01:45
<Hixie>
any opera people here? i'm trying to work out how to complete this sentence:
01:45
<Hixie>
Chaals could improve the Opera intranet if he had a mechanism for identifying the original source of various parts of a page, because...
01:46
<Hixie>
it's based on something chaals said but i don't seem to be able to find anything about how knowing the original source of a page would actually help improve the intranet
04:51
<Hixie>
i love this use case (paraphrasing): "how come if a tv guide site doesn't include a link to imdb my browser can't detect that it's a tv show from the microdata the author puts in the page and create an implied link to imdb?"
04:53
<Hixie>
if the tv guide author isn't including the link you want... what makes you think he'll include the microdata you want?
04:59
<MikeSmith>
Hixie: because experience has shown us that right-thinking authors will always go out of their way to embed hidden content that's only intended to be exposed to users of specialized tools, but not to normal users
04:59
<MikeSmith>
and they will never make any mistakes when they do it
05:00
<Hixie>
well actually some of the summary="" data did show that that does hapen sometimes!
05:00
<Hixie>
they have to be pretty motivated authors though
06:47
<MikeSmith>
Hixie: am I'm reading it incorrectly, or does abarth's content-sniffing data indicate that 0x00000100 signature almost never actually triggers in practice?
06:47
<MikeSmith>
http://webblaze.cs.berkeley.edu/2009/content-sniffing/metrics.html
06:48
<Hixie>
the metric counter is probably not being triggered for favicons
06:48
<Hixie>
but you'd have to speak to abarth for details
06:49
<MikeSmith>
I see
06:49
<MikeSmith>
just asking because that seems the be the only remaining case where Gecko content-sniffing behavior is different from the spec
06:50
<MikeSmith>
and abarth recommendation in the related Mozilla bug is that signature just be removed
06:51
<MikeSmith>
https://bugzilla.mozilla.org/show_bug.cgi?id=465007
07:13
<MikeSmith>
http://twitter.com/200ok/status/1592050739
07:13
<MikeSmith>
"why isn't NOSCRIPT valid inside HEAD? there are perfectly legit uses..."
07:14
<MikeSmith>
noscript is allowed in head, right?
07:16
<MikeSmith>
ah, I guess it's not allowed in HTML4
08:02
<Hixie>
Lachy: have you heard any news of a a requiem that works with the latest iTunes?
08:33
<Lachy>
Hixie, no
08:34
<Lachy>
Hixie, my only source of information is that forum thread. So if you checked there, I'm unlikely to know anything more than you
08:35
<Lachy>
but I just started freenet to see if the page has been updated since the last version was released
08:46
Philip`
wonders why they don't just publish all the news and updates on, like, a web page
08:50
<Lachy>
Philip`, because Requiem is techically illegal to distribute, and so an ordinary web page is subjected to DMCA take down notices and subsequent legal issues for failing to comply
08:50
<Lachy>
Using Freenet helps to avoid that by keeping anonymous
08:51
<olliej>
Lachy: requiem?
08:51
<Lachy>
it's illegal because of the anti-circumvention clauses in the DMCA and similar laws in other countries
08:51
<Lachy>
olliej, an iTunes DRM removal utility
08:53
<olliej>
happily at least some of that drm is gone
08:57
Philip`
recently discovered get_iplayer which can download BBC programmes via the streaming Flash RTMP protocol, including a few in HD (1280x720 3Mbps H.264), which is pretty nice, and much more useful than the official Flash interface
08:58
<Philip`>
(They have a cross-platform programme downloader built on Adobe Air, but Air only installs onto deb/RPM-based Linuxes, not Gentoo, so I can't use it)
08:59
Philip`
would prefer it if they didn't have to make things so unnecessarily complex (rather than e.g. providing a download link to an .mp4 file)
10:04
jgraham
realises that one of his scripts has broken because he installed a namespace-aware html5lib
10:05
<jgraham>
except s/one/many/, just that I haven't tried to use the others yet
10:27
<annevk2>
lol
10:27
<annevk2>
check http://www.stockholmgamlastan.se/ in Opera
10:28
<jgraham>
It would be much less funny if the photographs on the site weren't so crappy
10:29
<krijnh>
Nice, simpel protection :)
10:29
<svl>
heh
10:29
svl
toggles dom.event.contextmenu.enabled and happily rightclicks a couple of times
10:29
<krijnh>
That'll teach them!
10:30
<olliej>
wow, that's stupid
10:30
<olliej>
errr
10:30
<MikeSmith>
annevk2: hallvord should write a patch for browser.js that unblocks opera and then replaces all the images on the pages with lolcats or something
10:31
Philip`
left-clicks and drags the image into Firefox's location bar, and then uses Ctrl+S
10:32
<MikeSmith>
speaking of stupid: http://www.torchmobile.com/blog/?p=29
10:33
<jgraham>
Philip`: Why not jut drag the image to your folder of choice?
10:40
<Philip`>
jgraham: Because I don't trust drag-and-drop to work in any scenario more complex than from one application to itself
10:41
<Philip`>
(Actually it often works between native KDE applications too, but Firefox isn't one of those so I still wouldn't trust it)
10:43
<annevk2>
http://sourcefrog.net/projects/meantime/ is interesting
10:43
<annevk2>
basically cookies through HTTP caching
10:44
<annevk2>
so you'd have to disable your cache altogether if you want to prevent tracking
11:11
<annevk2>
it might also explain why Hixie got such strange values back for Last-Modified although somehow I doubt a lot of sites would be using the technique
11:16
svl
set his cache to 2MB a couple of years ago as a compromise between mostly preventing this and still having some use of the cache while staying within a single website.
11:18
Philip`
enables cookies and uses a mostly-static IP, so people don't have to bother wasting time on such tricks to track people
13:28
<Lachy>
re the right clicking on stockholmgamlastan.se , I guess they're not aware that Firefox has had the option to prevent context menu blocking for years
13:32
<annevk2>
or aware of browser caches...
13:34
<Lachy>
generally, people who try to prevent right clicking just want to use it as a deterrent
13:34
<jgraham>
Or drag and drop of view-page info or any of the other hundreds of ways of getting at those images
13:34
<Lachy>
besides, simply dragging the image from the browser to the desktop is the quickest way to save images now anyway
13:35
<Philip`>
The illusion of security is more valuable than security itself
13:36
<jgraham>
Philip`: To whom? In this case I would prefer the ability to access the site and use the context menu than the ability to save the files
13:37
<Philip`>
jgraham: To the people who are choosing to implement illusory security, I guess
13:38
<Philip`>
(I am, of course, entirely wrong)
13:38
<Lachy>
it would be interesting to see some sort of usability study that analysed common techniques real users used when saving images, to find out if using the context menu is still a significantly relevant technique compared with others
13:38
<Philip`>
(because they'd prefer real security, but the cost is too high (since it's impossible), so they make do with the best can they easily implement)
13:39
<Philip`>
s/can they/they can/
16:04
<annevk2>
hurray: http://www.w3.org/TR/css3-mediaqueries/
16:09
<Philip`>
I assume someone has already pointed out that the brackets in media="print and (color)" look weird and stupid and unintuitive? :-)
16:10
<annevk2>
not sure, but at this point the comment comes about seven years too late
16:11
Philip`
wonders why it's restricted to disjunctive normal form
16:13
<mpilgrim>
recommendations for "this week" happenings?
16:13
<mpilgrim>
i'm covering the datagrid and the keygen elements
16:13
<annevk2>
HTML5 was just published
16:13
<mpilgrim>
ah, indeed it was
16:13
<mpilgrim>
and presumably your html-differences document as well?
16:13
<annevk2>
also the drafts that were splitted out from HTML5 as well as Web Workers have been published too
16:13
<annevk2>
yes it was
16:13
<mpilgrim>
ooh
16:13
<mpilgrim>
excellent
16:13
<hsivonen>
mpilgrim: it looks a lot like there's going to be a parsing quirk (unless Hixie changes Acid2 and all browsers agree)
16:14
<mpilgrim>
is this the <p><table> thing?
16:14
<hsivonen>
mpilgrim: yeah
16:14
<mpilgrim>
yes, i saw your blog post on it
16:14
mpilgrim
adds that link to the list of "around the web"
16:14
<hsivonen>
mpilgrim: zcorpan's public-html post has the conclusion, though
16:15
<mpilgrim>
checking
16:15
<Philip`>
annevk2: It may have been seven years but pretty much nobody uses it, so it'd be easy to change without breaking compatibility much :-)
16:16
mpilgrim
can't tell if Philip` is joking
16:16
<billyjackass>
mpilgrim: dunno if you'd find this newsworthy
16:16
<mpilgrim>
it's used quite often for mobile stylesheets for iphone, ipod, android
16:16
<billyjackass>
http://hg.mozilla.org/mozilla-central/rev/e6812f79d2ba
16:16
<billyjackass>
"Harmonize content sniffing in HTML5 and Firefox"
16:16
<billyjackass>
abarth checking in on April 5
16:16
<mpilgrim>
billyjackass: absolutely!
16:17
<Philip`>
As far as I can see, the only examples more complex than comma-separated tokens are (I guess) trying to select mobile devices based on screen width
16:17
<Philip`>
(and they're not using the weird single-token-in-parentheses syntax)
16:17
<billyjackass>
mpilgrim: https://bugzilla.mozilla.org/show_bug.cgi?id=465007 is the related bug
16:17
mpilgrim
first learned about css media queries from an article on iphone-specific stylesheets
16:17
<Philip`>
1 only screen and (max-device width:480px)
16:17
<Philip`>
31 only screen and (max-device-width: 480px)
16:17
<Philip`>
1 screen and (min-device-width: 481px)
16:17
<Philip`>
1 screen and (-webkit-min-device-pixel-ratio:0)
16:18
<Philip`>
(Total occurrence counts, and values, from 130K pages)
16:18
<mpilgrim>
i was ready to rant about how apple was making shit up again and polluting the web, until i checked discovered that it was a standard that nobody else was using
16:18
<annevk2>
:)
16:18
<annevk2>
it's been in Opera for a long time
16:18
<mpilgrim>
Philip`: that's interesting
16:19
<annevk2>
Firefox 3.5 has it too
16:19
<Philip`>
Polluting the web with standards isn't really any better than polluting the web with proprietary technologies
16:19
<hsivonen>
mpilgrim: media queries are nice with Opera Mini, too
16:19
<mpilgrim>
Philip`: blasphemy!
16:19
<mpilgrim>
;)
16:19
<Philip`>
(They might be a bit better documented but that's the only real difference, and they equally add to complexity and not-working-in-everyone's-browser-ity and so on)
16:22
<mpilgrim>
where is the new html5 draft published?
16:22
<mpilgrim>
http://www.w3.org/TR/html5/ still lists the feb 12 draft
16:23
<Rik|work>
http://www.w3.org/TR/2009/WD-html5-20090423/
16:23
<annevk2>
seems the latest version links aren't updated yet
16:23
<annevk2>
MikeSmith, ^^
16:23
<annevk2>
it is however announced on http://www.w3.org/
16:23
<Philip`>
Hmph, the multipage splits are all stupid now
16:24
<MikeSmith>
yeah, TR symlinks are not set up yet
16:24
<MikeSmith>
webmaster is working on it now
16:24
<annevk2>
hmm, webstorage, websockets, and workers
16:24
<annevk2>
nobody did a sanity check on the shortnames? :)
16:24
<MikeSmith>
annevk2: what's wrong with the shortnames?
16:25
<mpilgrim>
Rik|work: thanks
16:25
<annevk2>
MikeSmith, nothing much, but if either all were prefixed with web or none would've been better (preferably all, given webidl)
16:26
<MikeSmith>
annevk2: point taken
16:27
<Philip`>
Hmm, www.w3.org actually links to http://www.w3.org/TR/2009/WD-html5-20090421/ which redirects to ...23
16:39
<MikeSmith>
the /TR versions are up now
16:39
<MikeSmith>
Philip`: I'll get that fixed
16:55
<remysharp>
Is there any way, or current discussion on the audio and video tags to prevent the browser from downloading the content?
16:56
<gsnedders>
No
16:56
<remysharp>
I'm pretty new to how this works, so please humour me - but how do I start that discussion (or is it just via here!)?
16:57
<gsnedders>
remysharp: http://wiki.whatwg.org/wiki/FAQ#Is_there_a_process_for_adding_new_features_to_the_spec.3F
16:57
<annevk5>
yeah, don't like to any resource
16:59
<remysharp>
annevk2: I mean to prevent the tag from saturating the bandwidth when the tag's in heavy used on the page
16:59
<gsnedders>
remysharp: Why are you using it without wanting it to be downloaded?
16:59
<remysharp>
but then, thinking out loud, there's no way to say start loading without explicitly using some JS function
16:59
<remysharp>
gsnedders: if there's 50 audio tags on the page
16:59
<remysharp>
and they all come down at once
16:59
<gsnedders>
remysharp: But why would there be?
16:59
<remysharp>
then my bandwidth is going to go loopy
16:59
<remysharp>
one second - I'll give you a link
17:00
<remysharp>
http://huffduffer.com/tags/sxswi2009
17:00
<remysharp>
so this page - not 50, but a lot right?
17:00
<remysharp>
these used to be audio tags
17:00
<remysharp>
and it would fail back to flash for the play buttons
17:01
<annevk5>
that seems like something the browser should optimize
17:01
<Rik|work>
http://www.whatwg.org/specs/web-apps/current-work/multipage/video.html#attr-media-autobuffer
17:01
<remysharp>
as in, don't download them in parallel?
17:01
<Rik|work>
remysharp: isn't it the purpose of autobuffer ?
17:01
<annevk5>
remysharp, as in, do whatever is best for the user
17:01
<remysharp>
Rik|work: just reading that now...
17:02
<remysharp>
so would I set this to false to tell the UA to not go and get all this data?
17:02
<remysharp>
(for example)
17:03
<Rik|work>
from what I read, it is just a hint
17:03
<Rik|work>
maybe a browser could first download media with autobuffer=true and then those with false
17:04
<remysharp>
again, just thinking out loud, but what about when mobile devices support html5 -
17:04
<remysharp>
bandwidth is a cost for some plans
17:04
<remysharp>
wouldn't we want to be able to prevent this?
17:04
<remysharp>
I guess this could boil down to the UA having to have a setting in the phone I guess.
17:04
<jmb>
is this not just something which should be left to UAs to sort out?
17:04
<remysharp>
(I guess...I guess)
17:05
<remysharp>
I'm starting to see that for the mobile case, yes.
17:07
<gsnedders>
remysharp: It's a problem for any number of images, videos, and audio files
17:08
<remysharp>
gsnedders: yeah, you're definitely right, and putting images in that context, kind of makes me see that it's definitely down to the browser.
17:08
<remysharp>
that said...
17:09
<gsnedders>
remysharp: Limited bandwidth is not a new problem, we already had it with img (esp. in the days of dial-up being normal)
17:09
<remysharp>
didn't the image tag used to have a low-rez attribute or something...
17:09
<remysharp>
like a preview on the image tag...am I making this up??
17:09
<gsnedders>
remysharp: Um, maybe HTML 3.0? It did sometime. Nobody ever implemented it. Nobody cares about it.
17:09
<remysharp>
:D
17:09
<remysharp>
okay, I was harking back a bit then!
17:10
<gsnedders>
Before I had ever used the web, I think :)
17:10
<Philip`>
Nobody implemented it?
17:11
<Philip`>
I'm sure I remember it as actually working
17:11
<Philip`>
(a long long time ago)
17:11
<gsnedders>
Well, nobody somebody. Nobody relevant with enough influence, though.
17:11
<gsnedders>
s/nobody/maybe/
17:11
<remysharp>
Philip`: I'm pretty sure it did, I'm sure I used it back then
17:11
<Philip`>
Apparently it worked in Netscape
17:11
<remysharp>
nobody relevant == me :)
17:11
<Philip`>
who had influence
17:11
<Philip`>
http://www6.uniovi.es/gifanim/gifhtml.htm
17:12
<Philip`>
"Only Netscape (that I know of) supports the Netscape extension of LOWSRC"
17:12
<remysharp>
that was the one lowsrc
17:13
<Philip`>
I guess that technique for GIF animation didn't take off
17:13
<remysharp>
I just wonder whether there's a usecase for some attribute like 'loadondemand'
17:13
<remysharp>
(for video + audio)
17:47
<annevk5>
remysharp, you start with the use case, not the solution
17:47
<annevk5>
remysharp, i.e. you're doing it wrong :)
17:47
<remysharp>
annevk5: okay, so I should start with my own experience when I visited the page, and it tried to load all the audio thus slowing my connection and killing my experience
17:48
<remysharp>
it's a bit over the top - but that's the point you're making right
17:48
<remysharp>
how do we make sure the experience is still fast and not hammered by lots of things downloading that I don't want.
17:48
<annevk5>
by getting a better browser
17:49
<remysharp>
it's not the browser - it was my connection
17:49
<Philip`>
remysharp: That seems like a reasonable statement of the problem
17:49
<remysharp>
Philip`: okay, cheers.
17:49
<remysharp>
at least I post together my thoughts on it, *then* it can be poo-pooed :-)
17:50
<remysharp>
I take back what I said about connection - you could be right, maybe it's the browser, perhaps not. That I can test myself fairly easily.
17:51
<annevk5>
my point is that the browser is more likely aware of the user's available bandwidth than the website owner
17:51
<Philip`>
I guess the solutions are something like (1) require page authors to add an attribute to most of their audio elements when they've got quite a few and they think it might be bad for most of their users; (2) make browsers manage resources sensibly, by not opening a hundred socket connections at once or downloading dozens of megabytes automatically or whatever; (3) something else
17:51
<annevk5>
so if someone (between the browser and website owner) needs to be in control of bandwidth usage (and disk usage, and memory usage, etc.) it's the browser
17:52
<Philip`>
and (2) seems a solution that's likely to work better for users
17:52
<remysharp>
Sure, but as a developer of page, I'm creating a listing page with video tags, I know the user doesn't want them all at once down the wire though.
17:52
<remysharp>
(in response to annevk5)
17:52
<annevk5>
yeah, and the browser can know that too
17:53
<remysharp>
Philip`: I definitely agree that browsers should need to be wary of concurrent downloads for audio and video tags -
17:53
<Philip`>
I assume browsers already apply their 6-connections-per-domain-name limit to media downloads?
17:53
<remysharp>
but, right now, that's limited to 2 for IE and 4(?) for FF (I think?)
17:54
<remysharp>
I thought FF was on 4.
17:54
<Philip`>
remysharp: I think it's 6 in IE8 and recent FFs etc
17:54
<remysharp>
Right, cool.
17:54
<Philip`>
at least for normal page content
17:54
<Philip`>
("recent" might mean "not yet released"; I'm not at all sure)
17:54
<remysharp>
I guess that would actually be a big problem point then too
17:55
<remysharp>
big-ish
17:55
<remysharp>
because if there's images or scripts below the 6th video, they're all going to hang if they're the same domain
17:55
<remysharp>
obviously solved by splitting your content - but I don't think that's the point
17:55
<Rik|work>
http://stevesouders.com/ua/
17:55
<Rik|work>
check the conns/host column
17:56
remysharp
bookmarks - cheers
17:56
<Philip`>
Rik|work: Nice, thanks
17:56
<Rik|work>
btw, webkit will be 6 conns/host soon
17:57
<Philip`>
remysharp: Only while the videos are actively downloading - the connections should be released when it's stopped buffering, and it probably shouldn't be buffering the entire video before you've started playing it
17:58
<remysharp>
are you saying that they would buffer a bit, then move on to the next video then?
17:58
<Philip`>
I suppose I don't see why they'd buffer anything at all
17:58
<Philip`>
other than the poster image
17:59
<remysharp>
Yeah, that's the problem though - in WebKit, last I saw, the just download the whole shebang
18:00
<remysharp>
which is why Jeremy's site (the link I posted above) was changed from the audio tag to just flash
18:15
<Philip`>
mpilgrim: "Adam Barth's [PDF] whitepaper" seems misleading, since I interpret 'whitepaper' as being a marketing document (and Wikipedia seems to agree so I must be right), whereas this thing is an actual proper academic paper (so I'd just call it a 'paper')
18:16
<annevk5>
proper response :)
18:17
<Philip`>
(?)
18:17
<annevk5>
joking
18:17
<annevk5>
<datagrid> is also not quite as new as mark said
18:18
Philip`
is confuzzled
18:18
<annevk5>
you'll get over it
18:18
<Philip`>
I may
19:35
<gmiernicki>
<audio> +1 <video> +1
19:35
<gmiernicki>
flash -1
21:47
<gsnedders>
Hixie: You _would_ just reference Avenue Q in your email.
21:47
gsnedders
now really wants to see Avenue Q after all your references to it
21:47
<Hixie>
i've seen it THREE times
21:47
<Hixie>
it's that good
21:47
<zcorpan_>
http://www.w3.org/TR/2009/CR-CSS2-20090423/ still hasn't fixed "BODY element in XHTML" :-(
21:48
<gsnedders>
Hixie: I mean, I was thinking about going to London in June, and there's a production there…
21:53
<annevk5>
zcorpan_, really?
21:53
<annevk5>
zcorpan_, might have been a resolved issue without edits, but I actually thought the edits were made too
21:54
gsnedders
realizes how crazy his June will be
21:54
<zcorpan_>
annevk5: if an issue is resolved but not changed in the spec, could the spec go to rec without the change on the basis that there are no unresolved issues?
21:54
<annevk5>
zcorpan_, the actual text is updated I think
21:54
<gsnedders>
I have an exam on the 4th, I am away 12th to 14th at leadership training thing, and Hadrian's Wall trip is 19th to 21st, then I start internship on the 29th… And I was going to go to London too…
21:55
<zcorpan_>
oh i didn't check further than the "features at risk" section
21:55
<annevk5>
zcorpan_, "For documents whose root element is an HTML "HTML" element or an XHTML "html" element [...]"
21:55
<annevk5>
zcorpan_, I guess somebody forgot to remove that text there
21:55
<zcorpan_>
ok
21:55
<zcorpan_>
i guess they'll notice
21:58
<annevk5>
zcorpan_, I guess the spec could go to rec btw if nobody noticed
22:07
<Hixie>
i wonder if the e-mail i sent will trigger 15,000 replies, or zero.
22:07
<Hixie>
i guess i'll poke at websocket then get back to microdata
22:27
<gsnedders>
Weee… More bug reports of XML parse errors!
22:27
<gsnedders>
annevk2: Finish XML5, kthxbai.
22:32
<Philip`>
XML parse errors should be rebranded as "XML parse bonuses", so people don't think it's so bad to encounter one and report them as bugs
22:33
<jmb>
that'll lead to competition to get the highest bonus score, though
22:34
<annevk5>
gsnedders, it seems hardly anyone cares anymore about XML these days
22:34
<annevk5>
gsnedders, I mean sure, there's the occasional issue, but nothing compared to CSS/HTML/DOM/ECMAScript
22:34
<gsnedders>
annevk5: Should we just use Aaron Swartz's suggestion of RSS 3.0?
22:34
<gsnedders>
http://www.aaronsw.com/weblog/000574
22:34
<annevk5>
gsnedders, I'd rather we just give up on feeds and use HTML; HTML files are not that much larger
22:35
<gsnedders>
RSS 3.0 is simpler.
22:35
<annevk5>
gsnedders, though meanwhile I guess we're stuck with several flavors of RSS and Atom
22:35
<gsnedders>
(Read that page.)
22:36
<annevk5>
(I know RSS 3.0.)
22:36
<gsnedders>
(Ah, OK.)
22:36
<gsnedders>
(It wasn't obvious.)
22:38
<jwalden>
(I wonder why all recent lines of discussion have been parenthesized.)
22:41
<Philip`>
(It helps prevent the characters from falling off the edges of the sentence and making a terrible mess on the channel's carpet)
22:43
<Philip`>
(Well, it's either a carpet or a thick layer of hair and biscuit crumbs, and I'm not quite sure which)
23:23
<annevk5>
http://sharovatov.wordpress.com/2009/04/22/relcanonical/ more evidence as to why rel= fails
23:29
<Dashiva>
annevk5: Heh, I didn't even notice the error until I went to read the comments on your post
23:29
<annevk5>
and I made a mistake too, I meant to say rev=
23:30
annevk5
blinks