15:07:00.0000] <Lachy>
gsnedders, to illustrate, consider a page which contains these elements throughout the page: <span title="foo"/>...<span title="bar"/>...<span title="bar"/>...<span title="foo"/>

[15:08:00.0000] <Lachy>
per the algorithm, the IDs will be: foo, bar, bar-1, foo-2

[15:08:01.0000] <BenMillard>
Hixie, me neither...but I remembered it talked about entry points and exit points of algorithms, so it seemed on-topic :)

[15:08:02.0000] <Hixie>
BenMillard: :-)

[15:12:00.0000] <Hixie>
BenMillard: turns out i replied: http://lists.w3.org/Archives/Public/public-html/2007Aug/0993.html

[15:12:01.0000] <Hixie>
looks like he ignored my e-mail


2008-11-30
[16:05:00.0000] <yecril71>
Maybe English is better for describing flow charts, but HTML is better for blocks of code.

[16:05:01.0000] <yecril71>
Such a block of code can become a nested list.

[16:05:02.0000] <Lachy>
gsnedders, ignore my last comments. I now see that it resets i to 0 for every new id and keeps incrementing it until it reaches "foo-#" where # is a unique number, so my previous interpretation was wrong

[16:06:00.0000] <yecril71>
HTML is not a programming language; if it were one, it would be called HTPL instead.

[16:06:01.0000] <yecril71>
HTML is a markup language.

[16:07:00.0000] <BenMillard>
really? who knew! :D

[16:07:01.0000] <yecril71>
Calogero apparently did not.

[16:10:00.0000] <gpy>
its a markup language...

[16:14:00.0000] <Philip`>
yecril71: Technologies are not defined by their names

[16:15:00.0000] <Philip`>
e.g. Javascript has very little to do with Java - you have to look at more than the name to work out what something really is

[16:17:00.0000] <yecril71>
And, moreover, Java has very little to do with Java :-)

[16:17:01.0000] <yecril71>
They are just codenames.

[16:18:00.0000] <Philip`>
/me unexpectedly discovers that window.location.replace(/pattern/, 'thing') really doesn't do what he intended

[16:18:01.0000] <Dashiva>
This is where I go "Don't do that"

[16:18:02.0000] <yecril71>
But the choice of the words in the term HTML is very appropriate.

[16:19:00.0000] <Dashiva>
You should always use location.href because otherwise you'll get bit in the tail by things like that

[16:19:01.0000] <Philip`>
Dashiva: The problem is that I don't know I can use location.href without spending ages (multiple seconds) trying to find documentation that tells me I can use that

[16:20:00.0000] <yecril71>
location is a DOM object, it does not replace

[16:20:01.0000] <Philip`>
Dashiva: but now I know, so thanks :-)

[16:20:02.0000] <Hixie>
the name "HTML" says very little about what HTML is

[16:21:00.0000] <Dashiva>
Philip`: The problem is that you were taught (or otherwise learned) to use just location in the first place

[16:21:01.0000] <yecril71>
But it covers the basics.

[16:23:00.0000] <yecril71>
If the browsers tosses the options, the script cannot consider the options between.

[16:23:01.0000] <yecril71>
That means the browser really cannot toss the options for presentation only.

[16:24:00.0000] <Philip`>
Dashiva: I've never been taught HTML or JS - I just learned through copy-and-paste :-)

[16:25:00.0000] <Philip`>
The web is large enough that somebody has already solved any problem you might have, so the trick is just to find that and copy it

[16:27:00.0000] <Dashiva>
Bad practice breeds bad practice, yes

[16:27:01.0000] <yecril71>
Unless you are on an "original work" contract.

[16:29:00.0000] <yecril71>
In that case the trick is to find that and understand it.

[16:29:01.0000] <Philip`>
/me attempts to tweak his vastly-incomplete localStorage test framework so it can handle tests that require the user to exit and restart their browser, as automatically as possible, which seems to be quite reasonable as long as there's an unlimited number of domain names so each test can run independently in parallel...

[16:30:00.0000] <Philip`>
yecril71: No, in that case the trick is to find some code and copy it and replace the variable names and reformat so nobody knows it wasn't yours

[16:30:01.0000] <yecril71>
This works only if you copy from one source.

[16:31:00.0000] <yecril71>
Otherwise it gets too inconsistent.

[16:38:00.0000] <yecril71>
"Select one" as an invalid option looks rather like a generic help baloon.

[16:39:00.0000] <yecril71>
This is rather obvious and the browser can do it itself, without HTML.

[16:39:01.0000] <yecril71>
Anything more specific should be in the label, or perhaps the title attribute.

[16:44:00.0000] <yecril71>
I do not think CSS can do anything about inheriting attributes that belong to HTML.

[16:45:00.0000] <yecril71>
A mechanism that would allow this to HTML is closer to templates.

[17:52:00.0000] <wilhelm>
Is there a Perl version of html5lib yet?

[17:54:00.0000] <Philip`>
wilhelm: There's http://suika.fam.cx/www/markup/html/whatpm/readme though I have no idea how well it works or whether it works at all

[17:55:00.0000] <wilhelm>
Thanks.

[19:42:00.0000] <Lachy>
Hixie, you might find this useful http://hymn-project.org/forums/viewtopic.php?p=19078#19078

[19:43:00.0000] <Hixie>
cool

[19:44:00.0000] <Hixie>
my hacked up 7.x install is still working great

[19:44:01.0000] <Hixie>
i'll wait til something breaks before updating it

[19:44:02.0000] <Lachy>
which version of itunes are you running?

[19:45:00.0000] <Hixie>
latest

[19:46:00.0000] <Hixie>
i get a weird error message about "constant sized" when decrypting, but the files are still drm-free and work, so...

[19:46:01.0000] <Lachy>
8.0.2? I have that, but my backed up 8.0.1 keys don't work with songs downloaded with 8.0.2.

[19:46:02.0000] <Hixie>
oh haven't tried songs, only video

[19:46:03.0000] <Hixie>
i'm using the 7.x keys

[19:47:00.0000] <Lachy>
ok. I haven't tried video. Only the free song of the week

[19:47:01.0000] <Lachy>
I don't have my 7.x keys any more

[19:47:02.0000] <Hixie>
where does one get that?

[19:47:03.0000] <Lachy>
from the iTunes store home page. There's an ad for it in the middle

[19:47:04.0000] <Hixie>
aah yes

[19:49:00.0000] <Hixie>
/me downloads it and runs dedrmify.sh

[19:50:00.0000] <Hixie>
decrypting...

[19:53:00.0000] <Lachy>
did it work?

[19:55:00.0000] <Hixie>
restarting iTunes...

[19:56:00.0000] <Hixie>
(sorry my script takes a while because it scans my entire collection for new files)

[19:56:01.0000] <Lachy>
oh

[19:56:02.0000] <Hixie>
hm, no error message, but the file is still protected according to itunes

[19:56:03.0000] <Hixie>
it did say it had decrypted it

[19:57:00.0000] <Lachy>
try running requiem on the file directly and see if it outputs an error messae

[19:57:01.0000] <Hixie>
/me tries copying it to an unauthorised machine

[19:57:02.0000] <Lachy>
or try playing it with VLC or some other player

[20:00:00.0000] <Hixie>
it claims it's protected but it doesn't seem to be

[20:01:00.0000] <Hixie>
this computer isn't authorised

[20:01:01.0000] <Hixie>
and it's playing fine

[20:01:02.0000] <Lachy>
maybe it claims it's protected based on the file extension

[20:02:00.0000] <Hixie>
maybe

[23:12:00.0000] <BenMillard>
hsivonen, this survey seems relevant given your work figuring out how ARIA can be integrated with HTML5: http://www.w3.org/2002/09/wbs/32212/axs-API-impl-Org/

[02:18:00.0000] <mookid>
http://www.b-list.org/weblog/2008/nov/29/multiresponse/

[03:00:00.0000] <gsnedders>
Anolis 1.0RC3 available now

[03:00:01.0000] <gsnedders>
(nothing has changed since yesterday)

[03:43:00.0000] <nlogax>
hi. looking at html 5 and playing around a bit. i noticed that html5 "valid global date and time strings" are not understood by the js date object. is there a reason for this? i was trying to make nice dates like everyone else, "2 hours ago" etc, using the datetime attribute value. maybe i'm abusing it? :)

[04:07:00.0000] <Hixie>
nlogax: datetime on which element?

[04:09:00.0000] <nlogax>
Hixie: the datetime attribute on <time>

[04:10:00.0000] <nlogax>
with values like "2006-09-24T05:00-07:00"

[04:11:00.0000] <Hixie>
you'll be able to get a Date object off it using t.date+t.time, once it's implemented

[04:11:01.0000] <Hixie>
though maybe we should offer a simple way of getting the exact corresponding Date

[04:15:00.0000] <nlogax>
i see. i understand most of it is not implemented yet, but with the focus on backward compatibility i wasn't sure if it was on purpose or not to use a format Date doesn't understand :)

[04:16:00.0000] <Hixie>
it's the format HTML4 used for datetime="" on <ins> and <del>

[04:16:01.0000] <Hixie>
(ISO8601)

[04:16:02.0000] <nlogax>
oh.. i didn't even know about that

[04:17:00.0000] <Hixie>
i wouldn't be surprised if js didn't change to support it at some point, it's a pretty standard format

[04:17:01.0000] <Hixie>
but anyway, there'll be DOM APIs to give you a Date in due course

[04:17:02.0000] <Hixie>
they're just not implemented yet

[04:18:00.0000] <Hixie>
bed time for me now

[04:18:01.0000] <Hixie>
nn

[04:18:02.0000] <nlogax>
gn, thanks for the help!

[05:46:00.0000] <erlehmann>
any idea how to mark up the author of spoken words ? cite seems to be insufficient.

[08:38:00.0000] <hsivonen>
/me wonders about the validator perf impact of r2486 that requires scanning text content for quotes

[08:38:01.0000] <hsivonen>
are all quote characters on the BMP?

[08:40:00.0000] <Dashiva>
What if the quoted text itself begins with a quote character?

[08:40:01.0000] <hsivonen>
Good question.

[08:41:00.0000] <hsivonen>
Hixie: the spec now effectively prohibits using <q> for outer and punctuation for inner quotes when the inner quote touches the boundary of the outer quote

[08:42:00.0000] <gsnedders>
/me doesn't get what hsivonen means

[08:44:00.0000] <hsivonen>
My favorite passage from novel Foo is <q>‘Quote!’, she said.</q>

[08:44:01.0000] <gsnedders>
ah

[08:47:00.0000] <hsivonen>
/me wonders if there's a way to find out if a given legacy lock can be opened by the TSA without breaking it

[08:47:01.0000] <gsnedders>
hsivonen: Try taking it into the US?

[08:48:00.0000] <hsivonen>
gsnedders: well, I've taken the luggage in questing to the U.S. twice already, but I thought they changed their policies recently

[08:48:01.0000] <hsivonen>
s/questing/question/

[08:48:02.0000] <gsnedders>
/me is generally trying to avoid US immigration (by not going to the US)

[08:49:00.0000] <hsivonen>
/me wouldn't want to buy a new luggage only to get the right logos on the locks

[08:49:01.0000] <gsnedders>
/me wonders whether to be pedantic about the "a" that hsivonen used

[08:50:00.0000] <hsivonen>
s/new/new piece of/

[08:50:01.0000] <gsnedders>
(luggage is uncountable, and therefore takes no article directly)

[08:52:00.0000] <hsivonen>
the legacy lock is really simple and has a number on it, the purpose of which I infer to be that the number tells the authorities which passepartout to use

[08:52:01.0000] <hsivonen>
but it doesn't have the new TSA lock logo

[08:59:00.0000] <hsivonen>
/me assumes that the TSA can pick the simple legacy Samsonite latch locks

[09:02:00.0000] <Philip`>
/me finds it peculiar that 'easy to open without the key/code' is a marketable feature for locks

[09:06:00.0000] <hsivonen>
Philip`: it depends on what your threat model is.

[09:07:00.0000] <hsivonen>
Philip`: are you more concerned with damage done by official authorities or unofficial thieves?

[09:54:00.0000] <Lachy>
Last time I travelled, the lock I used on my luggage was one of those ones you can pick with a paper clip

[09:55:00.0000] <Lachy>
the purpose of it was only to hold the zippers together to stop them accidentally opening.

[09:56:00.0000] <Lachy>
I guess I'm just too trusting of airport staff to not put anything into or take anything out of my luggage

[09:57:00.0000] <svl>
Given the really high theft rates from luggage by baggage handlers at some airports... probably.

[09:59:00.0000] <hsivonen>
I think I'm going to trust that they can pick the latch locks without damaging them and then I'll write the combination of the numeric lock next to the lock so they don't need to break it

[09:59:01.0000] <hsivonen>
I totally don't trust the idea that only good TSA personnel has keys to the TSA locks now

[10:00:00.0000] <Philip`>
Rather than using a numeric lock and telling everyone what the number is, why not just leave it unlocked?

[10:00:01.0000] <gsnedders>
hsivonen: You're assuming they won't break the numeric lock anyway

[10:00:02.0000] <hsivonen>
so one might as well assume that writing the combination of a legacy lock next to the lock is as safe as using a new TSA lock

[10:00:03.0000] <hsivonen>
Philip`: stuff might fall out accidentally

[10:01:00.0000] <Philip`>
hsivonen: You could tie it closed with a piece of string, which is much cheaper to replace if they decide it needs to be forcibly opened

[10:02:00.0000] <Lachy>
does the TSA open all luggage, or do they randomly select bags to check?

[10:03:00.0000] <hsivonen>
Lachy: I think they non-invasively scan them all and open the ones they doubt based on x-ray

[10:03:01.0000] <hsivonen>
or perhaps they open some at random, too

[10:08:00.0000] <hsivonen>
anyway, speaking of going to the U.S.: are there any WHATWG-related events taking place on the week of Dec 8th in the Bay Area?

[10:23:00.0000] <gsnedders>
/me should just totally change his AH computing project

[10:26:00.0000] <Philip`>
/me agrees

[10:29:00.0000] <gsnedders>
biblio of HTML files! :P

[10:30:00.0000] <Philip`>
/me isn't quite sure what gsnedders is talking about

[10:30:01.0000] <gsnedders>
I have a computing project to do for school.

[10:30:02.0000] <Lachy>
gsnedders, I have a feature request for the spec gen...

[10:30:03.0000] <Philip`>
I got that bit :-)

[10:30:04.0000] <gsnedders>
I'm meant to have half the write up and the implementation done by 9th Jan.

[10:31:00.0000] <gsnedders>
I'm meant to have been working on it since August.

[10:31:01.0000] <gsnedders>
I haven't. I hate what I've chosen and have no enthusiasm for it :P

[10:31:02.0000] <gsnedders>
Lachy: What?

[10:31:03.0000] <hsivonen>
gsnedders: do you keep the bibliography data in a BibTeX .bib file?

[10:31:04.0000] <Lachy>
if I use the HTML5 <section> and <h1> elements for all headings in the src document, can the spec gen automatically work out the heading levels and replace the h1's with the appropriate backwards compatible numbered headings?

[10:31:05.0000] <gsnedders>
hsivonen: I don't have any biblio data.

[10:32:00.0000] <gsnedders>
Lachy: no.

[10:32:01.0000] <Lachy>
why not?

[10:32:02.0000] <gsnedders>
Lachy: That's out of scope.

[10:32:03.0000] <Lachy>
how is that out of scope? It would make the job of restructuring a spec so much easier

[10:32:04.0000] <gsnedders>
hsivonen: The one grip I have with BibTeX is that virtually all Python things for dealing with it suck

[10:33:00.0000] <gsnedders>
*gripe

[10:33:01.0000] <Philip`>
gsnedders: I have a feature request - could you implement some plugin system so people can implement these out-of-scope features as separate modules that can be easily inserted into the appropriate processing step?

[10:33:02.0000] <Lachy>
don't you have an implementation of the heading algorithm already anyway?

[10:33:03.0000] <hsivonen>
I used a Java-based parser that a friend wrote for TeXlipse

[10:33:04.0000] <gsnedders>
Philip`: I have implemented that all ready

[10:34:00.0000] <gsnedders>
Lachy: Moving section and stuff around is a lot of effort, and I don't think it's relevant

[10:34:01.0000] <gsnedders>
Lachy: But yes, I do. It's only dependancy is having an lxml tree, AFIAK

[10:34:02.0000] <gsnedders>
*AFAIK

[10:34:03.0000] <Philip`>
gsnedders: Ah - then you could implement some module that does the heading stuff Lachy wants, as an independent project, and it wouldn't be out of scope

[10:34:04.0000] <gsnedders>
Philip`: I don't want to implement it :)

[10:35:00.0000] <Philip`>
gsnedders: That's a more compelling argument than "That's out of scope" :-)

[10:35:01.0000] <Lachy>
gsnedders, the spec gen wouldn't need to move anything around. It would just need to replace <h1> elements with appropriately numbered <h2>-<h6>

[10:35:02.0000] <gsnedders>
Lachy: And leave the section elements et al as is?

[10:35:03.0000] <Lachy>
yes

[10:35:04.0000] <gsnedders>
Lachy: And what do I do when I need h7?

[10:35:05.0000] <Lachy>
just use h6

[10:35:06.0000] <gsnedders>
Lachy: That breaks stuff.

[10:36:00.0000] <Lachy>
no it doesn't

[10:36:01.0000] <gsnedders>
It would change the outline.

[10:36:02.0000] <gsnedders>
I'd rather throw a fatal error then

[10:36:03.0000] <Lachy>
because the h6 limitation applies even if I use h1-h6 manually

[10:37:00.0000] <gsnedders>
Lachy: Then don't use too many sections :)

[10:37:01.0000] <Lachy>
I'm not intending to use too many sections

[10:37:02.0000] <Philip`>
If the outline is correct in the source which uses <h1> everywhere, wouldn't it still be correct in the output if it uses <h6> everywhere instead of <h7>?

[10:37:03.0000] <Lachy>
I have no intention of using more than 6 levels of headings, and if I do, I'll restructure so I don't

[10:37:04.0000] <gsnedders>
Philip`: Yeah, sure. But it would have the side-effect of screwing stuff up using the HTML 4.01 algo

[10:37:05.0000] <hsivonen>
gsnedders: the grammar for .bib seems reasonably short: http://texlipse.cvs.sourceforge.net/viewvc/texlipse/net.sourceforge.texlipse/source/bibtex6.sablecc?revision=1.6&view=markup

[10:38:00.0000] <hsivonen>
gsnedders: you might be able to adapt it for your Python parser generator of choice

[10:38:01.0000] <Lachy>
gsnedders, I thought the HTML5 heading algorithm was backwards compatible with HTML4, and it wouldn't screw up anything that didn't use <section>

[10:38:02.0000] <Philip`>
gsnedders: If it's using <section> then it can't be HTML 4.01, so why does it matter what HTML 4.01 tools would do?

[10:38:03.0000] <gsnedders>
hsivonen: <http://code.google.com/p/bibstuff/source/browse/trunk/bibgrammar.py> is shorter

[10:39:00.0000] <gsnedders>
Lachy: Yeah, sure. It'll be fine provided the HTML 5 algo is used

[10:39:01.0000] <Lachy>
gsnedders, one benefit of implementing this would be that it gives a clear demonstration of a use case for the new heading algorithm

[10:39:02.0000] <gsnedders>
Lachy: The use case exists anyway :P

[10:39:03.0000] <gsnedders>
hsivonen: (and Python)

[10:39:04.0000] <gsnedders>
hsivonen: The majority of the rest of the codebase sucks though

[10:40:00.0000] <Lachy>
gsnedders, I know. And I want to make use of the use case :-)

[10:40:01.0000] <gsnedders>
Lachy: Ask me when I'm in a good mood after 1.0 has shipped :P

[10:41:00.0000] <Lachy>
but if anolis supports plugins, do you have any documentation for how I can write a plugin that does this?

[10:41:01.0000] <Lachy>
when is 1.0 expected to ship?

[10:41:02.0000] <gsnedders>
Lachy: second paragraph http://hg.gsnedders.com/anolis/raw-file/1.0RC3/README.html#using-anolis

[10:42:00.0000] <Lachy>
wait, when are you ever in a good mood?

[10:42:01.0000] <gsnedders>
Lachy: I had originally said when Hixie started using it, but I haven't. Really soon now, though.

[10:42:02.0000] <gsnedders>
Lachy: Maybe next weekend?

[10:42:03.0000] <gsnedders>
Lachy: I hope 1.0RC3 to be the final RC, and I just want to see if anyone bitches about bugs in that before tagging 1.0

[10:44:00.0000] <Lachy>
ok

[10:44:01.0000] <Lachy>
how do I update anolis? Do i just download the new one and run python setup.py install?

[10:45:00.0000] <gsnedders>
yeah

[10:45:01.0000] <Lachy>
gsnedders, from that doc: foo.foo(ElementTree, **kwargs) - what is the **kwargs parameter?

[10:47:00.0000] <gsnedders>
Lachy: keyword arguments, so a dictionary of unknown arguments, and if any keyword arg matches on what it's calling, that

[10:47:01.0000] <gsnedders>
e.g., def foo(bar=False)

[10:47:02.0000] <gsnedders>
kwargs = {"bar": True}

[10:47:03.0000] <gsnedders>
foo(**kwargs)

[10:47:04.0000] <gsnedders>
that's the same as foo(bar=True)

[10:48:00.0000] <Lachy>
so if I understand correctly, if I write my process "fixheadings" and run with --enable fixheadings, then anolis will call fixheadings.fixheadings(ElementTree, **kwargs). Then my process just needs to make changes to the ElementTree and return?

[10:48:01.0000] <gsnedders>
It doesn't need to return anything

[10:49:00.0000] <Lachy>
ok

[10:49:01.0000] <gsnedders>
(this other modules function is completely untested, so it might be broken)

[10:50:00.0000] <Lachy>
gsnedders, http://anolis.gsnedders.com/#obtaining-a-copy is still linking to RC1

[10:50:01.0000] <gsnedders>
ah, I never updated that

[10:50:02.0000] <gsnedders>
/me yawns

[10:51:00.0000] <gsnedders>
fixed

[10:52:00.0000] <gsnedders>
you probably want to from anolislib.processes import outliner

[10:52:01.0000] <Lachy>
ok

[10:52:02.0000] <gsnedders>
(a section is just a list FYI)

[10:52:03.0000] <gsnedders>
(possibly with section.header)

[10:52:04.0000] <Lachy>
how do I deploy the script? Does it just look for it in the current working directory, or do I need to put it somewhere special?

[10:53:00.0000] <gsnedders>
Lachy: Somewhere Python will try and import

[10:53:01.0000] <Lachy>
ok. Does that include the current working directory?

[10:53:02.0000] <gsnedders>
It might

[10:53:03.0000] <gsnedders>
/me dunnos

[10:54:00.0000] <Lachy>
alright. I'll experiment with it

[11:39:00.0000] <Lachy>
I think it's ironic how Calogero wrote in the cite/blockquote thread: "I hope now that's more clear" immediately after one of the most confusing, grammatically incorrect and incomprehensible sentences I've ever tried to read.

[11:44:00.0000] <hsivonen>
Hixie: is it intentional that HTML5 doesn't associate <legend> with the form pointer even though Gecko does?

[11:47:00.0000] <hsivonen>
hmm. the form association is annoying. now I need to split separate magic numbers for fieldset and object

[12:10:00.0000] <hsivonen>
also had to introduce a OUTPUT_OR_LABEL magic

[14:00:00.0000] <annevk3>
/me never locks his luggage

[14:21:00.0000] <Lachy>
gsnedders, yt?

[14:25:00.0000] <Lachy>
gsnedders, I'm getting an error when I try to use anolis --enable to run a plugin

[14:25:01.0000] <Lachy>
$ anolis --enable=AdjustHeadings Overview.src.html test.html

[14:25:02.0000] <Lachy>
Traceback (most recent call last):

[14:25:03.0000] <Lachy>
  File "/usr/local/bin/anolis", line 218, in <module>

[14:25:04.0000] <Lachy>
    main()

[14:25:05.0000] <Lachy>
  File "/usr/local/bin/anolis", line 48, in main

[14:25:06.0000] <Lachy>
    tree = generator.fromFile(input, **kwargs)

[14:25:07.0000] <Lachy>
  File "/Library/Python/2.5/site-packages/anolislib/generator.py", line 84, in fromFile

[14:25:08.0000] <Lachy>
    process(tree, processes, **kwargs)

[14:25:09.0000] <Lachy>
  File "/Library/Python/2.5/site-packages/anolislib/generator.py", line 36, in process

[14:25:10.0000] <Lachy>
    process)

[14:25:11.0000] <Lachy>
AttributeError: 'module' object has no attribute '--enable'

[14:27:00.0000] <BenMillard>
Lachy, e-mail might be the best way to work through that so multi-line samples remain intact?

[14:27:01.0000] <BenMillard>
Lachy, also my impression from the earlier logs is gsnedders had higher priorities, so an e-mail would let him get to it at a time of his convenience.

[14:28:00.0000] <Lachy>
gsnedders, I'm aware of gsnedders priorities

[14:29:00.0000] <Lachy>
I also figured there may be someone else in here familiar with python who might understand what's causing the error

[14:29:01.0000] <Lachy>
s/gsnedders,/BenMillard,/

[14:30:00.0000] <BenMillard>
Philip` & annevk3, any ideas about the error Lachy is seeing?

[14:30:01.0000] <hsivonen>
BenMillard: I didn't fill out the accessibility API survey, at least not yet, because I figured that I should defer to aaronlev on API things

[14:31:00.0000] <BenMillard>
hsivonen, I thought it was about individuals' participation rather than organisations'?

[14:32:00.0000] <hsivonen>
BenMillard: it had a question on patent policy, which looks like an organizational thing

[14:36:00.0000] <BenMillard>
hsivonen, ah

[14:36:01.0000] <Lachy>
hsivonen, the way the patent policy question is worded, it looks like it's an individual question

[14:36:02.0000] <BenMillard>
since Mozilla won't renew my funding, I answered it as an individual

[14:43:00.0000] <hsivonen>
I'm sorry to hear your work is no longer funded

[14:45:00.0000] <BenMillard>
hsivonen, apparently the "overlap is too thin" between what I was doing and the scheme my funding was granted under, but that my work was useful and I could ask other people in Mozilla for funding (but no clue was given about whom)

[14:45:01.0000] <BenMillard>
so once the comparison document is done, I'll be knocking on doors :)

[14:50:00.0000] <Philip`>
/me tries to find a font to use in Keynote that isn't Gill Sans (because that's a blatant "look at me I'm a Mac user" sign) and isn't really boring (Arial, Helvetica, etc), and ends up finding only Trebuchet MS

[14:51:00.0000] <heycam>
Philip`, i like Myriad (if you have it) as a nice sans serif font

[14:51:01.0000] <Philip`>
heycam: Hmm, that doesn't seem to be installed here

[14:51:02.0000] <heycam>
i don't know what the big deal with gill sans is, i think it looks a bit ugly, especially in the heavier weights

[14:52:00.0000] <Lachy>
don't go with Trebuchet MS. That's one of the ugliest fonts ever

[14:52:01.0000] <BenMillard>
Philip`, got any from the Lucida family?

[14:52:02.0000] <heycam>
or take a look at http://www.webfonts.info/wiki/index.php?title=Fonts_available_for_%40font-face_embedding#Fonts_with_an_Apache_Licence

[14:52:03.0000] <heycam>
(not necessarily with that fragid tho)

[14:53:00.0000] <Philip`>
BenMillard: Only Lucida Grande, which looks a bit boring and/or ugly

[14:53:01.0000] <Philip`>
Lachy: What's wrong with it? :-)

[14:54:00.0000] <hsivonen>
/me uses Gill Sans on the default Steve background when presenting

[14:54:01.0000] <BenMillard>
Philip`, how about Georgia? That's a tidy yet attractive (imho) serif which is quite common.

[14:55:00.0000] <Philip`>
BenMillard: It seems unnatural to use a serif font for a presentation

[14:56:00.0000] <gsnedders>
Lachy: um, yeah. bug in my code.

[14:56:01.0000] <BenMillard>
Philip`, you could try browsing around Joe Clark's various websites and stylesheets, see if any fonts he lists float your boat.

[14:58:00.0000] <gsnedders>
Lachy: yt?

[14:58:01.0000] <Lachy>
yes

[14:59:00.0000] <Lachy>
when will you have time to fix the bug?

[15:00:00.0000] <heycam>
Lachy, do you have any objection to my following the approach (for idl module -> package mapping) in http://www.w3.org/mid/20081129021607.GF16515⊙amia ?

[15:01:00.0000] <Lachy>
heycam, do you mean the hard coded prefix approach you described?

[15:02:00.0000] <heycam>
yeah

[15:02:01.0000] <heycam>
with ability to override it with [Prefix]

[15:02:02.0000] <Lachy>
so if I understand correctly, then I wouldn't need to make any more changes to selectors api?

[15:02:03.0000] <heycam>
what's in there at the moment?  just:  module dom { interface ... } ?

[15:02:04.0000] <Lachy>
using: module dom {...} is enough?

[15:02:05.0000] <heycam>
then yes

[15:02:06.0000] <heycam>
no changes required

[15:02:07.0000] <Lachy>
ok. That's fine with me

[15:02:08.0000] <heycam>
:)

[15:02:09.0000] <heycam>
ok thanks

[15:04:00.0000] <gsnedders>
Lachy: committed and pushed

[15:07:00.0000] <gsnedders>
Lachy: http://hg.gsnedders.com/anolis/archive/85a6d32479c4.tar.bz2

[15:07:01.0000] <gsnedders>
(or .zip or .tar.gz)

[15:08:00.0000] <Lachy>
gsnedders, ok, now it's complaining that there is no module named "adjustheadings". So I guess I'm doing something wrong

[15:10:00.0000] <Lachy>
I have a file named adjustheadings.py in my working directory, which contains this:

[15:10:01.0000] <Lachy>
import sys

[15:10:02.0000] <Lachy>
import os

[15:10:03.0000] <Lachy>
from anolislib.processes import outliner

[15:10:04.0000] <Lachy>
class adjustheadings:

[15:10:05.0000] <Lachy>
 """Change the heading elements to use the backwards compatible, numbered headings"""

[15:10:06.0000] <Lachy>
    def __init__(self, ElementTree, **kwargs):

[15:10:07.0000] <Lachy>
        self.ElementTree = ElementTree

[15:10:08.0000] <Lachy>
 def adjustheadings(self, ElementTree, **kwargs):

[15:10:09.0000] <Lachy>
  print "test"

[15:11:00.0000] <gsnedders>
it'll be relative to the installed generator.py, or absolute

[15:11:01.0000] <gsnedders>
easy solution: put it in an absolute location

[15:11:02.0000] <gsnedders>
where: python -c "import sys; print sys.path"

[15:12:00.0000] <Philip`>
Lachy: Maybe run 'PYTHONPATH=. anolis ...'

[15:13:00.0000] <gsnedders>
Philip`: That won't help. It's relative to the anolislib file.

[15:14:00.0000] <Lachy>
ok, so if I just place a copy in the same directory as anolis is installed in it should work.

[15:14:01.0000] <Lachy>
I just have to find where that is

[15:15:00.0000] <gsnedders>
Lachy: No, it's easier if you place it in the normal python location

[15:15:01.0000] <gsnedders>
Lachy: On OS X? Using default Python install (and not a macport/fink one)?

[15:15:02.0000] <Lachy>
ok, which location exactly?

[15:16:00.0000] <gsnedders>
Lachy: /Library/Python/2.5/site-packages

[15:16:01.0000] <Lachy>
I think its the macports one

[15:16:02.0000] <Lachy>
how do I tell?

[15:17:00.0000] <gsnedders>
Lachy: which python

[15:17:01.0000] <Lachy>
ah, that's where anolislib is installed

[15:17:02.0000] <gsnedders>
Lachy: Just put it in that folder

[15:18:00.0000] <Lachy>
do you mean the site-packages folder, or the site-packages/anolislib folder?

[15:18:01.0000] <gsnedders>
Lachy: site-packages

[15:18:02.0000] <Lachy>
ok

[15:20:00.0000] <gsnedders>
Now, I'm really not here till after school tomorrow

[15:21:00.0000] <Lachy>
it's running without errors now. I'm not sure if it's actually working though

[15:21:01.0000] <gsnedders>
is adjustheadings.__init__ being called?

[15:22:00.0000] <Lachy>
I wonder if it'll work if I place a symlink in there instead, so I can keep the actual file in my working directory

[15:22:01.0000] <Lachy>
I don't know. what's the easiest way to find out?

[15:22:02.0000] <Lachy>
is there some way to output debug messages?

[15:22:03.0000] <gsnedders>
print "magic foo" in that method?

[15:22:04.0000] <gsnedders>
now, I really really really must go

[15:23:00.0000] <gsnedders>
Ask Philip`. He's clever. He's got a degree. He's finished this whole school thing. :)

[15:23:01.0000] <Lachy>
yes

[15:23:02.0000] <Lachy>
alright, now I just have to learn the ElementTree API

[15:58:00.0000] <Philip`>
Don't accuse me of knowing things!