02:09
<karlcow>
http://benfry.com/writing/archives/597
02:12
<AryehGregor>
Interesting.
02:12
<AryehGregor>
Computers in other games have much more human-like play styles, I think. Like backgammon.
04:25
<Traveler>
morn
04:40
<jtbandes>
I'm trying to use font-variant:small-caps; on a @font-face embedded font which does not have a specific small caps style. Sometimes font-variant:small-caps; seems to "fake" small caps, and on some fonts it doesn't. Should/when should I expect it to work? What alternatives do I have?
08:25
<Dashiva>
> But he stressed that if he did a deal with one content owner, any other content owner would be able to get exactly the same deal at the same price. "Non-discriminatory conditions are very important."
08:26
<Dashiva>
How do you guarantee non-discriminatory terms for a deal like "Give my component highest priority in the network"?
09:55
<MikeSmithX>
r4730 is a pretty significant change
09:56
<MikeSmithX>
http://html5.org/tools/web-apps-tracker?from=4729&to=4730
10:00
MikeSmithX
wonders if hsivonen is around and if he has seen r4730
10:15
<hsivonen>
MikeSmithX: I hadn't seen it, but it had been a long time coming
10:17
hsivonen
has always wondered how high and low directions are supposed to go in the phrase "higher-level protocol"
10:17
hsivonen
thought an encapsulating container was lower in the stack
10:19
<MikeSmith>
yeah, it seems like the spec should say "lower-level protocol"
10:20
<MikeSmith>
hsivonen: maybe at least emit a warning for documents lacking a title element?
10:20
<MikeSmith>
or not?
10:21
<hsivonen>
MikeSmith: as far as I can tell, the spec still requires <title> for documents served over HTTP
10:22
<hsivonen>
and neither v.nu nor the w3c validator supports the validation of email messages
10:22
<hsivonen>
so as far as I can tell, no cange to the schema is needed until srcdoc checking is implemented
10:22
<MikeSmith>
I see
10:22
<hsivonen>
*change
10:23
<MikeSmith>
hsivonen: the spec also mentions the iframe case
10:24
<MikeSmith>
oh
10:24
<MikeSmith>
sorry, that's what you already mentioned
13:39
<MikeSmith>
othermaciej: I have a question about the Webkit default user-agent stylesheet: I notice that it doesn't set text-align on the th element, but if I open Web Inspector and show the inherited properties, it indicates that th has text-align: center inherited from.. somewhere
13:39
<MikeSmith>
I don't understand where it's inheriting it from
13:40
<othermaciej>
MikeSmith: good point - lemme see if I can find out
13:51
<othermaciej>
MikeSmith: after looking for a while, I can't figure out how th elements get their centering
13:54
<othermaciej>
MikeSmith: my guess is that it does it in an odd way to avoid interfering with <th align=whatever>
13:54
<othermaciej>
MikeSmith: since a UA stylesheet rule would (I think) take priority over a presentational attribute, which is supposed to be treated at "presentational hint" priority
13:54
<othermaciej>
but, I can't see the code that does it at the moment
13:55
<annevk>
presumably th[align] would be implemented as a UA style sheet rule too
13:59
<othermaciej>
this is supposed to be the precendence for presentational hints: http://www.w3.org/TR/CSS2/cascade.html#preshint
13:59
<othermaciej>
I guess the only thing special about them is they have specificity 0
14:00
<othermaciej>
I do not think th[align] in the UA style sheet would correctly implement the requirements for "presentational hints"
14:00
<othermaciej>
that would have specificity 0,0,1,1
14:01
<annevk>
it's a ua style sheet, they're not part of the CSS cascade
14:01
<othermaciej>
I see, origin takes priority over specificity
14:01
<othermaciej>
I guess the only technical problem with putting them in the UA stylesheet is that they would override the user stylesheet
14:01
<othermaciej>
anyway
14:02
<othermaciej>
I can't tell where the text-align: center comes from through code inspection
14:02
<othermaciej>
I could debug it but I don't feel up to it right now
14:03
<MikeSmith>
othermaciej: thanks for looking at it
14:03
<MikeSmith>
I had just been wondering if there was something obvious I was missing, or misunderstanding
14:04
TabAtkins
is late to the party, and just now realized that Opera's implemented Transitions and 2D Transforms.
14:05
<othermaciej>
table cells can inherit some properties from their table in a weird way but I don't think that's what is causing this
14:05
<othermaciej>
oh I found it
14:05
<othermaciej>
this is hella weird
14:06
<othermaciej>
// Table headers with a text-align of auto will change the text-align to center.
14:06
<othermaciej>
if (e && e->hasTagName(thTag) && style->textAlign() == TAAUTO)
14:06
<othermaciej>
style->setTextAlign(CENTER);
14:06
<othermaciej>
in CSSStyleSelector.cpp
14:06
<othermaciej>
I have no idea why it's done that way
14:07
<othermaciej>
I guess we want a table header to center even if you explicitly set text-align: auto?
14:10
<othermaciej>
"auto" isn't even a standard text-align CSS value
14:10
<othermaciej>
it's our implementation of the secret nameless initial value of text-algin
14:25
<jgraham>
*
14:26
<Philip`>
14:29
<annevk>
\
15:01
<MikeSmith>
othermaciej: thanks for the details on th
15:01
<MikeSmith>
I'm wondering if other engines maybe do something similar for th
15:02
MikeSmith
goes to look at gecko default ua stylesheet
15:08
<othermaciej>
I'm honestly not sure why it is done in this odd way
15:08
<othermaciej>
would have to ask hyatt
15:08
<MikeSmith>
othermaciej: fwiw, gecko default ua stylesheet also doesn't set text-align: center
15:08
<MikeSmith>
http://mxr.mozilla.org/mozilla/source/layout/style/html.css
15:08
<MikeSmith>
..on the th element, that is
15:08
<othermaciej>
it might just be to make <th align=auto> simpler to implement
15:24
<boblet>
hey all, what are context menus? Also, is the contextmenu attribute only for inputs?
15:26
<TabAtkins>
right-click menus, generally.
15:28
<boblet>
TabAtkins: aah, ok
15:29
<TabAtkins>
Also, no, @contextmenu is a global attribute.
15:30
<TabAtkins>
It's just an attribute that points at a <menu> somewhere in the document, which should be presented as a context menu in a UA-specific manner when the user triggers it in a UA-specific manner. (Which is just a popup menu when you right-click in typical UAs.)
15:30
<boblet>
they’d probably normally contain buttons huh
15:30
<boblet>
huh
15:31
<TabAtkins>
Not necessarily. They could contain <command> or whatnot too.
15:31
<TabAtkins>
\http://www.whatwg.org/specs/web-apps/current-work/multipage/interactive-elements.html#context-menus
15:31
<TabAtkins>
Check that example.
15:31
<boblet>
TabAtkins: that’s the trouble, I did :)
15:31
<TabAtkins>
Haha.
15:31
<boblet>
guess I don’t make enough web applications
15:32
<TabAtkins>
Nah, it's something that isn't used enough in typical webapps today.
15:32
<boblet>
ok, thanks for explaining
15:32
<TabAtkins>
But yeah, just assume that it's for producing custom right-click menus on specified elements. It happens to have usability/accessibility benefits beyond that which come for free.
15:33
<boblet>
the UAs will present in relevant way part? or something more?
15:34
<TabAtkins>
Yeah, they can present it in whatever way makes sense for them.
15:34
<TabAtkins>
And trigger it based on whatever interaction makes sense for them.
15:34
<boblet>
cool
15:35
boblet
pines for the web of pages
15:35
<boblet>
I could understand that one
15:35
<boblet>
:)
15:41
<TabAtkins>
Now you've got a web of applications, which happens to contain the web of pages as a strict subset.
15:44
<boblet>
TabAtkins: don’t forget the web of data. sheesh, where do I file a complaint that my knowledge is being made less relevant? ;)
16:35
<TabAtkins>
Man, I have zero self control when it comes to discounted video games. I went to Best Buy with a $100 bill and the intention of buying Mass Effect 1, expecting it to be $30 or $40. It was $20. So was Mirrors Edge. And Dragon Age was $40. I now have to make the difficult decision of which to play first.
16:35
<TabAtkins>
(Probably gonna be ME1.)
16:37
<Philip`>
$20?
16:37
<Philip`>
I got Mirror's Edge for about £3 on Steam
16:38
<TabAtkins>
My computer is weak and sickly.
16:38
<TabAtkins>
I must play these things on the xbox.
16:40
<TabAtkins>
I expect to parlay my new salary into a new desktop, since the last of my birthday just got spent on video games and extending my tattoo.
16:40
<Philip`>
Hmm, I can't see the exact price but it was £4.98 for Mirror's Edge plus The Secret of Monkey Island: Special Edition
16:41
<Philip`>
That reminds me, I've still got three Monkey Island games I need to play
16:41
<TabAtkins>
Though I am physically unable to play Oblivion on the xbox. Ffs, every fps in the entire world needs to come with the option to invert BOTH x and y. I can't play unless they're both inverted.
16:41
<Philip`>
Hold the controller upside down
16:42
<TabAtkins>
Then I have to do the thumbstick with my right hand, which is just wrong.
16:42
<Philip`>
Hold it behind your back so your arms are the other way round
16:42
<TabAtkins>
Then my hands are upside down.
16:43
<annevk>
http://twitpic.com/13rmjt has a pic of that Japanese HTML5 book
16:43
<TabAtkins>
That's a badass-looking book.
16:45
<Philip`>
TabAtkins: Alternative solution: View the game in a mirror
16:46
<TabAtkins>
That one would actually work, since I *can* invert the Y-axis.
16:53
<TabAtkins>
Ah, there we go. Hacking on Chatzilla's CSS is difficult when you can't use Firebug on it.
16:58
<svl>
dom inspector
16:59
<TabAtkins>
Does that work on arbitrary XUL apps, or just on ones running within Firefox?
16:59
<TabAtkins>
'cause I've got cz running standalone.
17:00
<paul_irish>
doesnt ChromeBug work well for that?
17:00
<TabAtkins>
I wouldn't have any idea.
17:03
<svl>
TabAtkins: should work on arbitrary xul apps (afaik, no personal experience)
17:04
<svl>
https://developer.mozilla.org/en/XULRunner_tips#DOM_Inspector has instructions (seems to be more convoluted than I'd thought/expected, but still doable)
17:05
<paul_irish>
check out chromebug, it's designed explicitly for hacking on XUL
17:14
<TabAtkins>
Ah, excellent. Chromebug is doing what I need.
19:07
<AryehGregor>
Am I the only one who thinks that the existence of Ctrl+scroll wheel in browsers probably causes many more users to find websites mysteriously the wrong size than it actually helps anyone?
19:07
<Steve^>
you are not
19:14
<annevk>
AryehGregor, it helps me
19:14
<AryehGregor>
Well, it probably doesn't *hurt* anyone in this channel, because we all know the magic Ctrl+0 trick.
19:15
<AryehGregor>
But I constantly see people asking how to reset their browser to the right zoom level.
19:15
<AryehGregor>
I use zoom shortcuts too, but would it really hurt to have them less accessible? Like Ctrl-Shift-mouse wheel or something?
19:16
<Steve^>
this is why Opera disables single key shortcuts by defaults - they are too easy to hit by accident
19:16
<Steve^>
its easy to be using ctrl for something and hit the scroll wheel by accident
19:19
<AryehGregor>
Oh, it does that for all single-key shortcuts?
19:19
<AryehGregor>
I recall it requires Shift-Esc or something for accesskeys, which always struck me as inconvenient.
19:19
<Steve^>
theres an option that simply re-enables single key shortcuts, like z for back and x for forward
19:25
<daedb>
I remember when Opera had a single-key shortcut for turning images on and off, it was bloody annoying :)
19:26
<AryehGregor>
Oh, so we were talking about double-key shortcuts, not single-key.
19:26
<AryehGregor>
Backspace is an annoying shortcut.
19:27
<Steve^>
I find it more annoying that my firefox doesn't have that shortcut
19:27
<tantek>
AryehGregor - strongly agreed, backspace is a very annoying shortcut.
19:27
<tantek>
really wish I could disable that
19:28
<tantek>
have lost content too many times due to that shortcut
19:28
<AryehGregor>
I've used backspace on a laptop before, when I didn't have a mouse with back/forward buttons.
19:28
<AryehGregor>
The back button on my current mouse hasn't worked for a long time, though, and I mostly just click the back button when I want to go back.
19:28
AryehGregor
should look into that.
19:28
<tantek>
e.g. you think you're focused in a text field, and want to delete characters you just typed, but it turns out the "document" has focus, and then backspace/delete goes back, and loses your textarea edits.
19:29
<Steve^>
your browser doesn't remember the field contents?
19:29
tantek
wishes he could patch is browsers to *not* treat backspace/delete as "go back" in navigation.
19:30
<tantek>
Steve^ depends on if the page was the result of a form submit etc.
19:30
<tantek>
then if you try to go forward, you get asked, do you want to resubmit? etc.
19:30
<tantek>
= dataloss
19:30
<tantek>
or worse, AJAX hacks to edit content that don't recover from back/forward
19:31
<AryehGregor>
Anyway, one of my first MediaWiki commits was removing an annoying shortcut. https://bugzilla.wikimedia.org/show_bug.cgi?id=8203
19:32
<Steve^>
apart from the AJAX wizardry, your browser doesn't account for form submit pages?
19:32
<AryehGregor>
tantek, well, if you use an open-source browser, you can!
19:33
<tantek>
AryehGregor - not without lots of work recompiling and merging with updates.
19:33
<AryehGregor>
Probably wouldn't have to do much work merging with updates, just recompiling. It's unlikely those lines would get changed often. But, yeah, just saying.
19:33
<Steve^>
Maybe I'll just point out that Opera has none of these problems, without you patching the browser?
19:34
<Steve^>
might be simpler in the short-term
19:34
<tantek>
seems to me like shortcuts should be in a separate config file anyway
19:34
<AryehGregor>
Yeah, every game out there lets you change all your shortcuts -- why not browsers too?
19:34
<tantek>
I'd be fine even with a "defaults write com.apple.webkit ... " type solution (for webkit, firefox etc.)
19:34
<Philip`>
Opera has a handy keyboard shortcut editing dialog so you can disable backspace there if you want
19:35
<tantek>
Philip` - nice - thanks I'll check that out.
19:35
<Philip`>
(in Preferences/Advanced/Shortcuts)
19:35
<Steve^>
hmm, Philip stealing my ideas!
19:35
<Steve^>
So, <video> replaces flash on the ipad. What's going to replace PDFs?
19:36
<Philip`>
Who wants to replace PDFs?
19:36
<Steve^>
i.e, when are we going to able to print properly from the net?
19:36
<AryehGregor>
PDF is a standard format, why do you need to replace it?
19:36
<Steve^>
doesn't work on my phone
19:36
<AryehGregor>
Just use a different implementation if you don't like Adobe's.
19:37
<Steve^>
if its a matter of time for it to get to all devices, then that's probably ok
19:37
<Steve^>
I assumed it was more of a royalty deal
19:37
<Steve^>
or implementation problem
19:37
<AryehGregor>
PDF can be implemented royalty-free.
19:37
<AryehGregor>
And is.
19:37
<Philip`>
There's already lots of reader and writer implementations
19:37
<AryehGregor>
Tons of programs support it.
19:37
<Steve^>
does Android have one?
19:37
<daedb>
My phone has a pdf reader installed by default.
19:38
<daedb>
I've never used it though :p
19:38
tantek
thought HTML+CSS was supposed to replace the need for PDF. ;)
19:39
<Steve^>
tantek, that's what you'd expect, but you can't print from the web - I can't do my CV in HTML
19:39
<AryehGregor>
Steve^, why not?
19:39
<Steve^>
it's ideal in that you can format exactly how you want and yet the browsers let you down
19:39
<AryehGregor>
I've done a resume in HTML before.
19:39
tantek
remembers helping to write some pretty damn good print (including print preview and size/image tweaking) in IE5/Mac ~ 10 years ago.
19:40
<daedb>
I write my CV in HTML and print it to a pdf...
19:40
<tantek>
Steve^ re: CV in HTML: http://microformats.org/wiki/hresume
19:40
<Steve^>
I'm speaking from experience in trying to get something I know will print to 2 pages
19:42
<Steve^>
ok, Opera printing a full-screen CV document: the right margin is about 3 inches wide
19:42
<Steve^>
same as the bottom
19:43
<Steve^>
ooo, chrome does a pretty good job
19:45
<TabAtkins>
I dont' think I've tried printing my resume before, I just send it around as a link. And it's all marked up with hresume. ^_^
19:46
<Steve^>
I think the printing is more for myself than employers
19:46
<Steve^>
I want to send a PDF around but Open Office is useless
19:47
<Steve^>
looks like Chrome produces a good quality doc though, so I might try this in the future
20:16
<ment>
Steve^: why would you want pdf replacement? there's no way to comfortably read documents with page-wide fixed layout on phone anyway.
20:16
<Steve^>
but I dislike meeting a dead-end when browsing the web
20:17
<ment>
what do you mean?
20:18
<Steve^>
I needed the days that my recycling is picked up from my site, the council gives the information as a pdf
20:18
<Steve^>
I browse their site, find the pdf and have to give up
20:19
<tantek>
sounds like an opportunity for a PDF->plain text (or semantic markup) web service
20:20
<ment>
steve^: that has nothing to do with PDF, they might as well supply the informations in .xls or dBase 3 format
20:21
<tantek>
Steve^ ask the council to support more and better OpenGov: http://tantek.pbworks.com/OpenGov - specifically through step 5
20:21
<Steve^>
pdf is the standard "I want the information laid out like this" format, just as flash is the "I want a video" format
20:24
<Philip`>
Google already does PDF -> HTML
20:24
<Philip`>
which often works okay
20:28
<ment>
Steve^: pdf is standard for "i want this document to be printed exactly like this", flash is the "there's no other way to include video in web page" standard
20:29
<Steve^>
Philip`, is there a way to easily access that function?
20:30
<ment>
pdf could be composed of scanned handwritten pages
20:34
<Philip`>
Steve^: Put e.g. cache:http://www.whatwg.org/specs/web-apps/current-work/html5-letter.pdf into the search box
20:36
<AryehGregor>
Argh, I have 110 threads in my spec inbox. :(
20:42
AryehGregor
just starts skipping all procedural threads
20:43
<AryehGregor>
And accessibility, and CSS. That cuts out about 90%.
20:55
<AryehGregor>
TabAtkins, "note, though, that I can count pretty high on two hands" -- using binary?
20:55
<TabAtkins>
Indeed.
20:57
<TabAtkins>
That line was a joke, but yes, I regularly count on my fingers in binary.
20:57
<TabAtkins>
It's very useful to be about to count to 30 on one hand, or 1000 on two.
20:57
Philip`
doesn't have enough independent control over each finger to count in binary :-(
20:58
<TabAtkins>
Really?
20:58
<TabAtkins>
Well, you can always just represent less digits. Find out which groups of fingers you *can* control independently, and just use them.
20:58
<Philip`>
I can't bend my little finger without the adjacent finger bending
20:59
<AryehGregor>
So you can only count to 256.
20:59
<AryehGregor>
TabAtkins, I've never actually found much use for it . . .
20:59
<AryehGregor>
If I need to count higher than ten, I usually go by fives and just remember how many fives I've done.
20:59
<AryehGregor>
Requires less brainpower.
20:59
<TabAtkins>
I use it to count how long to use my mouthwash, frex. ^_^
20:59
<TabAtkins>
And after you've done it for a while, it doesn't require any thinking at all. Muscle memory takes care of changing your fingers.
20:59
<Nikita>
so is the device spec ever going to be reviewed and/or possible to implement?
21:00
<TabAtkins>
Nikita: It just (a few weeks ago) got introduced at all. At some point, yes, more work will be done on it. It's just an idea for now.
21:03
<Nikita>
ok, who is involved in the committs and specs? Like if I or someone else wanted to committ code like that to the repository is that possible? I'm only now starting to get into the group
21:03
<TabAtkins>
Ian is the primary (and currently sole) editor of the spec. However, if you wanted to start proposing stuff, go ahead and do so. Just pop it onto the list.
21:04
<TabAtkins>
Right now, afaik, that feature is *extremely* experimental, and would benefit tremendously from some actual browser experimentation in the area.
21:05
<AryehGregor>
It's not even really specced, is it?
21:05
<Nikita>
ok, I would love to try to experiment with it, any guidelines? I think since its not a published spec? is there anything I need to do, like is there a resource on how to make pages with the latest code/experimental things?
21:05
<AryehGregor>
Not fully/properly, at least.
21:06
<Nikita>
not sure if this article is true http://www.ozcarguide.com/technology/web/2136-google-gears-html5
21:06
<Nikita>
but I do believe gears had an Audio API for recording, don't remember about video
21:06
<TabAtkins>
AryehGregor: Yeah, nothing's specced except for some very basic concept work on it. Basically "Hey, it should would be cool if there was a <device> element.".
21:07
<TabAtkins>
s/should //
21:09
<Nikita>
well since its just cool (it hasnt been updated since last year according to that side note thing on the left), is there any technology that can be used for live video recording and broadcasting WITHOUT flash/flex/silverlight (i was thinking a java applet might do it), but I reallly wish that spec was available
21:10
<TabAtkins>
Nothing exists yet, afaik.
21:11
<AryehGregor>
Java might work, no?
21:11
<AryehGregor>
I'm not sure what it supports.
22:08
<AryehGregor>
Yay, finished reading spec mail.
22:08
<AryehGregor>
Took an hour and a half. :/
22:09
<Philip`>
Was it worth it?
22:09
<AryehGregor>
Maybe.
22:09
<AryehGregor>
I guess so.
23:03
<Hixie>
TabAtkins: you didn't mail hte list
23:03
<TabAtkins>
Craps.
23:04
<Hixie>
:-)
23:04
<TabAtkins>
Danke, sent.
23:05
Hixie
learnt last week that the IETF submission process actually has a human in the loop
23:05
<Hixie>
and apparently submitting websocket so much is tiring them out
23:05
<Hixie>
:-/
23:05
<Hixie>
it's like the IETF stopped evolving in the 70s or something
23:21
<tantek>
Hixie every standards organizations' processes tends to be a time capsule of the culture/technologies in common use at their inception. Hence the need for new standards organizations every few years.
23:21
<Hixie>
seems that way
23:41
<AryehGregor>
Blargh, why does type="search" not support width in WebKit?
23:42
<Hixie>
width?
23:42
<AryehGregor>
Hmm, maybe I'm wrong?
23:43
<Hixie>
what do you mean by width?
23:43
<Hixie>
CSS property, IDL attribute, content attribute, something else?
23:43
<AryehGregor>
The width CSS property. But my accusation was scurrilous.
23:43
<Hixie>
ah ok
23:46
<AryehGregor>
Hmm, I guess scurrilous doesn't mean what I think it means.
23:47
<AryehGregor>
My accusation was groundless.
23:48
<Hixie>
is there a spec for how HTTP CONNECT works anywhere?
23:50
<Lachy>
Hixie, maybe this http://www.ietf.org/rfc/rfc2817.txt
23:50
<Lachy>
I haven't read it to see how well it defines it. just found it
23:50
Hixie
looks
23:55
<Hixie>
weird
23:55
<Hixie>
if i connect to hixie.ch port 80 and send:
23:55
<Hixie>
CONNECT hixie.ch:80 HTTP/1.1 [CRLF][CRLF]
23:56
<Hixie>
i get back a 400 Bad Request
23:56
<Hixie>
but if i connect to hixie.ch port 80 and send:
23:56
<Hixie>
CONNECT hixie.ch:80 garbage [CRLF][CRLF]
23:56
<Hixie>
i get back a 405 Method not Allowed