01:19
<MikeSmith>
Hixie: editorial -
01:19
<MikeSmith>
"The reversed attribute is a boolean attribute. If present, it indicates that the list is an ascending list. If the attribute is present, the list is a descending list."
01:20
<MikeSmith>
2nd sentence should start "If absent,"
01:52
<MikeSmith>
hmm, the algorithm for determining what an object element represents has a requirement to check the value of a "classid" attribute
01:52
<MikeSmith>
but nowhere is such an attribute defined
01:53
<MikeSmith>
I guess it's meant to be a content attribute
01:53
<MikeSmith>
but it's not in the content model for the object element
01:54
<Philip`>
The content model only lists conforming attributes, and classid is non-conforming
01:54
<MikeSmith>
ah
01:55
<MikeSmith>
Philip`: this is some MS proprietary attribute?
01:55
<Philip`>
so it's the same as how e.g. <isindex> is used in the parsing algorithm, but is not defined anywhere
01:55
<MikeSmith>
OK
01:55
<Dashiva>
Oh boy, more alt fun
01:57
<Dashiva>
If the designers of these tools cannot at this time
01:57
<Dashiva>
make their tools ATAG-compliant
01:57
<MikeSmith>
seems like for those kinds of cases, to avoid confusion (mine), the spec should explicitly mention that those are things that are not conformant for authoring purposes and so not documented in the spec
01:57
<Dashiva>
That line is such a wonderful thing
01:57
<Philip`>
MikeSmith: I think it's mostly used for ActiveX GUIDs, but I can see some people suggesting things like classid="java:Example.class" too but I don't know if that's implemented anywhere
01:58
<Philip`>
MikeSmith: How you can document something to say that it's not documented in the spec? :-p
01:58
<MikeSmith>
heh, true I suppose
01:59
<hsivonen>
I misread the commit log and thought Hixie made classid conforming :-(
01:59
<MikeSmith>
I think it would be appropriate for all instances of those kinds of things to be enclosed in <blink> tags
01:59
<hsivonen>
I think making classid non-conforming at this time does not make sense, even though it is product-specific
02:00
<Philip`>
MikeSmith: I agree it'd be nice to be explicit when things mentioned in the spec are non-conforming, because otherwise it's not at all clear that the spec author didn't just forget to add that attribute to the list
02:00
<MikeSmith>
Philip`: exactly
02:01
<hsivonen>
also, to me it would make sense for non-IE browsers to map the Flash classid to an NPAPI Flash plug-in, etc.
02:04
<csarven>
re: http://www.whatwg.org/specs/web-apps/current-work/multipage/section-global.html#embedding "So here we're saying that microformats shouldn't use this... is that what we want?" -- I suppose, but, I don't believe microformats is particularly interested in "Embedding custom non-visible data". One of its main goals is to mark "visible" data.
02:04
<csarven>
I don't know who wrote that or if there is a backlog for it
02:05
<Philip`>
csarven: Hixie wrote it
02:14
<csarven>
microformats is still open to finding a better way to do datatime-design-pattern. For a second there I thought data-* could assist that but realised it is probably not the right fit.
02:14
<csarven>
"User agents must not derive any implementation behavior from these attributes or values. Specifications intended for user agents must not define these attributes to have any meaningful values." -- I suppose this is to prevent authors from creating custom markup and relying on some UA behaviour.
02:16
<takkaria>
csarven: datetime-design-pattern is solved with <time>, no?
02:16
<csarven>
I don't know. Can you give me an example using <time>?
02:18
<csarven>
takkaria I see.
02:18
<takkaria>
csarven: <time datetime="2007-08-20T23:30Z">20 Aug</time>
02:19
<csarven>
Certainly.
02:20
<csarven>
As far as I can tell. The existing <abbr> approach would probably need to me replaced with <time>
02:20
<takkaria>
I believe that was one of the reasons <time> was invented
02:22
<Philip`>
As far as I'm aware, that was the only reason <time> exists
02:23
<csarven>
Then I'm not sure if data-* would at all be a use to microformats.
02:24
<Philip`>
If there was something equivalent to data-* but without the 'implementations must ignore these attributes' requirement, would that be useful to microformats?
02:29
<annevk>
csarven, I suggested to Hixie that data-* must not be used by UAs
02:29
<annevk>
csarven, this excludes UA extensions such as versioning and also excludes microformats
02:29
<annevk>
csarven, if microformats feel they need some attribute other than what they already got (such as a solution for time -> <time>) I think they can easily negotiate that with the HTML WG
02:30
<csarven>
Philip` Not necessarily because of the "visible" bit. data-* almost falls on meta-data.
02:30
<csarven>
annevk Right on.
02:30
<csarven>
<time> is perfect IMHO
02:31
<annevk>
Jacques Distler: "HTML5lib sucks." :-(
02:32
<annevk>
Oh, he's complaining about performance...
02:33
<annevk>
I think that's more or less a non-goal at this point. We need bindings to a C implementation to get that going strong...
02:33
annevk
adds a comment
02:33
<smedero>
Anne: are you talking about this post? http://golem.ph.utexas.edu/~distler/blog/archives/001689.html
02:36
<annevk>
yeah
04:12
<andersca>
Hixie: what was the rationale for making the dynamic entries ordered?
04:28
<hsivonen>
do C extensions to Ruby have to use UTF-8?
04:29
<hsivonen>
C extensions to Python should use UTF-16 or UTF-32, right?
04:29
<jgraham>
hsivonen: I think that's right (the python bit)
04:30
<jgraham>
http://docs.python.org/api/unicodeObjects.html
04:30
<hsivonen>
the difficult bit with recasting the V.nu tokenizer into UTF-8 C would be that REPLACEMENT CHARACTER would take more space than \0 and the tokenizer expects them to fit into the same space
04:31
<jgraham>
If that's the only difficult bit it doesn't sound too bad
04:31
<jgraham>
:)
04:37
<takkaria>
I should mention I'm being paid to work on a C html5 parser this summer
04:37
<hsivonen>
takkaria: what's your implementation strategy? from scratch? adapting an existing implementation?
04:39
<jgraham>
takkaria: I mentioned that on Jacques' blog
04:39
<jgraham>
Is it a Summer of Code thing?
04:39
<jgraham>
IIRC you are using lxml compatible interfaces?
04:39
<jgraham>
sorry libxml2
04:43
<takkaria>
yes, summer of code
04:44
<takkaria>
and there will be a libxml binding, yes
04:44
<takkaria>
hsivonen: well, jmb (on this channel) has started, so it's not entirely from scratch
05:09
<Hixie>
to anyone who sent feedback over the past few days, please send mail
05:10
<Hixie>
andersca: well it has to have some stable order for enumeration
05:10
<Hixie>
andersca: apart from that, i don't really cre
05:10
<andersca>
Hixie: OK
07:18
Hixie
commits <iframe seamless>
07:21
<hober>
Hixie: formatting error
07:21
<hober>
I see "he seamless/code> attr..."
07:21
<Hixie>
yeah i'm on it
07:25
<Hixie>
ok fixed those
07:25
<Hixie>
that was a mess :-)
07:25
<Hixie>
ok let's see
07:26
<Hixie>
i guess sandbox="" is next
07:38
<MikeSmith>
Hixie: small but important change:
07:38
<MikeSmith>
"The reversed attribute is a boolean attribute. If present, it indicates that the list is an ascending list. If the attribute is present, the list is a descending list."
07:38
<MikeSmith>
2nd sentence should start "If absent,"
07:38
<MikeSmith>
can send you mail about it if you want
07:40
<Hixie>
actually i'm between edits right now so i can change it
07:40
<Hixie>
but in general e-mail is definitely better
07:40
<Lachy>
Hixie, when the src of a seamless iframe is changed to a document that is not the same origin (either via the dom or <a href=... target=the-iframe>), is that expected to change from a seamless frame to a normal iframe?
07:53
<jwalden_>
http://pastebin.mozilla.org/436577
07:53
<jwalden_>
I'm not getting a message saying it'll appear after moderation or anything like that
07:53
<jwalden_>
so I don't know what's up with it
07:54
<jwalden_>
just redirects back to the main blog page
07:56
Hixie
tries to comment too
07:57
<Hixie>
wfm
07:58
<Hixie>
(i used safari 3)
07:58
<Hixie>
(trunk
07:58
<Hixie>
)
07:59
<Hixie>
one day i'll stop working on browsers and then i'll have to pick a preferred browser that i use everywhere, instead of switching browsers every 2 minutes...
08:00
<gavin_>
when you have to do something on the web that's not testing browsers, do you really just randomly pick one?
08:01
<gavin_>
you don't care about any of the stuff that browsers keep track of for you?
08:04
<Hixie>
gavin_: i use multiple machines and i nuke my profiles regularly, so no
08:08
<Hixie>
jwalden: wfm using safari trunk; might be a browser problem.
08:08
<Hixie>
jwalden: good comment btw
08:08
<jwalden>
Firefox 2? that would be pernicious
08:08
<Hixie>
oh i assumed you were using ff trunk
08:08
jwalden
tries Safari 3.latest
08:08
<Hixie>
afk, brb
08:08
<jwalden>
I'm lame
08:13
<jwalden>
would have been nice for MSFT to point out their implementation isn't up to latest spec, but I'm guessing that's just their paranoia about saying what they're doing before they do it :-\
08:14
<jwalden>
or rather, to say that anyone who experiments should expect future breakage
08:16
<Hixie>
jwalden: when you're in an abusive relationship, and your spouse has been beating you for 15 years, flinching when they raise their hand to hug you isn't paranoia
08:16
<Hixie>
it's good sense
08:17
<MikeSmith>
jwalden: "Our policy gives our customers more predictability in their planning."
08:17
<MikeSmith>
http://blogs.msdn.com/ie/archive/2008/05/13/may-chat-with-the-ie-team-on-thursday.aspx#8502193
08:18
<Hixie>
MikeSmith: that answer is especially stupid given microsoft's policy of announcing everything years before they have implemented it
08:18
<MikeSmith>
predictability meaning, I guess, that their customers can consistently plan on them not giving any information about anything until they ship it
08:18
<MikeSmith>
yeah
08:19
<Hixie>
(my favourite example being winfs, which iirc was announced around when Windows 95 was in beta)
08:19
<Hixie>
(13 or 14 years ago)
08:20
<MikeSmith>
announce ahead when it's convenient and useful (even vaporware), don't announce anything else ahead unless it gives some potential market advantage to do so
08:20
<jwalden>
I think there's a double standard here, tho, of announcing something (so committing to it being there) before it's ready and not announcing it until it's complete
08:20
<roc>
Hixie, I think you're mixing up WinFS with the "Cairo" "OLE file system"
08:20
<roc>
I think they cancelled the project at least once in between :-)
08:21
<Hixie>
roc: right
08:21
<Hixie>
roc: i just mean their "next generation" file system to come after NTFS which is somehow "revolutionary"
08:22
<Hixie>
the actual project has, as you say, been announced and canceled and reannounced many times since the original announcement
08:22
<MikeSmith>
I think in part that policy is just a general kneejerk policy of unnecessary secrecy -- of keeping everything in development (across all their product lines) secret by default
08:22
<Hixie>
i really do think the abusive relationship analogy is appropriate here, btw, right down to people being willing to forgive them any time they claim to have turned over a new leaf
08:22
<MikeSmith>
it reminds me of the Bush administration secrecy policy
08:23
<MikeSmith>
which has resulted in more stuff being made secret in the last eight years than in the previous 40 or something years combined
08:23
<Hixie>
(and right down to the feeling i have of watching the relationship -- i've seen abusive relationships and it's pretty much exactly the same feeling)
08:23
<Hixie>
MikeSmith: but the point is they DON'T keep everything secret!
08:24
<MikeSmith>
yeah, same as Bush admin
08:24
<MikeSmith>
they leak stuff when they want to
08:24
<roc>
Silverlight is the best current example of this
08:24
<MikeSmith>
like the "Valerie Plame is a CIA agent" thing
08:24
<jwalden>
everybody does that, a way of saying something without saying it
08:24
<roc>
its feature set has been announced for years and it hasn't even shipped yet
08:24
<jwalden>
it's kinda smarmy, but it's common practice
08:25
<Hixie>
i thought silverlight 2 had shipped
08:25
<Hixie>
am i falling prey to the fud too?
08:25
<roc>
don't think so
08:25
<Hixie>
so what is it that microsoft.com asks me to install every time i go there?
08:25
<Hixie>
oh sl2 is in beta
08:25
<Hixie>
ok
08:25
<MikeSmith>
while I can see a policy like that making sense for their other products -- or for other classes of products in general -- I think it makes a lot less sense for IE, or for a browser product in general
08:26
<MikeSmith>
because for one thing they're not selling it, nor is anybody else selling a desktop browser
08:37
<MikeSmith>
Hixie: btw, you got me in the Acks twice, as "Michael Smith" and "Mike Smith".. not sure I merit being in there at all (since the stuff I've commented on has mostly just been minor editorial things) but if you do have me in there, "Michael(tm) Smith" would be preferred (to make it clear which Michael Smith among the nation of Michael Smiths it is)
08:39
<Hixie>
(tm) or &trade; ? :-)
08:40
<Lachy>
MikeSmith, what's the purpose of the (tm) in your name?
08:41
<Hixie>
he just said what the purpose was
08:41
<Hixie>
"to make it clear which Michael Smith among the nation of Michael Smiths it is
08:41
<Hixie>
"
08:42
<Lachy>
ok, but why use TM? Is it your middle initials or something?
08:42
<Hixie>
"trademark"
08:42
Hixie
upgrades Lachy's sense of humour to the latest security update :-P
08:43
<Lachy>
yeah, I get that it means trademark. But I don't get the relevance
08:44
<Hixie>
it's irony
08:44
<Hixie>
but i don't know how to explain it
08:45
<Hixie>
like, "I'm going to go and eat now, because I'm Hungry(tm)"
08:45
<MikeSmith>
Lachy: what Hixie said :)
08:47
<Lachy>
ok, so I guess there's no logical or rational explanation for it and I just need to accept it :-(
08:49
<Hixie>
it's irony because his name is so common, yet he claims it as a trademark
08:49
<Hixie>
it's perfectly rational, but i don't know how to explain it :-)
08:49
<Hixie>
MikeSmith: spec is updated btw
08:50
<Lachy>
ok, that's better explanation.
08:53
<MikeSmith>
Hixie: cheers
08:53
<MikeSmith>
saw your checkins
08:56
Lachy
wonders why he didn't receive twitter updates for the recent spec checkins
08:56
<Hixie>
did twitter get them?
08:56
<Lachy>
yeah, the iframe seamless one is there, but I didn't receive it
08:58
<Lachy>
ah, it's twitters crappy jabber implementation that stops sending me twitter updates sometimes
09:01
<Hixie>
i haven't been able to get the jabber to work, like, ever
09:01
<Lachy>
what do you use to receive updates?
09:02
Philip`
would have assumed Hixie already knew about updates to the spec without having to get Twitter notifications :-p
09:02
<Hixie>
Lachy: i don't follow twitter
09:03
<Hixie>
i only use it to get direct twitters when i'm expecting them (and then i just look at the site, half the time i don't get e-mail notifications)
09:04
<Hixie>
other than that, for me, it's a write-only medium
09:09
<Hixie>
http://www.whatwg.org/specs/web-apps/current-work/#xxxoriginpleasesuggestabettername
09:09
<Hixie>
just for the record
09:09
<Hixie>
that is not a proposal for the name
09:09
<Hixie>
and if anyone implements that keyword i will personally hunt them down
09:24
<Lachy>
Hixie, add sandbox to "The DOM attributes src, name, and seamless must reflect the content attributes of the same name. "
09:26
<Hixie>
will do
09:26
<Hixie>
haven't got that far yet
09:26
<Lachy>
ok.
09:26
<Lachy>
will there be a DOMTokenList attribute for the sandbox?
09:28
<roc>
it would be useful if the spec document was published in multiple versions with different stability
09:29
<Hixie>
roc: hm, interesting idea
09:29
<Hixie>
roc: how would that work?
09:29
<roc>
not sure
09:29
<jwalden>
html5-mm
09:29
<roc>
but right now, for authors and implementors its hard to tell what is set in stone and what is likely to change
09:29
<Lachy>
roc, why would that be useful?
09:29
<Lachy>
oh
09:30
<Hixie>
roc: well, the sections do have annotations
09:30
<Hixie>
roc: which should make it relatively easy
09:30
<roc>
and what is likely to get them hunted down :-)
09:30
<Hixie>
roc: but the presentation could do with some work, true
09:40
<roc>
not all the sections have annotations
09:41
<roc>
and it's hard to see what is covered by each annotation, especially if you happen to jump to the middle of the document and there's no annotation visible. I guess that's what you mean by the presentation needing work.
09:45
<Hixie>
roc: yeah
09:45
<Hixie>
roc: i'm slowly adding annotations
09:45
<Hixie>
roc: any help in either that or making the presentation better would be welcome
09:49
<jwalden>
Hixie: what's the "tests: ##" thing supposed to mean in the annotation boxes? some whatwg/w3-maintained test suite or something?
09:49
<Hixie>
links (one URI per line) to any tests you know of
09:50
<jwalden>
ah
09:52
<jwalden>
I suppose that doesn't work so well for cross-domain tests
09:53
<MikeSmith>
Hixie: the XXX convention in the spec is something you consistently use to indicates a placeholder ("we're not done here yet") thing?
09:53
<Hixie>
jwalden: ?
09:53
<Hixie>
MikeSmith: yes
09:53
<Hixie>
the blue line on the chart shows the count of XXX markers
09:53
<MikeSmith>
aha
09:54
<MikeSmith>
wondered what that blue line was
09:54
MikeSmith
is remined of Errol Morris film "The Thin Blue Line"
09:54
<jwalden>
er, yeah, I suppose you could set up tests on a bunch of real servers
09:54
<jwalden>
would be a pain, but you could do it, I guess
09:56
<Hixie>
not that much of a pain
09:56
<Hixie>
i do it for a bunch of tests on hixie.ch
09:56
<Hixie>
they use another.domain.libpr0n.com as the other domain
09:56
<Hixie>
just one symlink
09:57
<Hixie>
for the record, browsers all suck
09:57
<Hixie>
i just wasted like 30 minutes trying to debug something in safari using the safari web developer tools
09:57
<Hixie>
which frankly are crappy compared to firebug and other things available for firefox
09:57
<Hixie>
but god, safari loads the html5 spec so much faster than firefox
10:01
<Hixie>
(safari also reloads way more than necessary when all i do is change a fragment identifier, and doesn't reload the tabs when it crashes)
10:01
<Hixie>
(and occasionally it seems to get wedged in a state where it won't download any pages)
10:02
<Hixie>
but it feels much lighter than firefox
10:02
<Hixie>
at least on mac
10:54
<Hixie>
bbl
10:58
<MikeSmith>
I'm trying to figure out what the main use cases are the the Opera File I/O proposal would enable
10:59
<MikeSmith>
I mean what specific feature it would bring to users that existing applications can't
10:59
<MikeSmith>
features
16:45
<annevk>
Hixie, "an descending"
17:13
<annevk>
Hixie, seamless is not a boolean on the IDL
18:05
<hendry>
i need to draw a graph. what to library to use. seems to be so many nowadays.
19:07
<Philip`>
hendry: Gnuplot!
19:08
Philip`
has almost figured out how to use that, so he doesn't want to waste the investment by switching to anything else
19:16
<Lachy>
hendry, it depends what you need to draw the graph for and where you intend to use it.
19:16
MikeSmith
reads jwalden comment on IE Blog posting about cross-document messaging
19:16
<MikeSmith>
http://blogs.msdn.com/ie/archive/2008/05/19/enabling-mash-ups-in-internet-explorer-8-with-cross-document-messaging.aspx#8520209
19:23
<MikeSmith>
by way of summary at Ajaxian:
19:23
<MikeSmith>
http://ajaxian.com/archives/ie-8-and-cross-document-messaging
19:24
MikeSmith
wonders how Dion Almaer makes time to code and still blog & tweet as much as he does
19:50
<hendry>
Lachy: on the Web with Firefox 3. need to plot a simple approximate 2d graph to demonstrate an idea. I guess inkscape is probably best. I love xfig but it doesn't do SVG I guess.
19:52
<Lachy>
hendry, yeah, you could use SVG or <canvas>.
19:52
<Lachy>
If you want it generated dynamcially from a table of data, you could use PlotKit
19:52
<Philip`>
Or PNG
19:53
<Philip`>
Avoid overcomplication :-)
20:15
<hsivonen>
http://joeclark.org/access/webaccess/WCAG/response1_Understanding-WCAG2.html interesting points
20:30
<Dashiva>
hsivonen: There's also the whole "wcag process isn't accessible" thing that nobody wants to talk about :)
20:31
<Philip`>
"If data were understandable by themselves, we wouldn’t make a chart." seems untrue - we'd make a chart if the chart was more understandable than the data, even if the data was already understandable
20:32
<Lachy>
"A slider? That requires a mouse, not to mention full vision. " - How does a slider require a mouse? It can be used by keyboard users if implemented correctly
20:32
<hsivonen>
Dashiva: I don't have a diagnosed learning disability, but I find WCAG2 and ATAG2 hard to follow, because they are so abstract and then lawyerly on top of that
20:32
<Lachy>
maybe that was true when the document was written, and there was no way to add a slider to a web page accessibly
20:33
<hsivonen>
Lachy: it's from 2006
20:33
<Lachy>
ok
20:39
<Dashiva>
I thing I've been wondering is if (e.g.) red-green colorblind people are more likely to think "red looks like green", "green looks like red", "red and green are different, but I can't tell them apart" or "red and green are two names for the same color"
21:02
MikeSmith
notes Updated: 2006.05.23 at the bottom of that article
21:33
<Lachy>
http://css-tricks.com/html-5-vs-xhtml-2-an-article-roundup-and-poll/
21:35
<Lachy>
also note the poll at the bottom of the right column. XHTML2 (33%, 33 votes) is ahead of HTML5 (18%, 17 votes)
21:40
<Philip`>
Lachy: The people have spoken - HTML5 is doomed :-(
21:42
<Lachy>
indeed. I'm going to quit now and join the XHTML2 group before it's too late.
21:52
<Dashiva>
Darn, I wish someone told me the future of the web was decided on a first-come, first-vote, 50-votes-only basis
21:53
<hsivonen>
Dashiva: cf. The Workshop :-)
22:39
<Lachy>
Dashiva, there were more votes. But most people voted don't care.