00:22
<TabAtkins>
othermaciej: Sorry for the curtness on the lists. >_< Really didn't intend to be rude; I just shouldn't have responded at all. It was an obviously provocative statement that didn't deserve a response.
00:25
<othermaciej>
TabAtkins: just do your best to be courteous, please, even if you feel provoked for whatever reason
00:26
<TabAtkins>
Yeah, I know. I'm sorry I let myself go. I'll watch myself more in the future.e
00:29
<mpilgrim>
png optimization baffles me
00:30
<TabAtkins>
That it works?
00:35
<mpilgrim>
that i've optimized the same images with about six different tools, and each one seems to find a way to make each image slightly smaller than that last round
00:35
<TabAtkins>
You've discovered a perfect compressor.
00:35
<TabAtkins>
Keep going until the file is just 1 bit.
00:36
<mpilgrim>
apparently there's an element of randomness to it as well
00:37
<mpilgrim>
there are far too many ways to express the same image to try them all
00:37
<mpilgrim>
so it's a sort of monte carlo method
00:38
<mpilgrim>
which is quite maddening for those of us who are obsessive about making our sites load more quickly
00:38
<mpilgrim>
because, you know, you could always run it again, and maybe it'll be smaller
00:38
<mpilgrim>
you can't ever be sure that you're done
00:39
<Rik`_>
mpilgrim: do you know about http://imageoptim.pornel.net/ ?
00:39
<Rik`_>
it will take try several tools and pick the best
01:32
<mpilgrim>
Rik`: oh goodie, another tool to try
01:32
<mpilgrim>
you're not helping my neuroses
01:33
<Rik`>
mpilgrim: hopefully, this will be the last one
01:46
<erlehmann>
without dialog, what sensible markup options for an interview are there ?
02:09
<Hixie>
it has been brought to my attention that i was incorrect yesterday in saying RDFa was the work of the XHTML2 WG
02:09
<Hixie>
it was actually a joint project of the XHTML2 WG and the Semantic Web Deployment Working Group
02:09
<Hixie>
(I'd never heard of the Semantic Web Deployment Working Group before)
02:28
<TabAtkins>
erlehmann: There are many sensible options. The spec's suggest <p> and <b> method would work just fine. Ideally I might mark up the speaker with a <hx> and give it display:run-in.
09:54
<Lachy>
I don't get how Shelley can complain about the size of the spec being too large, and yet get all worked up about people suggesting that some features will end up being deferred to a future revision of HTML.
10:15
<Hixie>
it's probably similar to how she thinks that the "most compelling" argument for something she agrees with is my opinion, but that if she disagrees with something, that my opinion is worthless
10:52
<Hixie>
sorry, "a compelling", not necessarily the "most compelling"
10:53
<jgraham>
I don't get why people are so happy to let her dictate the terms of all discussion
11:00
<roc>
Hixie: isn't that common? If my enemy agrees with me *even though* he is evil, that is clearly strong evidence of the universal appeal of my position. On other hand, if my enemy disagrees with me, then that is only because he is evil.
11:01
<Hixie>
roc: I didn't say it was uncommon, but that doesn't make it acceptable
15:50
<mpilgrim>
i wonder if julian realizes he just walked into a trap
15:50
<mpilgrim>
probably not, since that's why they call them "traps"
15:51
<mpilgrim>
IMG isn't controversial *now*, but it was most certainly controversial when it was introduced (by a browser vendor, against the working group's wishes)
15:52
<mpilgrim>
it also wasn't in the first several drafts of XHTML 2, probably because XHTML 2 was TBL's recreation of what he thought HTML always should have been
15:52
<annevk2>
back back
15:52
<annevk2>
discussions on public-html are still somewhat painful to read
15:53
<mpilgrim>
references: http://lists.w3.org/Archives/Public/public-html/2009Dec/0173.html and http://lists.w3.org/Archives/Public/public-html/2009Dec/0174.html
15:54
<mpilgrim>
and TBL never wanted an IMG element
15:54
<annevk2>
I'm guessing some of the participants are not very happy with life
15:56
mpilgrim
is very happy with life, but finds that reading public-html distracts him from getting real work done
15:56
<mpilgrim>
like writing a book
15:56
<mpilgrim>
unlike irc, which isn't distracting at all :)
15:57
<annevk2>
IRC is nice
15:58
<annevk2>
I've never really seen significant process made on public-html discussions involving "two camps" with mutually incompatible goals
15:58
<annevk2>
I stopped participating in them, mostly
16:00
<annevk2>
The other reason is that most discussions feel like damage control, which I also tried to do in the CDF WG, and was no fun at all and mostly pointless in the end anyway
16:01
<annevk2>
Maybe we should just split the group and have two editions of HTML5. After a decade we can study the then contemporary practices and go to REC
16:02
<mpilgrim>
i'm pretty sure that's inevitable anyway
16:02
<mpilgrim>
maybe shelley et. al. could call their version "HTML5 Plus"
16:03
<annevk2>
sounds fine to me
16:03
<annevk2>
hopefully then everyone can be somewhat more happy instead of all this hostile crap
16:04
<mpilgrim>
i miss XHTML 2
16:04
<mpilgrim>
it provided a nice safe playground where all these like-minded people could play without hurting anyone
16:04
<timz>
mpilgrim: they'd better call it HTML5less
16:05
mpilgrim
wonders if timz knows the history of "HTML Plus"
16:05
<timz>
hehe no :-)
16:05
<timz>
i'm having a look
16:06
<daedb>
this http://www.w3.org/MarkUp/HTMLPlus/htmlplus_1.html ?
16:08
<timz>
haha, yes that would probably fit her/their ideas of what HTML5 should be
16:09
<mpilgrim>
the W3C has a long history of fucking up HTML by going off into la-la land for a few years while the world evolves without them
16:09
daedb
has never heard of HTML Plus before...
16:09
<mpilgrim>
HTML+ was one such effort (between HTML 1.0 and HTML 2.0)
16:09
<mpilgrim>
HTML 2.0 was basically "hey, let's see what browsers have actually implemented while we weren't looking"
16:09
<mpilgrim>
then they made HTML 3.0, which was similarly delusional
16:10
<mpilgrim>
then they threw that away and made HTML 3.2, which was basically "hey, let's see what else browsers have actually implmeented while we were busy masturbating"
16:10
<timz>
haha
16:11
<mpilgrim>
then they made HTML 4.0, which shockingly wasn't completely horrible (though it was horribly underspecified)
16:11
<timz>
you've been around all of those discussions or you just red the "history of html, a different perspective"
16:11
<mpilgrim>
then they gave up on HTML altogether, to the point of ignoring all W3C process and refusing to issue errata or integrate existing errata into point releases
16:12
<mpilgrim>
and started whole hog on XHTML
16:12
<mpilgrim>
XHTML 1.0 is literally HTML 4 + a few useless slashes
16:13
<mpilgrim>
oh, and a new MIME type and draconian error handling, which nobody uses (hi Sam)
16:13
<mpilgrim>
then XHTML 1.1, which nobody uses
16:13
<timz>
so is HTML5 a w3c idea or a community (vendors) idea ?
16:13
<mpilgrim>
then XHTML 2, which they spent almost a decade on and finally put out to pasture this year
16:14
<mpilgrim>
HTML5 was a community-led effort of browser vendors, standards wonks (like Hixie), and other interested parties
16:14
<mpilgrim>
all formerly quite active within the W3C
16:14
<mpilgrim>
they begged and pleaded with the W3C to let them work on HTML5 (or 4.1, or whatever, anything but XHTML 2)
16:14
<mpilgrim>
the W3C said no
16:15
<mpilgrim>
so they formed their own working group (WHATWG)
16:15
<mpilgrim>
and started from scratch (since the license of the HTML 4 specification forbids derivative works by non-W3C organizations)
16:16
<Lachy>
so then it seems this HTML+RDFa effort may just be the next round of crap to be attempted during the overall development of HTML, that will likely eventually die out like the rest.
16:16
<mpilgrim>
HTML+RDFa doesn't appear to have any chance at all of being a good spec until they redefine RDFa in terms of the DOM
16:16
<mpilgrim>
anything else is simply insane
16:17
<mpilgrim>
that's not to say people won't use it
16:17
<mpilgrim>
google rich snippets, etc.
16:17
<timz>
with limited vocabs that is
16:17
<mpilgrim>
but defining a new syntax "on top of" HTML in terms of a byte stream is just insane
16:19
<daedb>
heh, HTML+ had a <person> element :)
16:20
<mpilgrim>
http://diveintomark.org/archives/2009/11/02/why-do-we-have-an-img-element is fun related reading on the history of HTML
16:20
<Dashiva>
RDFa being based on the source text kinda makes me think of people parsing HTML with regex
16:20
<mpilgrim>
it (and the rest of this stuff) will form the basis of the introductory chapter of http://diveintohtml5.org/
16:20
<Dashiva>
It seems to work until you start using it for anything significant
16:21
<mpilgrim>
indeed
16:21
<daedb>
<RENDER TAG="PROPNAME" STYLE="I">... that's... interesting :O
16:22
<timz>
mpilgrim: "but defining a new syntax "on top of" HTML", you are referring to RDFa ?
16:22
<mpilgrim>
yes
16:23
<mpilgrim>
it would appear that the current HTML+RDFa effort has zero chance of producing a good specification
16:23
<mpilgrim>
and if RDFa sees enough implementation, we'll need to do an RDFa5 retro-spec in a few years
16:23
<timz>
because it so explicitly bound to xml and rdf ?
16:24
<mpilgrim>
which will consist of reverse engineering google's implementation (or whoever ends up being the dominant consumer of RDFa on the public web)
16:24
<timz>
hmm, sounds like a good argument for keeping microdata
16:24
<Dashiva>
Hmm, yeah
16:24
<mpilgrim>
i'm not a fan of microdata either
16:25
<mpilgrim>
it's a decent enough solution to the stated problem
16:25
<Lachy>
mpilgrim, that's not surprising. I read the HTML+RDFa spec, and it's clearly been written by people with little clue about how to actually write specifications
16:25
<mpilgrim>
i'm just not a fan of the stated problem
16:25
<timz>
hehehe
16:25
<Dashiva>
I'm sure there will be a decent amount of complaints that Google is wilfully sabotaging RDFa by having a non-perfect implementation
16:25
<mpilgrim>
when people say "i want to extend HTML in arbitrary, private, and probably proprietary ways," the appropriate answer is "fuck you"
16:27
<Dashiva>
Extensibility has become one of those words that doesn't mean anything anymore
16:28
<Dashiva>
Like when the topic was making sure HTML5 was easily extensible by HTML6, it turned into a complaint about lack of DE
16:28
<Lachy>
mpilgrim, if their intention is for those extensions to appear on the public web and have 3rd party implmentations and websites using them, then that sure is the right answer
16:29
<Lachy>
Dashiva, even the concept of distributed extensibility doesn't mean much in practice. It's basically become an alias for full namespace support and RDFa
16:30
<Lachy>
and that's despite all the extension points already in HTML5 that allow 3rd party extensions, without using either namespaces or RDFa
16:31
<mpilgrim>
Lachy: haven't you heard? those aren't "real" extensibility points
16:31
<mpilgrim>
Paging True Scotsman, party of 1... True Scotsman, party of 1
16:31
<Lachy>
the problem seems to be that proponents believe those to be the ultimate solution and are unwilling to accept any alternative solution, regardless of any evidence presented, nor evaluation of them against the use cases and problems they solve
16:36
<timz>
http://diveintomark.org/archives/2009/11/02/why-do-we-have-an-img-element <-- good read mpilgrim :-)
16:37
<mpilgrim>
'twas fun to research and write, too
16:39
<Dashiva>
http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2009Dec/0022.html
16:39
<Dashiva>
Validation is overrated, I guess
16:42
<annevk2>
Removing the test seems like a very poor solution to the problem...
16:43
<daedb>
http://www.w3.org/MarkUp/HTMLPlus/htmlplus_21.html Wow, HTML+ has both <img> and <image>... did the alt attribute not exist when that was written?
16:44
<annevk2>
alt was invented for lynx at some point, not sure when
16:44
<mpilgrim>
the alt attribute was first added in HTML 4.0
16:45
<mpilgrim>
lol
16:45
<mpilgrim>
i'd forgotten this salient little detail:
16:45
<mpilgrim>
when the XHTML 2 WG did finally re-introduce the <img> element,
16:45
<mpilgrim>
they gave it a different content type than <img> in XHTML 1 and HTML 4
16:46
<mpilgrim>
they made it a non-void element
16:46
<mpilgrim>
the child content was the fallback content
16:47
<mpilgrim>
which is actually a great idea, but (i'm told) it royally breaks existing web content
16:47
<annevk2>
yeah, it seems they admitted compat was needed, but then at the same time did not
16:47
<annevk2>
or something
16:47
<mpilgrim>
"Also, <img> was reintroduced to XHTML 2, no doubt after a vigorous and healthy debate in which all parties treated each other with mutual respect. But now it has a different content model than <img> in XHTML 1 and HTML 4, which just goes to show that mutual respect is for chumps."
16:47
<mpilgrim>
http://microformats.org/discuss/mail/microformats-discuss/2005-October/001783.html
16:56
<daedb>
Cool, HTML+ had an element for figures :)
17:27
<Lachy>
daedb, so did HTML3
17:40
<daedb>
Lachy: Hopefully it will stay this time.
17:41
<Lachy>
daedb, it might get dropped again, since we still can't resolve the caption element issue, and the current dt/dd model sucks
17:53
<timz>
what's the primary argument for not using caption ?
17:53
<timz>
in figure
17:54
<Lachy>
timz, compatibility with legacy parsing issues
17:55
<timz>
okay, so html5 documents have to be compatible with older parsers ?
17:56
<Dashiva>
No, html5 parsers have to be compatible with older documents
17:56
<Dashiva>
But the reverse is of course desirable
18:07
<daedb>
Lachy: I really hope not, it's one of my favourite new elements.
21:32
<zcorpan_>
http://www.w3.org/XML/2009/12/xml-stylesheet/
22:40
<AryehGregor>
jgraham, given the choice between using RDFa and using nothing, many people would use nothing. Except when RDFa is the only thing that works with CC license readers, or Rich Snippets, or . . .
22:41
<AryehGregor>
I'm not worried about individual authors saying "Let's use RDFa, that sounds great!" I'm worried about organizations with some clout basing new de facto standards on it.
22:41
<AryehGregor>
Thus forcing everyone to use it.
22:41
<AryehGregor>
That's how it's worked its way into MediaWiki.
22:42
<AryehGregor>
I couldn't really argue with "But you have CC license-scraping bots and this would be great for Commons."
22:42
<AryehGregor>
(Wikimedia Commons, that is)
22:42
<AryehGregor>
RDFa is the only Creative Commons-approved way to embed license metadata, AFAIK.
22:45
<hsivonen>
AryehGregor: FWIW, I think committees don't get to vote to kill specs. they only get to vote on whether they allocate their resources to developing a given technology
22:46
<hsivonen>
for example, OASIS doesn't get to vote OOXML out of existence
22:46
<AryehGregor>
Well, the W3C could still endorse one or the other exclusively.
22:46
<hsivonen>
and the W3C didn't successfully vote HTML out of existence
22:46
<hsivonen>
AryehGregor: sure
22:46
<AryehGregor>
So could the HTMLWG.
22:46
<hsivonen>
yes
22:46
<AryehGregor>
That would have an impact.
22:47
<AryehGregor>
Just like disbanding the XHTMLWG had an impact. We get to say "No, don't use that, it's obsolescent."
22:47
<AryehGregor>
I doubt it will happen, though.
22:47
AryehGregor
shakes fist dramatically at the W3C
22:47
<hsivonen>
it had a publicity impact
22:47
<hsivonen>
but XHTML2 was already dead
22:47
<AryehGregor>
Yes, that's my point.
22:47
<hsivonen>
or s still very much alive depending on point of view
22:48
<AryehGregor>
If the W3C said "Hey, we decided RDFa is a bad idea, use microdata instead", then we could say to Creative Commons "Hey, why don't you encourage people to use microdata instead of RDFa? The W3C says so!"
22:48
<hsivonen>
AryehGregor: I observe that the W3C didn't stop work on XForms when it invited HTML5 (incl. Web Forms 2.0) in
22:49
<AryehGregor>
And I think that's silly too. :)
22:49
<AryehGregor>
(plus, I'm sure that will die soon enough, unless it has significant non-HTML users)
22:51
<hsivonen>
XForms got baked into ODF 1.0
22:51
<hsivonen>
some approximation of RDFa got baked into ODF 1.1, according to rumors
22:52
<AryehGregor>
Yay.
22:53
<AryehGregor>
Oh well, none of this can possibly be more horrible than all the legacy HTML currently has to deal with.
22:53
jgraham
wouldn't put money on that
22:54
<jgraham>
(I am about to go sleep otherwise I might try to say something more useful)
22:57
<AryehGregor>
It's okay, it's unlikely there's anything useful to say in this discussion, except "Leave them alone for now and let God sort it out."
23:05
<Philip`>
Seems unfair to leave our mess for Him to clean up
23:07
<AryehGregor>
Well, it's really His mess, since He's the owner of the entire universe, after all.
23:07
<AryehGregor>
So everything is His.
23:07
<AryehGregor>
Also, it's His fault for telling us we can't murder people for promoting bad standards.
23:11
<Philip`>
That seems a little extremist
23:11
<AryehGregor>
Yes, I agree, it's just as likely they'd have murdered us, so we can't really blame Him on that score.
23:12
<AryehGregor>
We can blame the people responsible for the other standards, though. :)
23:12
<Philip`>
I don't think people who promote what you consider bad standards are that extremist either :-p
23:19
<zcorpan_>
"Sure; but if I've scheduled for it, I can deal with a hundred comments in
23:19
<zcorpan_>
a day -- that's not a problem." -- http://www.w3.org/mid/Pine.LNX.4.62.0912050854010.5629⊙hdc
23:22
<zcorpan_>
re rjavascript, i guess it would be useful with a linter for js performance pitfalls in modern implementations
23:22
<AryehGregor>
Haha, Google Site Performance is telling me my page is performing poorly because I don't have gzip compression enabled for http://www.google-analytics.com/urchin.js.
23:22
<AryehGregor>
\o/
23:22
<jgraham>
Die with die
23:23
jgraham
obviously got distracted from going to bed
23:23
<zcorpan_>
or maybe browsers should have a way to log in the error console when the js engine has to do something expensive
23:23
jgraham
was talking about the "with" statement in case it wasn't obvious
23:24
<jgraham>
zcorpan_: It seems like something Dragonfly et. al. could tell you (depending on how much attaching a debugger affects the actual code)
23:25
<jgraham>
(But it would be nice if they could do per-call profiling with big warnings for things that hit slow paths)
23:26
<Philip`>
"slow paths" include falling off the JIT, which might be really quite difficult to explain to an average script developer
23:26
<Dashiva>
I'm a bit rusty on the edges, but isn't there some overlap between rjavascript and stuff like caja?
23:27
<zcorpan_>
Philip`: no need to explain why it's slow, just state that it's slow and suggest fast alternatives
23:27
<Philip`>
(There's a thing that gives JIT profiling information for code in SpiderMonkey but it's not the most user-friendly thing imaginable, by necessity)
23:28
<jgraham>
Philip`: That was basically what I had in mind hen I said "slow paths". I'm not sure why it would be hard to explain to people sufficiently motivated to profile their code
23:30
jgraham
really is going to bed now
23:32
<Philip`>
jgraham: It seems hard because they'd first need to understand what a tracing JIT is and how it works
23:33
<Philip`>
http://blog.mozilla.com/dmandelin/2009/02/26/tracevis-performance-visualization-for-tracemonkey/
23:37
<Philip`>
It's easy to suggest "don't use with" and "don't use eval", but I don't really see how you could automatically give suggestions for the more complex problems that occur
23:40
<Philip`>
(It would still be nice, just maybe not possible)
23:56
AryehGregor
hasn't figured out what "with" actually does yet, but gets the impression that it's cool and trendy and is vaguely disappointed if he shouldn't use it in JS for performance reasons
23:56
<Dashiva>
It's horrible and evil
23:56
<AryehGregor>
But also cool and trendy?
23:56
<Dashiva>
No
23:57
<Dashiva>
It was cool and trendy a few years ago, when having the shortest bookmarklets possible was a goal
23:57
<Dashiva>
And it sometimes appears in golfing, I suppose?