00:54
<aruner>
Hixie, what's the history of the blocked header list in XHR in the spec.? How was that list arrived at?
00:54
<Hixie>
bjoern sent a detailed reply regarding this recently
00:55
<Hixie>
in other news i am especially amused that sunava just delayed another MONTH
00:55
aruner
kicks his mail client for retrieving messages slowly
00:56
<Dashiva>
Hixie: But otherwise it's going well. She's quoting properly, at least
00:56
<Hixie>
aruner: http://lists.w3.org/Archives/Public/public-webapi/2008May/0189.html
00:56
<Hixie>
Dashiva: i believe sunava is a he, am i wrong?
00:57
<Hixie>
annevk: yt?
00:57
<Dashiva>
I'm just guessing, so you're probably right
00:57
aruner
sighs. Though I *do* think that their general design principles are good (be conservative, err on the side of ultra safe) I think specifics help us understand their notions on threat modeling.
00:58
<Dashiva>
Flickr seems to support you
00:58
<Hixie>
aruner: their design principles resulted in xdr, which has at least three known security bugs, so... :-)
00:59
aruner
grins
01:00
<aruner>
Dashiva: Sunava Dutta is a he
01:00
<Dashiva>
He and anne should make a club
01:00
<Hixie>
i live in the bay area
01:01
<Hixie>
i don't make any judgements regarding gender
01:01
<aruner>
Hixie: are you putting in an appearance at the F2F in Redmond? I'm not sure.
01:01
<Hixie>
as it's not uncommon for people to change gender...
01:01
<Hixie>
aruner: i'd like the spec to be done by then.
01:01
<Hixie>
aruner: in which case i won't turn up
01:01
<Dashiva>
I try to use she for females because some get insulted by 'he' as gender-neutral
01:01
<Hixie>
aruner: we'll see nearer the time
01:01
<Philip`>
You can make a pretty good guess at someone's gender based solely on the fact that they're posting to a web specification mailing list
01:02
<Hixie>
sad but true
01:02
<Dashiva>
That's asking for a beatdown, Philip` ;)
02:21
<Hixie>
so i wonder how to define sandboxes
02:22
<Hixie>
such that if you have document A with iframe containing document B with iframe containing document C
02:22
<Hixie>
that:
02:23
<Hixie>
1. if A, B, and C are from the same origin, B can't play with C
02:23
<Hixie>
hm wait
02:23
<Hixie>
i was going to say that if B and C have hte same origin but aren't the same as A they should be able to talk
02:23
<Hixie>
but no, that would still allow problems
02:23
<Hixie>
so we really do want unique origins...
02:23
<Hixie>
hmm
02:23
<Hixie>
that makes it easier
02:24
<Hixie>
HMM.
02:24
Hixie
ponders
02:26
<othermaciej>
so I assume you are thinking sandboxes that still have script enabled?
02:26
<othermaciej>
(since for noscript iframes this would not matter)
02:27
<othermaciej>
maybe I should step back and ask what use cases you are targetting
02:27
<Hixie>
actually same-origin matters even without scripts
02:27
<Hixie>
a bunch of things use the origin concept
02:27
<Hixie>
but yes, i am considering scripts too
02:27
<othermaciej>
like frame targetting?
02:27
<Hixie>
yeah, that sometimes uses origin
02:27
<othermaciej>
I think the desired restrictions depend on the use cases
02:27
<othermaciej>
ones I can imagine:
02:28
<othermaciej>
- blog comments
02:28
<othermaciej>
- user-authored content in a profile on a social network site
02:28
<othermaciej>
- "widgets" embedded from off-site
02:28
<othermaciej>
- "widgets" hosted by the provider of the page containing them
02:28
<othermaciej>
- ads
02:29
<Hixie>
comments and other user-contributed content, gadgets (same site and other site; including ads), and (not sandboxing but still <iframe> related) client-side includes are the main ones
02:29
<othermaciej>
ads are complicated because at least some ad networks consider it important to support frameless flash ads that overlay the content
02:29
<othermaciej>
but anything that can overlay the content is not properly sandboxed by definition
02:30
<Hixie>
there are indeed various scenarios with varying levels of sandboxing
02:30
<Hixie>
http://www.whatwg.org/specs/web-apps/current-work/sandbox has my current strawman proposal (cut and pasted from the start of the e-mail i'm writing; note that i've used past tense because it'll be past tense once i send the mail, but the spec isn't actually updated yet)
02:30
<othermaciej>
for comments you would like them to be able to size naturally and seem to be inline as far as the end-user is concerned, but they should not be able to css position outside their area
02:32
<Hixie>
yeah the comments would be <iframe sandbox="unrestricted-origin" seamless src="data:text/html,(small html document)"></iframe>
02:33
<othermaciej>
I'm thinking about whether inheriting styles and applying outer stylesheets may be a security risk
02:33
<Hixie>
in certain cases it is
02:33
<Hixie>
in others not so much
02:33
<Hixie>
i can't see how it could be in a script-limited blog comment case
02:34
<othermaciej>
I'm thinking for the comment use case and the user-generated content on hosted profile page use case
02:34
<Hixie>
oh i have to disble forms too while i'm at it
02:34
<othermaciej>
and also plugins probably
02:34
<Hixie>
yeah
02:34
<othermaciej>
(including Java)
02:35
<Hixie>
updated sandbox file
02:35
<othermaciej>
I am nervous about the performance impact of using an iframe per comment with hundreds of comments
02:35
<Hixie>
yeah me too
02:35
<Hixie>
i'm not sure what else to do though
02:36
<othermaciej>
navigating the parent should be restricted (dunno if that falls out of other things), including back/forward
02:36
<othermaciej>
window sizing and positioning should be restricted
02:37
<Hixie>
any browsing context navigation outside the sandbox is blocked
02:37
<othermaciej>
does embedding further subframes in the content need to be restricted?
02:37
<Hixie>
though that does mean that links in comments won't work
02:38
<othermaciej>
or do they just inherit the sandbox restrictions already in effect?
02:38
<Hixie>
i don't see why it should be, but they certainly are also sandboxed
02:38
<Hixie>
and they're not in the same sandbox, either
02:39
<othermaciej>
links in blog comments not working is kind of a bummer
02:39
<othermaciej>
perhaps there should be a way to turn that on, since it can't be used as an annoyance in the noscript case
02:39
<Hixie>
yeah
02:40
<Hixie>
comments would now be <iframe sandbox="unrestricted-origin force-links-to-parent" seamless src="data:text/html,(small html document)"></iframe>
02:40
<Hixie>
it's getting long
02:41
<othermaciej>
the keywords are a bit verbose
02:41
<Hixie>
it's a security thing, i'm worried about making them too simple
02:41
<Hixie>
i'm especially worried about making them sound like they _increase_ security
02:41
<Hixie>
when all the keywords decrease it somewhow
02:41
<othermaciej>
perhaps there should be two addtributes
02:42
<othermaciej>
sandbox="" or restrict="" that turns everything off, and an allow="" attribute to turn things back on selectively
02:42
<othermaciej>
or sandbox-exceptions=""
02:43
<Hixie>
well "force-links-to-parent" isn't really an exception or a restriction, it just changes the rules to send all user-activated links to the parent browsing context
02:43
<othermaciej>
you could achieve that with <base target="parent"> and allowing targetted links to parent only
02:43
<othermaciej>
a bit more work but makes it strictly an exception, not a behavior change
02:43
<Hixie>
(<iframe seamless sandbox allow="..." src="data:text/html,(small html document)"></iframe> makes sense, but i'm not sure what the keywords would be)
02:44
<othermaciej>
though if it is not a restriction or an exception perhaps it should be a separate attribute like seamless
02:44
<Hixie>
well the idea is that you can allow whatever random content in the sandbox as you want
02:44
<Hixie>
so i'm not sure i want to be poking <base> into it
02:44
<Hixie>
ooo
02:44
<Hixie>
wait!
02:44
<Hixie>
it can just be part of seamless' behaviour!
02:44
<othermaciej>
for the comment use case presumably the data: URL is generated by software and the user's comment does not have <html> or <body>?
02:44
<othermaciej>
oh interesting
02:45
<Hixie>
comments would now be <iframe sandbox="unrestricted-origin" seamless src="data:text/html,(small html document)"></iframe> again
02:45
<othermaciej>
is it ever useful to allow gadgets to navigate the parent or open a window?
02:45
<Hixie>
probably, but i want to start with as small a set of allowed behaviours as possible
02:51
<othermaciej>
I guess for gadgets, anything forbidden can be selectively exposed via postMessage
02:51
<Hixie>
eah
02:51
<Hixie>
y
02:53
<Hixie>
maybe seamless should base its origin determination on the "real" origin, not the origin affected by sandbox=""
02:53
<othermaciej>
Hixie: i think the seamless navigation rule should only apply to navigations triggered by the content of the iframe
02:53
<Hixie>
as opposed to what?
02:53
<othermaciej>
Hixie: if the parent doc navigates a seamless iframe it should not navigate itself
02:53
<Hixie>
hm, that makes things more complex
02:53
<Hixie>
why would you use target="" on a seamless iframe?
02:53
<Hixie>
why not just change .data?
02:53
<Hixie>
or .src
02:54
<Hixie>
or whatever
02:54
<othermaciej>
changing .src is a navigation
02:54
<Hixie>
not per spec
02:54
<Hixie>
oh actually it is
02:54
<Hixie>
nevermind
02:54
<Hixie>
i'm thinking of <object>
02:54
<othermaciej>
well put it this way, changing iframe.src, changing iframe.contentWindow.location, and targetted link aimed at iframe all go at the same code path
02:54
<othermaciej>
at least in WebKit
02:55
<Hixie>
yeah
02:55
<Hixie>
spec too
02:55
<Hixie>
ok
02:55
<Hixie>
well
02:55
<Hixie>
i could just do it for <a> links
02:55
<othermaciej>
knowing the source frame of a navigation is not a big deal
02:55
<Hixie>
i guess
02:55
<othermaciej>
we always have to know anyway
02:55
<Hixie>
though what do you call the source if someone from another iframe clicks a link using .click() ?
02:56
<Hixie>
or what if frame A calls .click() on a link in frame B which has a target="" to A, which is seamless?
02:56
<othermaciej>
I would say only cases where the source frame is the seamless frame itself should get retargetted
02:57
<Hixie>
sure but what is the source frame in that example?
02:58
<othermaciej>
I don't know in general what a cross-frame click() call on a link considers the source frame
02:58
<othermaciej>
presumably something has to define which it is in any case
02:59
<othermaciej>
I don't think it matters which is chosen from security POV
02:59
<othermaciej>
since A has to have unrestricted origin in your scenario and thus could just modify B to do something bad
03:01
<Hixie>
well i've never had to define it before now :-)
03:01
<Hixie>
but i agree it doesn't really matter much
03:01
<Hixie>
so long as it is defined
03:02
<Hixie>
i guess i'll just add an "argument" to the "navigate" algorithm and make all the call sites say what the source is
03:04
Hixie
defines the term "source browsing context"
03:04
<Hixie>
(better terms welcome)
03:05
<Hixie>
hm, this causes <meta http-equiv=refresh> inside a seamless iframe to navigate the parent iframe
03:06
<Hixie>
oh hey, we do need this, to define .referer better
03:06
<Hixie>
:-)
04:54
<MikeSmith>
Hixie: why's the body element included in the "Obsolete elements, attributes, and APIs" section?
08:20
<othermaciej>
hey web expert people
08:21
<othermaciej>
is there any way in Firefox to do the equivalent of either word-wrap: break-word (breaks lines in the middle of long words that would overflow the box) or text-overflow: ellipsis (ellipsizes lines that are too long)?
08:28
<jwalden>
othermaciej: 3 you could use &shy; or the Unicode codepoint, I suspect
08:28
<othermaciej>
jwalden: between every character?
08:28
<jwalden>
wait
08:28
<jwalden>
that's what that does?
08:28
<othermaciej>
someone was asking me how to handle a long word in a fixed-width box gracefully
08:29
<jwalden>
"gracefully"
08:29
<othermaciej>
(url to be specific)
08:29
<othermaciej>
overflow: visible makes it stick out of the box
08:29
<othermaciej>
overflow: hidden silently truncates
08:29
<othermaciej>
overflow: scroll adds a scrollbar
08:29
<jwalden>
last I recall we were breaking URLs, actually; that may have changed
08:29
<othermaciej>
I suggested word-wrap break-word to wrap the word in the middle of the box
08:29
<othermaciej>
or text-overflow: ellipsis to truncate but with an ellipsis at the end
08:30
<othermaciej>
neither of those work in FF though
08:30
<jwalden>
I'd do &shy; at the / boundaries myself
08:31
<jwalden>
I don't remember anything particular that would work
08:31
<othermaciej>
(I just used a long word instead of a URL as my test case though so I could not tell if there was magical URL breaking, but I assume no or they would not have complained of this bug)
08:32
<jwalden>
so data:text/html,%3Cdiv%20style=%22width:%2010em;%20border:%201px%20dotted%20green;%22%3Ehttp://localhost:8888/tests/browser/microformats/test/test_Microformats_hCard.html%3C/div%3E does wrapping for me in 3+ source code, I think 3 does the same
08:32
<roc>
<wbr> or a zero-width unicode space?
08:33
<roc>
there is magical URL breaking in Firefox 3
08:33
<othermaciej>
I was testing with an FF3 beta but possibly an old one
08:33
<roc>
although it's not terribly magical, it works pretty well
08:34
<othermaciej>
I think it it desired for this to work for non-URL overflow too or overflow of a single url component (the column in question is not very wide)
08:35
<othermaciej>
and not sure that putting &shy; or <wbr> or zero-width space between every single character will fly
08:35
<othermaciej>
but I guess I can pass it along
08:35
<roc>
it doesn't have to be every single character, I guess
08:40
<othermaciej>
(both the things I mentioned work in IE as well as Safari fwiw, might be worth adding one or both to Gecko at some point)
08:40
<roc>
yeah, I should implement word-break
08:40
<roc>
text-overflow is totally underspecified
08:41
<roc>
and Webkit, at least, does weird things with it sometimes. Probably IE too although i test with IE less
08:41
<othermaciej>
word-wrap: break-word != word-break
08:41
<roc>
oh right
08:42
<othermaciej>
it is handy though cause it does what you expect in an editable area when your line is too long and is all one word (or at least what I expect)
08:42
<roc>
"Shaping characters are still shaped as if the word were not broken" ... that is weird
08:42
<othermaciej>
yeah, you would want arabic chars to still take medial form
08:42
<othermaciej>
not terminal or initial
08:42
<roc>
I guess that's the easy to implement way
08:43
<othermaciej>
it also seems like the correct way (at least for Arabic)
08:43
<roc>
but not for optional ligatures
08:43
<othermaciej>
presumably you can't break mid-ligature
08:43
<roc>
why not?
08:43
<othermaciej>
because it is one glyph
08:43
<othermaciej>
I mean
08:43
<othermaciej>
you could break between the chars
08:43
<othermaciej>
but the glyph is on one line or the next
08:44
<roc>
you can cut it in half
08:44
<roc>
we have to do that for selection and ligatures that cross span boundaries with different styles
08:47
<othermaciej>
if you can tell the difference between things that are ligature-like and things that are like Arabic I think you would be in your rights to undo the ligature, but I think breaking on one side of the glyph or the other should do, as perfecting the breakpoint when you already broke in the middle of a word seems less than essential
08:47
<roc>
right
08:47
<roc>
that's what I'd do
08:47
<annevk>
Hixie, am now
08:48
annevk
reads up on the sandbox discussion in the backlog
09:16
<annevk>
MikeSmith, I think it's there to define obsolete parts of the body element
09:24
<MikeSmith>
annevk: OK
09:28
<annevk>
I think it's confusing
10:01
<MikeSmith>
annevk: for the publication-notes doc, I wrote a short context description of that section as "contains descriptions of the obsoleted applet element and obsolete parts of DOM interface for the body element."
10:19
<MikeSmith>
so the table in section 8.6, "Named character references" was removed completely
10:19
<MikeSmith>
http://www.w3.org/html/wg/html5/#named
10:19
<MikeSmith>
wondering if we want to published without that
10:19
<MikeSmith>
if so, a note should probably be added there
10:20
<annevk>
that seems like a bug in the W3C script
10:20
<MikeSmith>
ah
10:21
<MikeSmith>
so still something that'll need to get fixed before we publish
10:21
<othermaciej>
annevk: I think Hixie's sandbox proposal is converging (at least I mostly liked the last thing I saw before I left work)
10:22
<othermaciej>
annevk: I fear the idea of using an iframe per comment will just not fly performance-wise (even with data: URLs)
10:22
<annevk>
yeah, it's such a hack
10:22
<annevk>
i doubt people will use it over better sandboxing on the server using real html5 parsers and tools
10:23
<othermaciej>
but you do need visual sandboxing I think (not let the content to overflow) as well as script, though I guess overflow: hidden does that
10:23
<othermaciej>
(does overflow: hidden hide relative, absolute or fixed position children that stick out?)
10:23
<othermaciej>
annevk: his design seems ok for gadgets/widgets embedded in web pages
10:23
<othermaciej>
or ads
10:24
<othermaciej>
not sure it flies for other use cases
10:45
<hsivonen>
Does Safari 3.1 support closed captiong for video?
10:51
<hsivonen>
othermaciej: Does Safari 3.1 support closed captiong for video?
10:51
<othermaciej>
oddly someone just asked me that
10:51
<othermaciej>
I have not tested
10:51
<hsivonen>
ok. thanks
10:51
<othermaciej>
I do not know how you turn it on in QuickTime
11:05
<annevk>
oh lol
11:06
<annevk>
Robert Burns is joining the ARIA discussion
11:06
<annevk>
guess I'm now more than done with that :)
11:07
hendry
wonders if there is a resource describing each test in Acid3 or perhaps some sort of source control access. best i've found is http://en.wikipedia.org/wiki/Acid3
11:08
<othermaciej>
oh no, Robert Burns
11:08
<othermaciej>
is there any clusterfuck of a discussion that he can't make worse?
11:10
<hsivonen>
sigh
11:10
<Dashiva>
hendry: You mean to show what spec part they test?
11:10
<annevk>
i wonder if i should reply
11:11
<annevk>
he's totally wrong, but i know that when i point that out he finds some way to turn the argument around
11:13
<Dashiva>
annevk: I suppose the question is if anyone else listens to him, and needs clarification :)
11:13
<annevk>
yeah, deleted the e-mail
11:13
<othermaciej>
" In my view the null namespace should not exist in a namespaced document processed by a namespace aware application. Instead unprefixed attributes should be attached to the same namespace as their parent elements."
11:14
<hsivonen>
he is not wrong about XMLNS having a design bug
11:14
<hsivonen>
it's just a bug that is too late to fix
11:14
<othermaciej>
he seems to think namespaces already work this way
11:15
<hsivonen>
many namespace advocates do
11:15
<othermaciej>
also seems to think adding namespaces (with : syntax) in HTML5 would solve anything with regards to compat issues with the proposal
11:17
<annevk>
the whole point about ARIA is that it doesn't require architectural changes
11:17
<annevk>
so arguing we first need to make architectural changes misses the point
11:18
<annevk>
but pointing that out to Robert Burns doesn't work based on past experience
11:18
<hsivonen>
I have to go buy a new display with a powered USB hub right now.
11:18
<hsivonen>
I need 3 USB devices to get work done and the MacBook can power only 2
11:20
<othermaciej>
however Robert Burns does have the ability to increase confusion of third parties to the discussion if not corrected
14:57
<hsivonen>
(I got an Apple display. thanks for the recommendations)
15:04
<annevk>
http://lists.w3.org/Archives/Public/public-forms/2008May/att-0022/2008-05-14.html#topic3 "Steven Pemberton: There was a huge amount of RDFa."
15:04
<annevk>
(over XTech)
15:05
<hsivonen>
annevk: different room, different amount of RDFa, I guess
15:06
<annevk>
Not sure what's happening to the Forms Task Force
15:07
<annevk>
Also seems they continue with copying HTML features. It seems though that they try to copy from the HTML4 specification rather than from implementations...
15:10
<hendry>
Dashiva: er, OK (sorry missed this IRC window somehow)
16:01
<Philip`>
http://doctype.googlecode.com/svn/trunk/tests/html/elements/plaintext-element-appears-in-dom-test.html - that's not quite right :-(
16:59
<JJ08>
Are there currently any problems with the <video> element in safari?
17:00
<JJ08>
i can get it to show the player but the .mpg wont play. Also the poster image covers the play controls sometimes.
19:14
virtuelv
curses at Google's newfound love for not making websites work in browsers instead of named product
19:14
<virtuelv>
+s
19:14
<JJ08>
lol. Sounds like your enjoying yourself
19:15
<virtuelv>
not at all
19:15
<virtuelv>
http://www.google.com/translate_t
19:16
<virtuelv>
they have opted for creating a rather ugly, custom widget instead of a standard <select>
19:16
<virtuelv>
continuing down that path, they might just as well create their own runtime, only viewable with the google client, on the google os
19:16
<MikeSmith>
the docreader thing looks like of a like a frankenstein monster that escaped from the lab and now out terrorizing the villagers
19:17
<virtuelv>
running on Google hardware
19:17
<virtuelv>
MikeSmith: yeah, and now they ruined translate as well
19:17
<virtuelv>
http://www.google.com/translate_t
19:17
<JJ08>
lol.
19:17
<JJ08>
i see what you mean now.
19:17
<virtuelv>
what's the use of standards when everything gets replaced with JavaScript that barely works anywhere
19:17
<JJ08>
its not very "standard" at all.
19:18
<JJ08>
oh, see you beat me to it
19:18
<JJ08>
and at my college Javascript is disabled .... so that goes to pot
19:18
<takkaria>
it's that bloody google web toolkit thing
19:18
<virtuelv>
if you turn off js, it sort of works
19:19
<virtuelv>
Sincerely, GWT should die
19:19
<JJ08>
there turning into Apple (even though i just spent over £1000 on one)
19:19
<JJ08>
tightly controlling everything they own
19:21
<JJ08>
googles source code is just one big line. They forgot to use "/n". lol
19:21
<virtuelv>
a single character costs google
19:22
<virtuelv>
removing redundant whitespace is IMHO acceptable for stuff that's production code, and has undergone proper QA
19:22
<JJ08>
i suppose if you put it like that.
19:23
<Lachy>
jgraham__, we're presenting on day 1 http://www.vivabit.com/atmedia2008/london/schedule/
19:24
<Lachy>
assuming the schedule doesn't change again between now and then
19:24
<virtuelv>
<DIV class="goog-inline-block goog-menu-button-caption" unselectable="on">Spanish</DIV>
19:24
<JJ08>
i wanted to go to @media. I had all the cash sorted then i had to move house! Bummer, well pissed off i was
19:38
<Dashiva>
When this is all over, we'll have to make medals for the people who risked mind and sanity in the alt wars
19:46
<smedero>
Dashiva: That or we'll all lose our minds and start implementing Dmitry Turin's HTML 6.
19:46
<Dashiva>
I don't see how that'll be much worse than our current heading
19:46
Dashiva
mumbles
19:47
<smedero>
I kinda wish he would email the list now with how he intends to support alt on his new picture input type: http://html6.by.ru/site/html60/en/author/inputpic_eng.htm
19:47
<smedero>
oh, i'm just being cruel now.
19:47
<smedero>
:(
19:48
<JJ08>
wtf is HTML 6?
19:50
<virtuelv>
JJ08: Dmitry Turin's private proposal for next-generation HTML
19:50
<virtuelv>
I'm amazed at how much people here actually know about it
19:50
<virtuelv>
http://html6.by.ru/site/html60/en/author/label_eng.htm
19:51
<JJ08>
i am taking this as a joke. This looks idiotic
19:51
<virtuelv>
<style>
19:51
<virtuelv>
form { next: execute }
19:51
<virtuelv>
</style>
19:51
<virtuelv>
<body>
19:51
<virtuelv>
<form action="program.exe">
19:51
<smedero>
You can also see his musings on SQL 5: http://sql50.euro.ru/site/sql50/en/author/index_eng.htm
19:51
<virtuelv>
</body>
19:51
<virtuelv>
Value next=execute orders to execute received information in local operational system (mainly value next=execute is intended for those cases, when form's address specifies into local file system).
19:51
<JJ08>
what the hell is this: <a href="./c.htm#:body/h2[@@~"part of header"]+h3[5]"> text </a>
19:51
<virtuelv>
http://html6.by.ru/site/html60/en/author/transfer_eng.htm
19:51
<JJ08>
is this some random guy that has set this uo?
19:51
<JJ08>
*up
19:52
<smedero>
Dmitry pops into various standards related mailing list from time to time and proposes one of his concepts.
19:52
<JJ08>
have any actually been used?
19:52
<smedero>
not that I know of, no.
19:53
<JJ08>
he looks serious about what he is trying to preach.
19:53
<smedero>
to be fair to Dmity, he at least provides examples and occasionally use cases.
19:53
<smedero>
however insane they may be
19:54
<virtuelv>
actually, I have seen a user-js implementation of his XPath-links
19:57
<virtuelv>
and the idea itself, for being able to link into specific locations of a document that doesn't have relevant anchors or id's has merit
19:59
<JJ08>
Maybe he will come up with something beneficial. I'll give him credit because he provides examples and is obviously working on ways to improve. Its more than what im doing.
19:59
<Dashiva>
It's not that his ideas necessarily are bad, but most of them are too different from everything we already have to fit in
20:00
<Dashiva>
He seems like he really likes stream processing, maybe we could get him to work on event-source :)
20:01
<JJ08>
but how can someone be considered credible if he isn't considering 'standards'. His source code is evidence to that.
20:01
<Lachy>
Hixie, yt?
20:01
<Dashiva>
JJ08: Well, our alt advocate friends are carrying that banner high themselves, so why can't Dmitry? :)
20:03
<JJ08>
any links to these 'friends' ? I know your telling the truth, id like to see thats all
20:04
<jgraham__>
Lachy: Yeah, they moved the progrmme around a lot at the same time. I wonder if Steve Faulkner will be around on Day 1
20:04
<jgraham__>
s/at the same time/ at the same time as scheduling us/
20:05
<Dashiva>
JJ08: If you can teach me how to search message contents in thunderbird, I'd probably find it.
20:05
<takkaria>
Dashiva: top right-hand corner search box, click on the magnifying glass, select "entire message"
20:06
<Dashiva>
Oh, swell
20:06
<Dashiva>
... but it seems it gladly hits the server instead of using the local copy.
20:06
<takkaria>
you using IMAP?
20:06
<Dashiva>
Yes
20:06
<JJ08>
lol, i cant :( So what your saying is that these people who are involved in the development of HTML 5 are not even following standards?
20:06
<smedero>
the public-html archive search works fairly well: http://lists.w3.org/Archives/Public/public-html/
20:07
<Dashiva>
JJ08: No, not like that
20:07
<takkaria>
Dashiva: IMAP has a built-in search command, so Tb uses that by default
20:07
<Dashiva>
Just that even advocates sometimes fail to go the extra mile in every single thing
20:07
<JJ08>
sorry. I get you now.
20:07
<Lachy>
jgraham__, yeah, Steve should be around.
21:12
<Lachy>
My podcast for boagworld.com that I recorded yesterday will most likely appear around the 4th of June
21:42
<mcarter>
good afternoon
21:43
<mcarter>
Hixie, I don't suppose you ever wrote that TCPConnection server perl script for your train set that I can test my implementation against?
23:27
<virtuelv>
http://www.regdeveloper.co.uk/2008/05/16/html5_xhtml_vendors_lockin/
23:30
<Lachy>
The article says: "For sophisticated cross-platform applications, there already exist several proprietary solutions (such as Mozilla's XUL, Adobe's Flash, or Microsoft's Silverlight)," the group said.
23:30
<Lachy>
Who in the group said that?
23:30
<takkaria>
that seems like a very confused bit of reporting
23:30
<Hixie>
the spec says that
23:31
<Hixie>
so much for the spec not being a position paper. :-D
23:31
<Hixie>
hahahahaha
23:32
<Hixie>
i love the picture of chris wilson with a caption "Microsoft's Wilson: modular HTML 5"
23:32
<Hixie>
that's hilarious
23:36
<Dashiva>
The picture reminds me of those british judges with wigs
23:36
<Dashiva>
Poor Chris
23:37
<Lachy>
LOL
23:39
<takkaria>
the IBM article that links to I haven't read before
23:39
<takkaria>
has some good lines
23:39
<takkaria>
"The HTML V5 specification was written using good communication with browser vendors ... The team is sceptical about official W3C approval, though. The FAQ doesn't even try to give a serious answer about the expected date of approval."
23:40
<Lachy>
IIRC, that IBM article wasn't very good
23:40
<takkaria>
it's not, no
23:41
<Lachy>
wow. I wrote that FAQ, and I certainly did try to give a very serious answer to that question
23:42
<Hixie>
i thought the answer in the faq was pretty accurate
23:42
<Lachy>
though, saying 15 years generally doesn't go down well with people and it works better if you emphasise that features can be used as soon as browsers implement them, rather than waiting for the finished spec
23:43
<Hixie>
it took us 18 years to get where we are now
23:43
<Lachy>
that's what I did in the podcast yesterday, and the host responded quite well to that answer
23:43
<Hixie>
and html5 is specifying all of that and enough to take us forward
23:43
<Hixie>
and they think 15 years to get two perfect implementations is too long? :-)
23:43
<Hixie>
some people have issues
23:44
<roc>
I've started having terrifying thoughts that I'm going to be working on Web browsers for the rest of my life
23:45
<Lachy>
re the modularisation, they fail to realise that there may be a lot to write specs for, but there's too few editors to do it.
23:45
<Hixie>
roc: why is working on the cutting edge of one of humanity's biggest achievements terrifying? :-)
23:45
<Hixie>
Lachy: who is "they"?
23:46
<roc>
I bet Pharoah said the same thing to the slaves building his pyramids
23:46
<Lachy>
the people who are pushing for splitting the spec up into modules
23:46
<Lachy>
chris wilson, included
23:46
<Hixie>
roc: :_P
23:47
<Hixie>
roc: i heard from a student who attended the talk you gave at stanford btw, remind me to slip you a fiver when we next meet :-)
23:47
<roc>
ooh, which one?
23:47
<Lachy>
oh, at least MikeSmith does: So for now, nothing changes as far as the W3C is concerned. Smith noted that while modularization of specification is a good idea in principle, the consortium needs a commitment from one or more member organizations to take over a part of the specification. “We don’t have that; nobody has stepped up,” he said.
23:47
<Lachy>
http://www.sdtimes.com/content/article.aspx?ArticleID=32067 (2nd last paragraph)
23:47
<Hixie>
roc: you gave more than one?
23:47
<roc>
no, which student?
23:47
<Hixie>
oh
23:47
<Hixie>
andy
23:47
<Hixie>
works on indexing at google
23:48
<Hixie>
andy hochhaus
23:48
<roc>
ok, don't know him