09:37
<zcorpan>
i tried bananabench and ammo.js in firefox with asm enabled and then disabled, but didn't notice much of a difference
09:39
<zcorpan>
oh, that's because i have a mac
10:53
<Ms2ger>
'I used to be down on Web IDL, until I realised I could pronounce it "Web Idyll".'
10:53
<Ms2ger>
Maybe we should tell TC39 that
11:14
<hsivonen>
https://s3.amazonaws.com/f.cl.ly/items/2L0a0B0R130G3u0k0b1y/Layering.pdf looks appealing but fails to discuss the benefits of CSS layout and painting not having to allow JS to plug into points inside the layout and painting process
11:53
<annevk>
hsivonen: From talking to wycats__ and slightlyoff it's not necessarily about exposing all the guts. First defining the layering better and second seeing what needs to be exposed and how. The hope is that implementors scream for future-hostile proposals.
11:53
<annevk>
Which seems pretty reasonable to me.
11:53
<annevk>
And I think it's something we've been doing to some extent already, although a lot of the focus has been on actually defining things in the first place too...
16:17
<dglazkov>
good morning, Whatwg!
16:22
<asmodai>
Well, the HTML5 DRM thing now reached defective by design
16:22
<asmodai>
http://www.defectivebydesign.org/no-drm-in-html5
17:05
<wycats__>
hsivonen: it's not about exposing the guts
17:05
<wycats__>
in fact, it's explicitly about NOT exposing the guts
17:05
<wycats__>
it's about defining a layered model that's architecturally sound but which may be implemented and optimized in different ways
17:05
<wycats__>
which means NOT exposing every single C++ function as a JS hook
17:05
<wycats__>
(that's what we're doing now :/ )
17:37
<Hixie>
asmodai: talk about missing the point
17:55
<Hixie>
anyone know how the content model is broken here? https://www.w3.org/Bugs/Public/show_bug.cgi?id=19650
17:55
<Hixie>
i don't understand the bug report
18:03
<Hixie>
cabanier: ping https://www.w3.org/Bugs/Public/show_bug.cgi?id=19818#c1 last paragraph
18:04
<Hixie>
Ms2ger: ping https://www.w3.org/Bugs/Public/show_bug.cgi?id=17844
18:19
<JonathanNeal>
Do web browsers download the first half of a video if I tell it to play from the second half on?
18:20
<JonathanNeal>
Say I tell a 10 minute video to play from the 5 minute mark, after, say, the loadedmetadata event, can it skip ahead without downloading the first 5 minutes?
18:20
<Hixie>
depends on many things
18:21
<Hixie>
browser, protocol used for downloading the file, server capabilities, etc
18:21
<JonathanNeal>
Hixie: are these written down somewhere in a friendly fashion? even just for a few of them, like which browsers support downloading fragments?
18:21
<Hixie>
no idea
18:32
<Ms2ger>
Hixie, as I said last time, I know nothing :)
18:33
<Hixie>
say it in the bug, otherwise i forget :-)
18:33
<Hixie>
sorry :-)
19:24
<asmodai>
Hixie: I didn't create it :)
19:24
<asmodai>
Hixie: But that's what the term DRM brings to the table nowadays, I guess.
20:17
<cabanier>
hixie: thanks!
20:27
<MikeSmith>
Hixie: I suspect the commenter doesn't understand "then transparent"
20:28
<MikeSmith>
what you have is clear enough if you understand what "transparent" means in the spec
20:28
<MikeSmith>
but I can imagine it being not very clear if you don's
20:42
<MikeSmith>
I wish we could get the FSF as interested in examining the W3C document license in the context of what kind of copyright/license for core Web standards is in the public interest.
23:17
<danbeam_work>
dglazkov, Hixie, hober, jgraham, matjas, othermaciej, ShaneHud1on, TabAtkins, tantek can haz wiki.whatwg.org account to ever-so-slightly edit RequestAutocomplete spec?
23:17
<danbeam_work>
(http://wiki.whatwg.org/wiki/RequestAutocomplete)
23:18
<othermaciej>
I don't think I have the power to make wiki accounts
23:18
<danbeam_work>
you might want to remove yourself from here then http://wiki.whatwg.org/wiki/IRC
23:18
<danbeam_work>
othermaciej: ^
23:18
<danbeam_work>
as it's in the header to ping somebody on that list
23:19
<danbeam_work>
doh, wrong list
23:19
<othermaciej>
wouldn't you want someone labeled "administrator" on this list? http://wiki.whatwg.org/wiki/Special:ListUsers%26group%3Dautoconfirmed
23:19
<danbeam_work>
yes, just found that, my bad
23:20
danbeam_work
doesn't see the point of multiple valid link colors
23:20
<danbeam_work>
that link looked invalid
23:20
<danbeam_work>
hsivonen, also ^ (can haz wiki.whatwg.org account?)
23:21
<othermaciej>
as a side note, it would probably be helpful to for the proposal to explain use cases
23:21
<danbeam_work>
othermaciej: agreed
23:21
<danbeam_work>
othermaciej: I'll borrow from some of the other emails we've sent and add on to that for #requestAutocomplete() (i.e. payments for now)
23:23
<othermaciej>
I would have thought autocomplete should always be triggered by the UA or the user, not the website
23:23
<danbeam_work>
(as soon as I get an account, of course)
23:24
<danbeam_work>
othermaciej: http://lists.w3.org/Archives/Public/public-whatwg-archive/2012Oct/0274.html "Some examples:"
23:25
<danbeam_work>
this isn't like autofill or the preview of autofill in chrome now
23:25
<danbeam_work>
it pops a modal dialog, which I wouldn't trust a UA to know what works for all sites' use cases
23:25
<othermaciej>
oh hey, I already commented on the email thread on this, assuming it is the same requestAutocomplete()
23:25
<danbeam_work>
yes
23:26
<danbeam_work>
it was before I mentioned the use cases, though, IIRC
23:27
<danbeam_work>
othermaciej: also, re: autofill vs. autocomplete, complete and fill seem to be used interchangeably here and HTML spec calls out autocomplete= for individual fields *and* a whole form, so I doubt they'll be cleanly separated any time soon
23:28
<othermaciej>
I still think it's a bad idea to let websites initiate autocomplete/autofill as per my earlier message
23:28
<danbeam_work>
othermaciej: there's a few other points I think we responded to from your email - http://lists.w3.org/Archives/Public/public-whatwg-archive/2012Nov/0105.html - it was a while ago, can't really remember
23:28
<othermaciej>
but I guess that can be discussed on the list
23:28
<danbeam_work>
othermaciej: probably the best place
23:29
<danbeam_work>
othermaciej: I should also mention that it's based on user gesture, so it's not solely up to a site
23:29
<danbeam_work>
othermaciej: it's up to both
23:29
<danbeam_work>
othermaciej: as it is currently (at least a site author has the ability to *not* to autofill with autocomplete=off)
23:32
<othermaciej>
probably not useful to discuss it here since I'm sure many people have an opinion