Options:
- # Session Start: Tue Mar 17 00:00:00 2009
- # Session Ident: #html-wg
- # [00:03] * Quits: Julian (chatzilla@217.91.35.233) (Ping timeout)
- # [00:06] * Joins: Julian (chatzilla@217.91.35.233)
- # [00:38] <takkaria> oh, someone embracing the impermanance of all things, even URIs
- # [00:38] <takkaria> oops
- # [01:01] * Quits: jwatt (roslea@83.87.4.17) (Client exited)
- # [01:49] * Quits: Lachy (Lachlan@85.196.122.246) (Quit: Leaving)
- # [02:09] * Joins: Lachy (Lachlan@85.196.122.246)
- # [02:26] * Quits: tH (Rob@129.11.83.14) (Quit: ChatZilla 0.9.84-rdmsoft [XULRunner 1.9.0.1/2008072406])
- # [02:32] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [02:35] * Quits: adele (adele@17.203.15.219) (Ping timeout)
- # [03:13] * Joins: rking3 (rking3@99.27.42.97)
- # [03:51] * Joins: karl (karlcow@128.30.54.58)
- # [03:56] * Quits: rking3 (rking3@99.27.42.97) (Quit: rking3)
- # [04:43] * Joins: xover (xover@193.157.66.22)
- # [05:23] * Joins: heycam (cam@130.194.73.110)
- # [05:31] * Quits: heycam (cam@130.194.73.110) (Ping timeout)
- # [06:07] * Joins: adele (adele@24.7.123.50)
- # [06:34] <MikeSmith> http://www.squarefree.com/2009/03/16/css-grammar-fuzzer/
- # [06:34] <pimpbot> Title: Jesse Ruderman » CSS grammar fuzzer (at www.squarefree.com)
- # [06:35] <MikeSmith> "Like any good fuzzer, it doesn't always follow the given context-free grammar. Sometimes it does weird stuff, such as inserting a random symbol, to throw the parser off. I was surprised that this only found one additional bug in Gecko. Perhaps this reflects the comprehensive error handling requirements of the CSS specifications and the corresponding test suites."
- # [06:41] * Joins: heycam (cam@130.194.73.110)
- # [06:50] * Quits: heycam (cam@130.194.73.110) (Quit: bye)
- # [06:50] * Joins: heycam (cam@130.194.73.110)
- # [06:55] * Quits: Dashiva (noone@80.202.223.46) (Quit: Dashiva)
- # [06:57] * Joins: Dashiva (noone@80.202.223.46)
- # [07:06] * Quits: heycam (cam@130.194.73.110) (Quit: bye)
- # [07:50] * Quits: gavin_ (gavin@99.226.207.11) (Ping timeout)
- # [07:55] * Joins: gavin_ (gavin@99.226.207.11)
- # [08:10] * Joins: heycam (cam@124.168.80.126)
- # [08:55] * Joins: tH (Rob@129.11.83.14)
- # [08:57] <MikeSmith> do we really want to people think it's OK to post messages to public-html and cross-post them to multiple other mailing lists at the same time?
- # [09:04] <hsivonen> depends on what other lists, I suppose :-/
- # [09:04] <hsivonen> it seems that the www-svg CC is relevant
- # [09:04] <shepazu> sorry for all the cross-posting :(
- # [09:05] <hsivonen> shepazu: I'm about to reply to you. should I drop the www-svg CC?
- # [09:06] <shepazu> hsivonen: it's not bothering me... I'd prefer the SVG community were in on this conversation, myself
- # [09:06] <hsivonen> shepazu: ok. I'll keep the CC.
- # [09:06] <shepazu> cool
- # [09:07] <shepazu> there may come a point where we need to break it off
- # [09:07] <hsivonen> personally, I don't like the <time> thread being CCed across public-html and whatwg
- # [09:07] <shepazu> MikeSmith: if this is bugging you, feel free to crack the whip
- # [09:07] * Quits: Julian (chatzilla@217.91.35.233) (Ping timeout)
- # [09:08] <MikeSmith> shepazu: wasn't talking about the www-svg cross-posting, actually
- # [09:08] <MikeSmith> nor about the <time> thread either
- # [09:09] <hsivonen> MikeSmith: WebAIM, gawds?
- # [09:09] <hsivonen> MikeSmith: www-tag, rdfa-tf?
- # [09:09] * MikeSmith wonders what the hell that gawds list even is
- # [09:11] <shepazu> "The Guild of Accessible Web Designers"
- # [09:12] <MikeSmith> guild
- # [09:12] <MikeSmith> christ
- # [09:12] <MikeSmith> does anybody take the word "guild" seriously any more?
- # [09:13] <hsivonen> HTML Writers' Guild?
- # [09:13] <MikeSmith> hsivonen: case in point
- # [09:13] <shepazu> guild of the beguiled
- # [09:24] <MikeSmith> anyway, I'm sure the Guild of Accessible Web Designers is a great group of people
- # [09:31] <MikeSmith> my point was just that cross-posting is usually just not good
- # [09:32] <shepazu> MikeSmith: nice email, well-stated
- # [09:32] <MikeSmith> shepazu: which one?
- # [09:32] <shepazu> begin the beguine
- # [09:32] <MikeSmith> I can't remember saying anything nice today
- # [09:32] <shepazu> the RDFa one
- # [09:32] <MikeSmith> ah
- # [09:34] <shepazu> man, what am I still doing up?
- # [09:34] <MikeSmith> .t EST
- # [09:34] <phenny> Tue, 17 Mar 2009 03:31:55 EST
- # [09:34] <shepazu> I'm stupid tired.
- # [09:34] <MikeSmith> just stay up and watch the sunrise
- # [09:34] <MikeSmith> make that your goal for the day
- # [09:35] <shepazu> hey, phenny, you need to update for daylight saving time... it's 4:30
- # [09:35] <MikeSmith> .t EDT
- # [09:35] <phenny> Tue, 17 Mar 2009 04:32:43 EDT
- # [09:35] <shepazu> nn
- # [09:35] <MikeSmith> oyasumi
- # [10:11] * Joins: Julian (chatzilla@217.91.35.233)
- # [10:17] * Joins: tlr (tlr@128.30.52.30)
- # [10:27] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Quit: Tomorrow to fresh woods, and pastures new.)
- # [11:05] * Joins: ROBOd (robod@89.122.216.38)
- # [11:08] * Quits: Julian (chatzilla@217.91.35.233) (Connection reset by peer)
- # [11:20] * Joins: maddiin (mc@87.185.244.62)
- # [11:22] <karl> http://blogs.adobe.com/jd/2009/03/on_flash_killers.html
- # [11:22] <pimpbot> Title: jd/adobe: On Flash Killers (at blogs.adobe.com)
- # [11:22] <karl> >There is no "HTML5 standard", only a process which might potentially lead to a Recommendation but more plausibly would lead to something like HTML 3, XHTML or ECMAScript 4. And its history doesn't seem very open.
- # [11:24] <annevk> jd is quite the troll
- # [11:24] <annevk> see also his twitter account
- # [11:25] <hsivonen> annevk: URL?
- # [11:25] <hsivonen> http://twitter.com/jdowdell/ it seems
- # [11:25] <pimpbot> Title: Twitter / jdowdell (at twitter.com)
- # [11:27] * annevk was still trying to find it :)
- # [11:30] * Quits: tH (Rob@129.11.83.14) (Quit: ChatZilla 0.9.84-rdmsoft [XULRunner 1.9.0.1/2008072406])
- # [11:47] * Quits: Lachy (Lachlan@85.196.122.246) (Quit: This computer has gone to sleep)
- # [12:01] <jgraham> karl: Is there any reason to believe that the behaviour of tools such as tidy needs to be interoperable rather than driven by the particular requirements of the situation in which they are being used
- # [12:02] <jgraham> ? And is there any reason to assume that the HTML spec would be the right place to specify such behaviopur if specification was desirable?
- # [12:04] <karl> jgraham: yes for the document I receive. It's all the principle of canonicalization
- # [12:04] <karl> but it could be done in another group :) I don't mind.
- # [12:04] * jgraham doesn't understand
- # [12:05] <karl> hmm
- # [12:05] <jgraham> I don't think it is possible to have a single canonical form for every document
- # [12:05] <karl> 1. it is not out of scope of html 5, not less than a parsing algorithm.
- # [12:06] <jgraham> Because of e.g. CSS, scripts
- # [12:06] <karl> 2. it is not necessary mandatory or feasible to make it in the current html5 wg
- # [12:06] <jgraham> karl: A parsing algorithm is needed for client interoperability. Who benefits from interoperability of tidy-like tools?
- # [12:07] <karl> jgraham: there will be always people asking for perfection, and binary see the world. perfect or nothing. I'm not interested by extremes :)
- # [12:07] <karl> jgraham: people benefits!
- # [12:07] <karl> :)
- # [12:07] <jgraham> karl: Which people? I really have no idea…
- # [12:07] <karl> you never used tidy?
- # [12:08] <jgraham> karl: Uh, sure, once or twice.
- # [12:08] <jgraham> But if I make htmlClean or whatever then why does it have to be compatible with tidy?
- # [12:08] <karl> I know people who are using a lot :) I do often when I receive documents from elsewhere.
- # [12:09] <karl> ?
- # [12:09] <karl> I have to go right now. I'll be back online in 2 hours-ish
- # [12:09] * Quits: karl (karlcow@128.30.54.58) (Quit: This computer has gone to sleep)
- # [12:11] * Joins: tH (Rob@129.11.106.50)
- # [12:12] * Joins: Lachy (Lachlan@213.236.208.22)
- # [12:21] * Quits: gavin_ (gavin@99.226.207.11) (Ping timeout)
- # [12:26] * Joins: gavin_ (gavin@99.226.207.11)
- # [12:59] * Quits: tH (Rob@129.11.106.50) (Connection reset by peer)
- # [13:13] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [13:32] * Joins: karl (karlcow@128.30.54.58)
- # [13:36] * Quits: xover (xover@193.157.66.22) (Ping timeout)
- # [13:36] * Joins: xover (xover@193.157.66.22)
- # [13:36] * Joins: anne (annevk@77.163.243.203)
- # [14:14] * Joins: Julian (chatzilla@217.91.35.233)
- # [14:19] * Joins: myakura (myakura@122.29.116.63)
- # [14:52] * Joins: aroben (aroben@71.58.77.15)
- # [15:17] * Quits: karl (karlcow@128.30.54.58) (Quit: This computer has gone to sleep)
- # [15:44] * Joins: karl (karlcow@128.30.54.58)
- # [15:55] * Joins: rubys (rubys@75.182.92.38)
- # [15:55] <MikeSmith> .t EDT
- # [15:55] <phenny> Tue, 17 Mar 2009 10:52:40 EDT
- # [15:55] <MikeSmith> rubys: I may not be around for much longer today
- # [15:55] * Joins: billmason (bmason@69.30.57.118)
- # [15:55] <rubys> ok
- # [15:56] <MikeSmith> rubys: you at home? traveling?
- # [15:56] <rubys> this week home
- # [15:56] <MikeSmith> OK
- # [15:56] <MikeSmith> I'm on a call right now that has sucked most of the life out of me for today
- # [15:57] * rubys hates when that happens
- # [16:20] <shepazu> http://blog.jclark.com/2009/01/relax-ng-and-xmlid.html
- # [16:20] <pimpbot> Title: James Clark's Random Thoughts: RELAX NG and xml:id (at blog.jclark.com)
- # [16:20] <anne> xml:id should never have happened
- # [16:29] <shepazu> anne, having looked pretty deeply into the issue, I now agree with that statement
- # [16:30] <shepazu> I'm curious why you think that, though?
- # [16:30] <anne> I think they should just have used id="" so we only get one ID per element without confusing if/else logic
- # [16:31] <shepazu> ok
- # [16:32] <anne> depending on what Gecko/WebKit/Trident do that might still happen I suppose, at least for the Web
- # [16:32] <shepazu> there are problems with that approach, but I think that in retrospect, co-opting "id" would probably have been a net benefit for XML and the Web
- # [16:32] <shepazu> though a distinct loss for CML
- # [16:33] <anne> The Council of Mortgage Lenders?
- # [16:33] <anne> Chronic Myelogenous Leukemia?
- # [16:34] <anne> Cinematography Mailing List?
- # [16:34] <shepazu> Council of Mortgage Lenders
- # [16:34] <shepazu> xml:id is the reason we have this budget crisis
- # [16:35] <shepazu> oh, also Chemical Markup Language, which uses @id in an interesting but problematic way
- # [16:37] * Lachy wonders how an attribute can cause a budget crisis
- # [16:38] <anne> ask Hixie
- # [16:38] <shepazu> Lachy, the classic draconian error handling argument by Bray should tell you
- # [16:38] <anne> they're very expensive to introduce :p
- # [16:43] <Philip> shepazu: The argument that assumes you're going to send highly critical data with not even a checksum, and will rely on syntax checking to detect any transmission errors despite it detecting only a small proportion of such errors?
- # [16:44] <shepazu> Philip: the very same
- # [16:44] <anne> :)
- # [16:44] <shepazu> for all we know, that's what caused the collapse of Wall Street... you permissive error-handlers are to blame!
- # [16:45] * shepazu looks around for some tar and feathers and a pitchfork
- # [16:45] <anne> or you "well-formedness checking is enough" people are
- # [16:47] * gsnedders notes his blog getting sent incorrectly will not cause a budget crisis, thus Bray's argument does not apply to his blog
- # [16:48] <shepazu> anne's a witch! burn him!
- # [16:49] <Philip> gsnedders: But there is a separate argument that it's good to share technologies between different parts of the world (so that everyone can benefit from shared tools and skills and expertise etc), and so arguments about other uses of the technology are still relevant to your blog
- # [16:49] <anne> actually, despite appearances, I weigh more than a duck
- # [16:49] <anne> so I can't be a witch
- # [16:50] <gsnedders> anne: Not in a vacuum
- # [16:50] <jgraham> anne can't be a witch in a vacuum?
- # [16:50] <gsnedders> (I meant the vacuum of space, but that's still untrue)
- # [16:50] <Philip> So you don't mean a vacuum at all?
- # [16:50] <Dashiva> I think gsnedders is the witch
- # [16:50] <jgraham> Anyway, having a girl's name seems prety suspicious on the witch-o-meter ;)
- # [16:51] <gsnedders> (Seeming although the effects of gravity are negligible, there is still a gravitational force, and it will be larger on anne than on the duck, thus meaning anne weighs more)
- # [16:52] <rubys> I thought that as long as a person was still alive, you couldn't prove non-witch-ness.
- # [16:52] * jgraham hypothesises that the collapse of wall street was caused by an xml document like <transaction><rating>insane<rating></transaction>
- # [16:58] <Julian> Philip: isn't the checksum provided by the transport (in this case HTTP over TCP/IP, if you#re referring to the RSS/Atom examples Tim Bray mentioned)?
- # [17:02] <Philip> Julian: I'm referring to the example that's about receiving a possibly-truncated message like "<transaction type='credit'><amount>100000", where draconian syntax error handling saves the day and prevents you from interpreting the message incorrectly
- # [17:04] <shepazu> anne, who are you, who are so wise in the ways of science?
- # [17:05] * Joins: rubys1 (rubys@75.182.92.38)
- # [17:05] <Philip> ...but that doesn't seem a very good way of ensuring integrity, and when you have a better way of ensuring integrity there's no longer that need for draconian error handling
- # [17:05] <Julian> Philip: i think it does. If the truncation occurs on the application layer.
- # [17:05] <Philip> (I think I've seen that example on Tim Bray's blog, though I don't know where exactly)
- # [17:05] <Philip> Julian: The integrity checking should be on the application layer too
- # [17:05] <anne> shepazu, I am ... nah, won't do this :p
- # [17:05] <Philip> Julian: (i.e. in addition to the TCP checksum stuff)
- # [17:05] * Parts: rubys1 (rubys@75.182.92.38)
- # [17:06] <Julian> Philip: well, one kind of that integrity checking *is* to rely on XML-wellformedness.
- # [17:06] <shepazu> :D
- # [17:06] <Philip> http://www.tbray.org/ongoing/When/200x/2004/01/11/PostelPilgrim - ah, that's it
- # [17:06] <pimpbot> Title: ongoing On Postel, Again (at www.tbray.org)
- # [17:06] <Julian> You may not like it, but it seems to do the job in this case.
- # [17:09] * Quits: Lachy (Lachlan@213.236.208.22) (Quit: This computer has gone to sleep)
- # [17:10] <shepazu> one solution, which I favor, is having an XML attribute in the root that indicates whether or not to apply error-correction, which defaults to yes
- # [17:10] <Philip> Julian: Hmm, I suppose I'm only thinking of transmission errors, in which case it works when the only possible error is truncation, but it seems unreasonable to assume no other errors are possible, and if you add a mechanism that handles the other possible errors (e.g. with a checksum of the whole document) then the wellformedness is redundant
- # [17:11] <Philip> But I'm probably wrong to only think of it in that way
- # [17:11] <shepazu> that way, all of Bray's financial data can say, "be strict", and everything else can do error correction
- # [17:11] <gsnedders> shepazu: Implementations will probably ignore that flag anyway
- # [17:12] <shepazu> gsnedders: why would they?
- # [17:12] <shepazu> there's value in it
- # [17:12] <Philip> because you're more likely to have some other code that produces the XML document string, then passes it to the communication layer which handles all the checksums and stuff, and if you e.g. ran out of buffer space in the code that constructs the XML document then no checksums will help
- # [17:13] <gsnedders> shepazu: There is always demand from users to be lax and to show something. With feeds, for example, so much is already non-draconian that the flag would be ignored.
- # [17:14] <Philip> shepazu: People think they can produce well-formed XML, so they would never bother setting the flag to enable error-correction, and would publish erroneous documents like they do now
- # [17:14] <jgraham> Philip: But you are relying on the clent to do the error checking anyway so you could simply make the client check for validity in cases where that is important
- # [17:14] <Philip> shepazu: Oops, I think I read your default the wrong way round
- # [17:14] <shepazu> gsnedders: not "always"... stock traders would have a very different sense of the relative importance of speed vs. accuracy
- # [17:15] <gsnedders> shepazu: My point is there would be parsers that would ignore the flag, making having the flag useless
- # [17:15] <Philip> jgraham: Validity checking is hard and nobody does it; well-formedness checking is easy because only the library writer has to care about it
- # [17:16] <shepazu> (actually, apparently in recent history, stock traders have shown pretty bad judgment there, but my point stands)
- # [17:16] <jgraham> Philip: But these are Super Importatn Fianacial Transactions. If they are really relying on well-formedness checking as a significant part of the validity checking they are doing it wrong
- # [17:17] <shepazu> gsnedders: I don't think you understand the use case here... there will be UAs that could enforce that, and for certain applications, people would use those *instead* of the lax ones
- # [17:17] * Philip wonders if stock traders actually use XML, or if they want higher-performance protocols (given how much a few milliseconds can matter)
- # [17:17] <hsivonen> shepazu: sure enough, CML was the first XML language to get a dedicated third-party Validator.nu front end...
- # [17:18] <shepazu> hsivonen: I talked to the CML guys, and their problem is pretty intractable from a universal shift to @id==ID, but I like your suggestion of simply grandfathering them
- # [17:33] * Disconnected
- # [17:33] * Attempting to rejoin channel #html-wg
- # [17:33] * Rejoined channel #html-wg
- # [17:33] * Topic is 'HTML WG 12 Mar http://lists.w3.org/Archives/Public/public-html-wg-announce/2009JanMar/0041.html (This channel is logged: http://krijnhoetmer.nl/irc-logs/ )'
- # [17:33] * Set by DanC on Thu Mar 12 17:00:49
- # [17:38] * Disconnected
- # [17:38] * Attempting to rejoin channel #html-wg
- # [17:38] * Rejoined channel #html-wg
- # [17:38] * Topic is 'HTML WG 12 Mar http://lists.w3.org/Archives/Public/public-html-wg-announce/2009JanMar/0041.html (This channel is logged: http://krijnhoetmer.nl/irc-logs/ )'
- # [17:38] * Set by DanC on Thu Mar 12 17:00:49
- # [19:40] * Disconnected
- # [19:40] * Attempting to rejoin channel #html-wg
- # [19:40] * Rejoined channel #html-wg
- # [19:40] * Topic is 'HTML WG 12 Mar http://lists.w3.org/Archives/Public/public-html-wg-announce/2009JanMar/0041.html (This channel is logged: http://krijnhoetmer.nl/irc-logs/ )'
- # [19:40] * Set by DanC on Thu Mar 12 17:00:49
- # [19:46] <karl> anne! go out of shepazu's body!
- # [19:47] <shepazu> SVG sucks!!
- # [19:48] <Lachy> oh, oops :-)
- # [19:55] * Quits: adele (adele@17.246.16.176) (Quit: adele)
- # [20:08] * Joins: Sander (svl@86.87.68.167)
- # [20:37] * Parts: asbjornu (asbjorn@84.48.116.244)
- # [20:51] * Joins: tlr (tlr@128.30.52.30)
- # [21:08] * Joins: adele (adele@17.246.16.176)
- # [21:13] * Joins: maddiin (mc@87.185.244.62)
- # [21:35] * Joins: dbaron (dbaron@198.214.235.51)
- # [21:37] * Parts: rubys (rubys@75.182.92.38)
- # [21:52] * Quits: karl (karlcow@128.30.54.58) (Quit: This computer has gone to sleep)
- # [22:25] * Joins: asbjornu (asbjorn@84.48.116.244)
- # [22:27] * Quits: heycam (cam@124.168.80.126) (Quit: bye)
- # [22:39] * Quits: ROBOd (robod@89.122.216.38) (Quit: http://www.robodesign.ro )
- # [22:42] * Quits: dbaron (dbaron@198.214.235.51) (Quit: 8403864 bytes have been tenured, next gc will be global.)
- # [22:46] * Quits: aroben (aroben@71.58.77.15) (Connection reset by peer)
- # [23:10] * Parts: billmason (bmason@69.30.57.118)
- # [23:23] * Quits: jgraham (jgraham@74.53.238.210) (Ping timeout)
- # [23:25] * Quits: Sander (svl@86.87.68.167) (Quit: And back he spurred like a madman, shrieking a curse to the sky.)
- # [23:27] * Joins: jgraham (jgraham@74.53.238.210)
- # [23:33] * Quits: jgraham (jgraham@74.53.238.210) (Ping timeout)
- # [23:54] * Quits: Dashiva (noone@80.202.223.46) (Quit: Dashiva)
- # Session Close: Wed Mar 18 00:00:00 2009
The end :)