| 00:01 | <Philip`> | gsnedders: If the spec-gen output well-formed XML, then the spec-splitter could load it in approximately no time and so the parsing cost would no longer be a problem |
| 00:01 | <annevk> | sicking, where does it say that? |
| 00:01 | <gsnedders> | Philip`: It can't easily output both |
| 00:01 | <annevk> | sicking, #Method means comma-separated list of Method names per RFC 2616 |
| 00:01 | <annevk> | sicking, same for #field-name |
| 00:01 | <gsnedders> | Hixie: Creating a really basic form isn't that much more than what I meant by a web UI :) |
| 00:01 | <Philip`> | gsnedders: Output one file that is a subset of both HTML and XHTML :-) |
| 00:01 | <annevk> | sicking, # has special meaning |
| 00:02 | <sicking> | annevk, ah, i missed the # thing |
| 00:02 | <Philip`> | Hixie: I probably won't be able to do anything with the spec-splitter in the next few days, but then I'll probably forget so someone will have to remind me |
| 00:03 | <Hixie> | k |
| 00:03 | <Hixie> | np |
| 00:03 | <sicking> | btw, has anyone checked what XDR in IE8b2 does? |
| 00:04 | <annevk> | Hixie, I'm interested in the Web service |
| 00:04 | <annevk> | (assuming it works with curl or something) |
| 00:05 | annevk | doesn't like how WS-Pain took "Web Services" |
| 00:05 | <Hixie> | gsnedders: http://junkyard.damowmow.com/340 |
| 00:05 | <Hixie> | gsnedders: there's your ui, fill in the "..." bit as needed :-) |
| 00:06 | <gsnedders> | Hixie: :D |
| 00:06 | <annevk> | Hixie, doesn't work |
| 00:06 | <gsnedders> | Hixie: The problem is I've never written any web thing before in Python, and I don't particularly care for a web ui |
| 00:06 | <annevk> | Hixie, http://junkyard.damowmow.com/...?url=http%3A%2F%2Fdev.w3.org%2F2006%2Fwaf%2Faccess-control%2FOverview.src.html |
| 00:06 | <Hixie> | annevk: :-P |
| 00:06 | <gsnedders> | annevk: look at the source. |
| 00:06 | <gsnedders> | annevk: It's a UI. Nothing more. |
| 00:06 | <gsnedders> | <form action="..." method=get> |
| 00:06 | <Hixie> | gsnedders: ok i can hook something up |
| 00:07 | <annevk> | oh, I see, lol |
| 00:07 | <Hixie> | gsnedders: is there a tarball somewhere? |
| 00:07 | <Hixie> | annevk: i said it was the ui, nothing else :-P |
| 00:07 | <gsnedders> | Hixie: See the download links from http://anolis.gsnedders.com/ |
| 00:07 | <gsnedders> | Hixie: See the "anolis" and "runtests.py" files |
| 00:08 | <gsnedders> | Hixie: If you wanted to be über-clever, you could probably actually run it as a pre-commit hook in SVN |
| 00:09 | <gsnedders> | Anyhow, I'm off to bed |
| 00:10 | <gsnedders> | http://stuff.gsnedders.com/spec-gen/html5.html is a version of the spec from earlier today |
| 00:11 | <Hixie> | anyone know how i install to a non-default location, like "." ? |
| 00:11 | <gsnedders> | Hixie: I could just send you the email I wrote a while ago, though one or two bits are a bit outdated, but if you ignore the stuff about 1.0b2, I think it's fine |
| 00:11 | <Hixie> | sure |
| 00:11 | <gsnedders> | Hixie: that's in the email |
| 00:11 | <Hixie> | send send send! :-D |
| 00:11 | gsnedders | sends |
| 00:11 | <Hixie> | wooo |
| 00:11 | <gsnedders> | it also refers to the program as spec-gen, while it's now anolis |
| 00:14 | <Hixie> | byte-compiling /home/ianh/bin/anolis-spec-gen/lib/python2.3/site-packages/anolislib/processes/toc.py to toc.pyc |
| 00:14 | <Hixie> | File "/home/ianh/bin/anolis-spec-gen/lib/python2.3/site-packages/anolislib/processes/toc.py", line 191 |
| 00:14 | <Hixie> | sections.extend((child_section, depth + 1) for child_section in reversed(section)) |
| 00:14 | <Hixie> | ^ |
| 00:14 | <Hixie> | SyntaxError: invalid syntax |
| 00:14 | <annevk> | (Anolis name flames can be e-mailed directly to annevan...) |
| 00:16 | <Hixie> | i'd fix that line but i have no idea what it's trying to do |
| 00:17 | <annevk> | create a list of tuples? |
| 00:17 | <gsnedders> | Hixie: Py 2.3 :( |
| 00:17 | <annevk> | Hixie, ask markp :) |
| 00:17 | <annevk> | prolly easiest |
| 00:18 | <gsnedders> | Hixie: Trying putting another ( at the beginning, and ), before the final ) |
| 00:18 | <gsnedders> | (note the comma) |
| 00:18 | <gsnedders> | Actually, that's nonsense |
| 00:18 | <gsnedders> | [ and ,] |
| 00:19 | <gsnedders> | peh. |
| 00:19 | <gsnedders> | I'll deal with this tomorrow |
| 00:19 | <gsnedders> | Hixie: Install a newer version of python :P |
| 00:19 | <annevk> | for child_section in reversed(section): |
| 00:19 | <annevk> | temp_list.append((child_section, depth + 1)) |
| 00:19 | <annevk> | sections.extend(temp_list) |
| 00:20 | <gsnedders> | annevk: I lost the whitespace in that |
| 00:20 | <annevk> | i didn't put any in |
| 00:20 | <Hixie> | annevk: i don't need to declare temp_list or something? |
| 00:20 | <annevk> | yeah, that too |
| 00:20 | <annevk> | temp_list = [] |
| 00:20 | <annevk> | and some spaces before temp_list.append |
| 00:20 | <gsnedders> | There probably is a better way of fixing that though |
| 00:21 | <Hixie> | woot it compiled |
| 00:21 | <annevk> | yay for remote debugging |
| 00:21 | <gsnedders> | sections.extend([(child_section, depth + 1) for child_section in reversed(section)]) |
| 00:21 | <annevk> | (aka guessing) |
| 00:21 | <gsnedders> | That should work |
| 00:22 | <annevk> | isn't it just that inline for doesn't work in Python 2.3? |
| 00:22 | <gsnedders> | (I think) |
| 00:22 | gsnedders | shrugs |
| 00:22 | <gsnedders> | I only claim to support 2.5 anyway :P |
| 00:22 | <gsnedders> | Mainly because I'm too lazy to test |
| 00:23 | <gsnedders> | Umm… |
| 00:23 | <Hixie> | something about a missing xProfile |
| 00:23 | <gsnedders> | It will fail to run on 2.3 |
| 00:23 | <Hixie> | cProfile even |
| 00:24 | <gsnedders> | reversed() doesn't exist either |
| 00:24 | <Hixie> | awww, failed upgrade. :-( |
| 00:24 | <Hixie> | oh well |
| 00:25 | <Hixie> | i look forward to RC2! |
| 00:25 | <annevk> | there's a way to get newer python versions on dreamhost |
| 00:25 | <Hixie> | oh? |
| 00:25 | <gsnedders> | annevk: sections.extend() (above) is a generator expression, which is 2.4 |
| 00:25 | annevk | looks for a wiki page |
| 00:25 | <gsnedders> | Hixie: supporting Py2.3 would mean moving back on to deprecated features |
| 00:26 | <Hixie> | well i have some stuff to do now anyway |
| 00:26 | <Hixie> | i'll be back in a bit |
| 00:26 | <annevk> | eg http://wiki.dreamhost.com/Python#Building_a_custom_version_of_Python |
| 00:27 | jgraham | has python2.5 at james.html5.org |
| 00:27 | <gsnedders> | It should be fairly easy to get it working on 2.4, it probably already does |
| 00:28 | <jgraham> | gsnedders: FWIW it would be nice if there were a utility function in anolislib that took a file-like object and a set of options and returned a processed document |
| 00:28 | <jgraham> | So the web interface would just be |
| 00:28 | <jgraham> | doc = urllib2.urlopen(url) |
| 00:28 | <gsnedders> | Yeah, it would be nice |
| 00:28 | <gsnedders> | A lot of things would be nice :) |
| 00:28 | <jgraham> | anolislib.gen(doc) |
| 00:29 | <jgraham> | gsnedders: Well all the code for thi thing is already in anolis. It's just refactoring |
| 00:29 | <gsnedders> | jgraham: The processed doc being a StringIO object, or an ElementTree, or…? |
| 00:29 | <annevk> | btw, the Web interface (whoever will set it up) needs file input as well |
| 00:29 | <annevk> | the Web service rather |
| 00:30 | <jgraham> | gsnedders: Don't really care. I guess an ElementTree is best but a string is more convenient |
| 00:30 | <jgraham> | (a StringIO would be weird) |
| 00:35 | gsnedders | heads off, again |
| 00:41 | <annevk> | Hixie, using #!/usr/bin/python2.5 at the start of the file should do the trick |
| 00:41 | <annevk> | Hixie, through shell I suppose you need to use python2.5 as programm |
| 00:42 | <annevk> | s/programm/program/ |
| 00:44 | annevk | -> bed |
| 02:06 | <Hixie> | jgraham: if you're still around, i could totally use a service on james.html5.org if you set one up |
| 02:06 | <Hixie> | i'd rather not have to set things up so that the script runs on my machine, for load reasons |
| 03:30 | <sicking> | Hixie, can you convince the gmail team to write a tool to import IMAP mail? |
| 03:30 | <sicking> | Hixie, pretty please? |
| 03:53 | <MikeSmith> | Hixie: <butt> |
| 03:53 | <MikeSmith> | I don't think there is likely much existing content with that in it |
| 03:57 | <roc> | sicking: gmail supports imap, so can't you just pull up an IMAP client (say Thunderbird), open both accounts, and drag-and-drop from one to the other? |
| 03:58 | <sicking> | roc, yeah, that should work in theory |
| 03:58 | <sicking> | roc, though last i tried thunderbird didn't like connecting to gmail as IMAP, dunno why |
| 03:59 | <sicking> | roc, and last i copied mail between accounts using thunderbird i basically lost the majority of mails due to thunderbird bugs |
| 03:59 | <roc> | hehe |
| 03:59 | <roc> | just write a script to talk IMAP directly to the two servers |
| 03:59 | <roc> | how hard could it be |
| 03:59 | <sicking> | dunno how hard imap is |
| 03:59 | <sicking> | you might be right |
| 03:59 | <roc> | it's ugly |
| 04:00 | <sicking> | i hear python solves all programming problems ever |
| 04:00 | <sicking> | import IMAP-simple |
| 04:00 | <sicking> | ? |
| 04:22 | <jacobolus> | roc, sicking: yes, opening up an imap client and moving emails into Gmail works just fine |
| 04:22 | <jacobolus> | sicking: if you want you can just add emails there using the IMAP protocol directly |
| 04:22 | <jacobolus> | write a little script or something |
| 04:22 | <sicking> | yeah |
| 04:23 | <sicking> | i just found that thunderbird has a copy function though, so i can probably use that |
| 04:23 | <sicking> | rather than the move which i tried last i did something like this |
| 04:59 | <roc> | is "seeked" really a word? |
| 05:00 | <jacobolus> | yes |
| 05:00 | <jacobolus> | erm, no |
| 05:00 | <jacobolus> | sought |
| 05:00 | <sicking> | sunk? |
| 05:00 | <sicking> | :) |
| 05:01 | <roc> | Hixie thinks it is :-) http://www.whatwg.org/specs/web-apps/current-work/#seeked |
| 05:02 | <jacobolus> | yeah, seeked sucks as a name for that |
| 05:02 | <jacobolus> | I'm not sure sought is much better though |
| 05:02 | <jruderman> | 'doneseeking'? |
| 05:05 | <jacobolus> | roc: then again, networking specs have a proud tradition of mis-spelling words, e.g. 'referer' |
| 05:13 | <Hixie> | seeked isn't misspelt. it's just a made up word. |
| 05:13 | <Hixie> | onsought would confusing the heck out of people |
| 05:14 | <Hixie> | confuse, even |
| 05:14 | <sicking> | doneseeking sounds like a good idea to me |
| 05:14 | <GregHouston> | Could throw some nice archaic terms in there like beseeched. |
| 05:14 | <Hixie> | what's wrong with 'onseeked'? |
| 05:15 | <sicking> | it's a made up word |
| 05:17 | <Hixie> | so? half of our industry uses made up words |
| 05:17 | <sicking> | no other event names do |
| 05:17 | <sicking> | s/do/are/ |
| 05:18 | <sicking> | why is onseeked better than ondoneseeking? |
| 05:19 | <Hixie> | there isn't a single other event of the form ondoneX |
| 05:20 | <sicking> | that seems like less of an addition than adding made up words |
| 05:20 | <Hixie> | we have lots of made up words :-) |
| 05:21 | <sicking> | we do? and not in event names |
| 05:21 | <sicking> | and are those truly made up, or using commonly used but not in marriam webster, words? |
| 05:22 | <GregHouston> | One convention is to add onsuccess or oncomplete to the event name, e.g., seekoncomplete, not that seeked bothers me any. |
| 05:23 | <sicking> | onseekcomplete sounds good too |
| 05:23 | <roc> | afterseek? |
| 05:23 | <GregHouston> | Sounds like a nightclub roc. :) |
| 05:24 | <Hixie> | i have approximately zero interest in worrying about event names, so if y'all realy do have a problem with 'onseeked', then implement something else and tell me what it is |
| 05:24 | <roc> | I don't care all that much either to be honest |
| 05:24 | <sicking> | same here, but it seems easy to fix |
| 05:24 | <Hixie> | (make sure the apple guys implement it too) |
| 05:24 | <Hixie> | (and opera i guess) |
| 05:25 | <sicking> | do either of them do seeked? |
| 05:25 | <roc> | Hixie, from my reading of this seek spec, if the seek fails for some reason, we should just fire a 'seeked' event and its up to the script to figure out that the seek didn't get to the right place |
| 05:25 | <Hixie> | sicking: no idea :-) |
| 05:25 | <roc> | is that correct? |
| 05:25 | <roc> | I mean if it fails around step 13 |
| 05:26 | <Hixie> | roc: correct; the event is really only meant for putting up "Seeking..." UI |
| 05:26 | <roc> | ok |
| 05:28 | <roc> | sicking: you might want to avoid posting with the name "Domain Admin" |
| 05:29 | <sicking> | roc, ugh! |
| 05:29 | <sicking> | roc, thanks |
| 06:40 | <Dashiva> | This has to be a new record in pointlessness: http://lists.w3.org/Archives/Public/public-html/2008Aug/0842.html |
| 06:41 | <jacobolus> | Dashiva: don't tempt the internets |
| 06:43 | <Dashiva> | I wear hubris like a crown already |
| 07:12 | <MikeSmith> | Dashiva: If they had a King of Fools, then I could wear that crown... and you can all die laughing because I'll wear it proudly. |
| 07:13 | <MikeSmith> | wow, a plus one with double exclamation |
| 07:13 | <MikeSmith> | I think that earns extra points in the olympics |
| 07:14 | <MikeSmith> | at least with the east German judges |
| 07:15 | <GregHouston> | It's really our fault as developers. We encourage everyone to grade every bit of content with little ajax star selectors. |
| 07:16 | <jacobolus> | GregHouston: as in, five stars, $('* * * * *').… ?? |
| 07:19 | <jacobolus> | ouch: http://intertwingly.net/blog/2008/08/28/Improved-Namespace-Support |
| 07:19 | <GregHouston> | Right, !! probably equates to "I wish I could have given it six stars!" |
| 07:20 | <jacobolus> | i guess ppl already saw that |
| 07:25 | <GregHouston> | Jacob, I'm a little slow, I just got the CSS selector part of your joke. |
| 07:26 | <jacobolus> | GregHouston: I thought it was *your* joke?! |
| 07:26 | <GregHouston> | Lol, I think you took my joke a step further and made it geekier. Haha. |
| 07:30 | <hdh> | howcome two votes make a 4.7 score? does microsoft round down? |
| 07:31 | <GregHouston> | Lol. After making fun of voting. I actually made that second vote, and I asked myself the same question. I thought maybe I didn't have cursor far enough over to the right when I clicked on the fifth star. |
| 07:34 | <GregHouston> | hdh, I'm assuming you followed the same path I did and landed here: https://connect.microsoft.com/IE/feedback/ViewFeedback.aspx?FeedbackID=364356 |
| 07:35 | <hdh> | right; that's the only thing interesting (apart from I can write ;-- in Connect password) |
| 07:50 | <Dashiva> | "So we should raise the conformance standards for accessibility, but also raise the conformance standard for authoring tool so that they re even required to produce non-conforming HTML whenever the author provides insufficient information to make the document conforming." |
| 07:50 | <Dashiva> | "We'll also raise the conformance standard for spec writers, by making it non-conforming not to shoot yourself in the shoot." |
| 07:57 | <GregHouston> | In Javascript when I shoot myself in the shoot I get a too much recursion error. |
| 07:58 | <Dashiva> | In spec-writing, that's just an implementation detail |
| 08:01 | <Dashiva> | Hixie: Has there been much brainstorming about <bb> names? |
| 08:02 | <Hixie> | some |
| 08:02 | <Hixie> | see the logs |
| 08:03 | Dashiva | tries to search for <bb> and gets every mention of bb.online.no |
| 08:03 | <zcorpan> | jgraham: fwiw, it's possible to exploit a security parsing bug in older operas (pre-9.52 i think) with unquoted attributes and html5lib, so using quoted attributes by default (at least if it contains any offending characters) would be good |
| 08:04 | <Hixie> | oooh, hsivonen's idea of an XSLT-specific PUBLIC FPI makes sense |
| 08:04 | <Dashiva> | Hixie: What about <ui> or <uibtn> or <uibutton> or something like that? Would that be misrepresenting the purpose? |
| 08:05 | <zcorpan> | jgraham: consider <a href=#{U+000A}onclick=doSomethingEvil()> ... or hmm, does html5lib emit quotes if the value contains an = ? |
| 08:05 | <Hixie> | Dashiva: maybe |
| 08:05 | <Hixie> | Dashiva: send mail on the thread, it's not really something i want to worry about now |
| 08:06 | <Dashiva> | Hey, not outright rejection. That's a good start :) |
| 08:06 | <Hixie> | :-) |
| 08:07 | <Hixie> | i'd want something that is clearly different from <command> and <button> if possible |
| 08:08 | <zcorpan> | <differentfromcommandandbutton> |
| 08:09 | <zcorpan> | clear enough? |
| 08:09 | <Hixie> | no :-P |
| 08:09 | <Hixie> | hsivonen: maybe "XSLT-generated"? |
| 08:10 | <zcorpan> | <clearlydifferentfromcommandandbutton> |
| 08:10 | <Hixie> | too long |
| 08:10 | <Hixie> | :-P |
| 08:10 | <zcorpan> | :( |
| 08:10 | <zcorpan> | <cdfcab> |
| 08:11 | <Dashiva> | <cdfcab> |
| 08:11 | <Dashiva> | dang |
| 08:12 | <Dashiva> | That's not so bad, either. CDF is a popular acronym, and it drives around in a cab. |
| 08:12 | <GregHouston> | It's sort of like YAML, "YAML Ain't a Markup Language" |
| 08:12 | <hdh> | CDF Cabinet, to go with the zipped-XML fever |
| 08:14 | <zcorpan> | hmm, crappy that one needs to reboot for ie8b2 |
| 08:14 | <Dashiva> | On the other hand, it might make people think it's for embedding widgets |
| 08:16 | <hdh> | I heard someone rebooted twice, and contemplated clicking "Cancel" after the old IE is gone and IE8 not yet installed |
| 08:24 | <zcorpan> | hmm. how does one know if ie8b2 was installed successfully? any characteristics to check for that differentiates it from ie8b1? |
| 08:25 | <zcorpan> | :first-line and :first-letter doesn't seem to work in standards mode |
| 08:28 | <Lachy> | zcorpan, Help>About |
| 08:28 | <Lachy> | you may need to turn on the menu bar. it's off by default |
| 08:29 | <zcorpan> | Lachy: it just says "Beta" |
| 08:30 | <zcorpan> | version 8.0.6001.17184 |
| 08:30 | <zcorpan> | hmm that's beta 1 it seems |
| 08:30 | <zcorpan> | why wasn't beta 2 installed then |
| 08:31 | <jacobolus> | the menu bar is off by default? |
| 08:31 | <jacobolus> | how does that work? |
| 08:32 | <zcorpan> | i unchecked the "install updates" during the install process, did that make the installer not do anything? |
| 08:33 | <Lachy> | zcorpan, mine says "Update versions: beta 2" |
| 08:33 | <Lachy> | Version 8.0.6001.18241 |
| 08:36 | <zcorpan> | Lachy: ok, running the installer again, this time with the checkbox checked |
| 08:37 | <zcorpan> | i would have presumed that unchecking the checkbox meant "only install ie8b2, don't update windows and everything else" |
| 08:37 | <zcorpan> | and not "don't do anything" |
| 08:38 | <hdh> | http://homepage.mac.com/bradster/iarchitect/controls.htm#element(/1/2/19) happened before |
| 08:38 | <hdh> | search for "Grrr" |
| 08:50 | <zcorpan> | still no luck :( |
| 08:50 | <zcorpan> | oh well |
| 08:55 | <Lachy> | zcorpan, try uninstalling beta 1 first, and then installing beta 2 |
| 09:00 | <collinjackson> | should the sandbox attribute of iframes be settable dynamically? |
| 09:18 | <zcorpan> | does someone in the html5 community know stuff about silverlight? |
| 09:19 | <Hixie> | what stuff? |
| 09:19 | <zcorpan> | i mean if we don't know anything about it how are we supposed to make html5 competitive? |
| 09:20 | zcorpan | rebooting |
| 09:22 | <roc> | people know things about it |
| 09:23 | <hsivonen> | it seems to me that the Open Web is lacking a killer IDE compared to Flash and Silverlight |
| 09:24 | <roc> | tools are definitely a big problem |
| 09:24 | <Hixie> | yes |
| 09:24 | <Hixie> | GWT and Firebug are basically it right now |
| 09:25 | <hsivonen> | GWT doesn't do vector graphics yet, though |
| 09:25 | <hsivonen> | and using GWT with Eclipse doesn't let you arrange SVG stuff graphically |
| 09:26 | <roc> | it's hard to figure out what to do in the tools space |
| 09:27 | <zcorpan> | so there's obviously a hole in the market for a standards based IDE |
| 09:27 | <zcorpan> | why isn't it filled? |
| 09:27 | <zcorpan> | i mean |
| 09:27 | <hsivonen> | HTML, SVG and CSS are too expressive to make a GUI tool import everything |
| 09:27 | <zcorpan> | an IDE for standards based apps |
| 09:28 | <hsivonen> | however, polished tool could output to HTML, SVG, CSS and JS |
| 09:28 | <roc> | there's no money in it and it's a lot of work |
| 09:29 | <zcorpan> | why is there no money in it? |
| 09:30 | <zcorpan> | would it be possible to write a plugin for existing IDEs that doesn't change the editing process, just the output? |
| 09:30 | <roc> | most tools are free |
| 09:31 | <roc> | only the most entrenched tools have people paying for them |
| 09:32 | <hsivonen> | we need a google search field in IDEs. :-) |
| 09:32 | <roc> | you might want to check this out: http://www.evolus.vn/Pencil/ |
| 09:32 | zcorpan | ponders whether it would be Good for Google to develop an IDE for standards based apps |
| 09:33 | <hsivonen> | GWT lacks an Interface Builder with flashy vector graphics eye candy goodness |
| 09:33 | <roc> | another problem is that for Web apps you usually have a big server piece |
| 09:33 | <Hixie> | there are plenty of wysiwyg tools |
| 09:34 | <roc> | so you need to integrate with PHP or Java or ASP.NET or whatever |
| 09:34 | <Hixie> | dreamweaver |
| 09:34 | <Hixie> | frontpage |
| 09:34 | <Hixie> | nvu |
| 09:34 | <Hixie> | etc |
| 09:34 | <roc> | but they suck |
| 09:34 | <Hixie> | but they suck, as you say |
| 09:34 | <roc> | NVu is obsolete |
| 09:34 | <hsivonen> | roc: but Flash and Silverlight apps need a server side whenever an Open Web app would need a server side |
| 09:34 | <Hixie> | what's needed is an IDE (app creator, not a web page creator) that does both client side and server side and does a good job |
| 09:35 | <Hixie> | that's non-trivial |
| 09:35 | <zcorpan> | Hixie: wouldn't providing such an IDE be Good for Google? |
| 09:35 | <Hixie> | yes |
| 09:35 | <Hixie> | google has GWT and App Engine in this space |
| 09:36 | <Hixie> | but we've only just come into this space |
| 09:36 | <Hixie> | i mean, in the last year or so |
| 09:36 | <Hixie> | we're still learning the ropes |
| 09:36 | <hsivonen> | interestingly, all of Flash, Silverlight and GWT treat semantic HTML as a bug |
| 09:36 | <Hixie> | and IDE's aren't our core competency (search and ads) |
| 09:36 | <Hixie> | yeah, actually doing it right (semantically) in a wysiwyg environment is an unsolved problem |
| 09:37 | <hsivonen> | I like the GWT compiler, but I have reservations about the toolkit side |
| 09:37 | <hsivonen> | but compilers are easier than user-facing stuff in a way |
| 09:37 | jgraham | notes that seked is a real word per OED |
| 09:37 | <jgraham> | s/seked/seeked/ |
| 09:37 | <hsivonen> | for example, the Google Docreader (a GWT app) breaks Google indexability |
| 09:38 | <hsivonen> | there's some irony there |
| 09:38 | <roc> | what I'd really like to do is something a bit more disruptive than taking on the IDE space head-on |
| 09:38 | <Hixie> | yeah i don't like the docreader |
| 09:39 | <Hixie> | roc: like what? |
| 09:39 | <roc> | and that is, leverage the browser by improving contenteditable/designmode |
| 09:39 | <roc> | get SVG into wikis and email that way |
| 09:39 | <Hixie> | well i'm all about that, but i think that would be solving a different problem |
| 09:39 | <roc> | user-created content |
| 09:40 | <roc> | part of the problem with IDEs is that a lot of the market is a big enterprise nightmare |
| 09:41 | <roc> | lots of stupid checkbox requirements, integration with byzantine processes, etc |
| 09:42 | <roc> | I should find out what the environments for developing for App Engine and Facebook are like |
| 09:42 | <Hixie> | hsivonen: btw i commented on the doctype idea in the logs while you were gone -- basically i think your idea is a good one and i'd suggest using "XSLT-generated" as the name |
| 09:43 | <hsivonen> | Hixie: nice. then one issue can be closed |
| 09:47 | <annevk> | (people might argue it doesn't quite close the issue with respect to <source>, <eventsource>, <command>, <embed>, etc.) |
| 09:47 | <Hixie> | well there's nothing we can do about those |
| 09:48 | <Hixie> | especially embed |
| 09:48 | <hsivonen> | Hixie: there is something you could do about the others but <embed>. it just wouldn't be elegant |
| 09:50 | <zcorpan> | what's special about embed? |
| 09:50 | <Hixie> | it's already implemented |
| 09:50 | <annevk> | actually, can't XSLT output <source /> |
| 09:50 | <annevk> | ? |
| 09:50 | <annevk> | because that would toaly be compliant |
| 09:50 | Hixie | admits not caring one iota :-) |
| 09:50 | <hsivonen> | annevk: it can, if you use XML output and you are very careful, IIRC |
| 09:51 | annevk | cares slightly about the endless whining |
| 09:51 | <zcorpan> | Hixie: you could allow an end tag that doesn't do anything |
| 09:51 | <Hixie> | wouldn't be elegant |
| 09:51 | <zcorpan> | indeed |
| 09:51 | <annevk> | and confusing as hell |
| 09:51 | <zcorpan> | indeed |
| 09:52 | <Hixie> | i find mail filters to be more effective at stopping the whining than spec changes |
| 09:52 | <zcorpan> | though, people already write <embed ...></embed> |
| 09:52 | <Hixie> | :-) |
| 09:52 | hsivonen | mumbles something about </p> being confusing |
| 09:52 | <zcorpan> | </br> anyone? |
| 09:52 | <zcorpan> | actually </p> is more confusing |
| 09:53 | <zcorpan> | with <p><table></table></p> being the winner |
| 09:53 | <zcorpan> | or <p><object></p></object> |
| 09:54 | <hsivonen> | zcorpan: that's Hixie's fault even! :-) |
| 09:54 | <hsivonen> | <p><table></table></p> that is |
| 09:55 | <Hixie> | yeah yeah |
| 09:55 | <Hixie> | :-P |
| 09:55 | <jgraham> | couldn't we use "legacy-doctype-compat" or something rather than "XSLT-compat" or "XSLT-generated" as the name. There are other tools that insist on having PUBLIC sections in the doctype |
| 09:55 | <Hixie> | i cared about standards back when i wrote acid2... |
| 09:55 | <zcorpan> | hsivonen: <p><ol></ol></p> |
| 09:55 | <Hixie> | jgraham: like which? |
| 09:55 | <jgraham> | Hixie: I think the serilization functions in libxml2 (or at least lxml), for example |
| 09:56 | <zcorpan> | actually i'm fine with table acting like ol if we can make it do so in quirks mode too |
| 09:56 | <Dashiva> | The RDF thread reaches new heights. "If the domain becomes unavailable, you just change the namespace!" |
| 09:56 | <Hixie> | hsivonen, zcorpan, Philip`: there was a thread about making unquoted attribute values containing "=" conforming. i'm thinking of going with no change (leave it non-conforming), based on the data in that thread showing that it does catch some subtle authoring mistakes. any objections? |
| 09:56 | <jgraham> | Hixie: btw I'm happy to run anolis off jame.html5.org |
| 09:56 | <Hixie> | jgraham: that would rock |
| 09:56 | <zcorpan> | Hixie: ok with me |
| 09:57 | <Philip`> | What limitations does html5.org have on long-running processes? |
| 09:57 | <annevk> | Dashiva, yeah, and then you take a third, unrelated domain, to describe the mapping |
| 09:57 | <Hixie> | jgraham: i'm not so worried about libxml2 serialisation, it doesn't have the high profile in w3c space that xslt does, and can be much more easily fixed (since it's just software, not a standard) |
| 09:57 | <jgraham> | Philip`: It's not particularly long running is it? |
| 09:57 | <Philip`> | jgraham: It takes tens of CPU-seconds just to parse the spec |
| 09:58 | <annevk> | I don't think I put limits in place... |
| 09:58 | <Hixie> | hsivonen, zcorpan, Philip`: ok i'm not making the aforementioned change. i'm also not replying to your e-mails on the subject. :-) |
| 09:58 | <Hixie> | (nobody else posted on the thread) |
| 09:58 | <Dashiva> | annevk: And you need a fourth domain to host the address of the third? |
| 09:58 | <annevk> | but I don't have DreamHost PS either |
| 09:59 | <annevk> | Dashiva, I think so, and then you google for prefix:localname to locate them all |
| 09:59 | <Philip`> | Dashiva: Have a persistent DNS service, which makes whateveryouwant.pdns.org be a CNAME to the actual domain the namespace content is hosted on, so it can be easily changed by someone other than the owner of the hosting domain |
| 09:59 | <jgraham> | Hixie: Well it seems to me that having a slightly more generic name that still emphasises legacy deals with any XSLT-like cases, thus not promoting XSLT as somehow special, but I don't care very much :) |
| 10:00 | <Dashiva> | It's like a giant game of pass-the-SPOF-buck |
| 10:00 | <annevk> | haha, jgraham is using more MB than everything I have on DreamHost combined |
| 10:00 | <zcorpan> | Hixie: nice to remove emails from your folders without changing the spec or replying to them, isn't it? :) |
| 10:01 | <Hixie> | zcorpan: :-D |
| 10:01 | <jgraham> | annevk: I have an entire custom python environment on there :) |
| 10:01 | <Hixie> | actually i'm going to reply to one of them after all |
| 10:01 | <annevk> | Philip` is slightly behind me |
| 10:01 | <Hixie> | since you proposed making <foo bar=> non-conforming, and that seems like a good idea |
| 10:01 | zcorpan | doesn't remember why he argued to make that non-conforming |
| 10:02 | <Hixie> | <foo bar=> add baz=1 becomes <foo bar= baz=1> |
| 10:02 | <annevk> | everything together it's still below 800MB out of 350GB |
| 10:02 | <zcorpan> | ah |
| 10:03 | <zcorpan> | wonder if i should set up a filter for "RDFa" |
| 10:04 | <Hixie> | doesn't work |
| 10:04 | <hsivonen> | Hixie: leaving = in unquoted attribute values non-conforming makes sense to me |
| 10:04 | <Hixie> | not enough of these e-mails mention the term rdfa |
| 10:04 | <Philip`> | Hixie: Unquoted attributes are ugly and cause errors, so I'm quite happy for any subset of them to be non-conforming :-) |
| 10:04 | <Hixie> | hsivonen: yeah, it was your e-mail arguing that that convinced me :-) |
| 10:04 | <zcorpan> | Hixie: they do in the subject line |
| 10:05 | <zcorpan> | hmm not all though |
| 10:05 | <zcorpan> | i guess holding down space is good enough for me |
| 10:06 | <Hixie> | gmail's "m" feature would come in handy for this kind of thing if i used gmail |
| 10:06 | <Dashiva> | At the moment, it's more like "every thread on whatwg that isn't about workers" |
| 10:06 | <zcorpan> | yeah |
| 10:06 | <zcorpan> | and i don't really care about workers either |
| 10:06 | <zcorpan> | :P |
| 10:06 | <annevk> | hah, html5.org has twice as much traffic as annevankesteren.nl in terms of bandwidth |
| 10:07 | <hsivonen> | Hixie: good to know I'm consistent enough that I don't need to recap what email I wrote :-) |
| 10:07 | <Hixie> | :-) |
| 10:07 | <Philip`> | annevk: Looks like a fifth of my space usage is logs, and half of the rest is canvas tests, and the rest is random stuff |
| 10:07 | <zcorpan> | annevk: from web-apps-tracker? |
| 10:07 | <Philip`> | (and it's only ~100MB in total) |
| 10:07 | <annevk> | looking into actual stats now |
| 10:07 | <Hixie> | oh hey |
| 10:07 | <Hixie> | anne |
| 10:07 | <Hixie> | while you're at it |
| 10:07 | <Hixie> | any chance you can set up a workers-tracker? |
| 10:07 | <annevk> | lately it has been 10.000 requests a day |
| 10:08 | <Hixie> | 10000?! or 10? |
| 10:08 | <annevk> | 10000 |
| 10:08 | <Hixie> | wow |
| 10:08 | <Hixie> | oh wait do you have test cases on there that go through spartan? |
| 10:08 | <annevk> | no |
| 10:08 | <jgraham> | For everything? |
| 10:08 | <annevk> | yes, but excluding subdomains |
| 10:08 | <Hixie> | all spiders then? :-) |
| 10:10 | <Philip`> | wc -l for philip.html5.org's logs says: 6270 access.log.2008-08-27 7969 access.log.2008-08-28 |
| 10:10 | <Philip`> | Uh, with newlines in appropriate places |
| 10:10 | <annevk> | most are actually browsers |
| 10:10 | <Philip`> | (Why does irssi keep stripping them out?) |
| 10:10 | <annevk> | with IE leading?! |
| 10:10 | <annevk> | lol, IE5 is leading, something is on crack |
| 10:10 | <zcorpan> | annevk: a bot masking as ie5? |
| 10:11 | <annevk> | I guess |
| 10:12 | <zcorpan> | i learned the other day that aria-hidden is semantically empty as implemented |
| 10:12 | <Hixie> | isn't all of aria semantically empty by definition? |
| 10:12 | <Philip`> | I see a couple of hundred requests from IE5 a day, almost all to http://philip.html5.org/tools/parser/ |
| 10:13 | <Philip`> | which doesn't make much sense |
| 10:13 | <Philip`> | (Must be bots since there's no requests to that page's scripts) |
| 10:13 | <Hixie> | aria is the equivalent of presentational attributes for accessibility apis, no? |
| 10:13 | <zcorpan> | Hixie: i mean, the presence of aria-hidden doesn't do anything *at all* |
| 10:13 | <Hixie> | oh |
| 10:13 | <Hixie> | why not? |
| 10:13 | <annevk> | then they might as well drop it in favor of hidden |
| 10:14 | <hsivonen> | zcorpan: doesn't aria-hidden=true prune the accessible tree? |
| 10:14 | <zcorpan> | firefox just looks at what is being rendered |
| 10:14 | <zcorpan> | if it's display:none, it's dropped from the accessible tree |
| 10:14 | <zcorpan> | "aria-hidden" isn't checked for in the code |
| 10:14 | <hsivonen> | wow |
| 10:14 | <zcorpan> | aria-hidden is for firevox basically |
| 10:15 | <Hixie> | i don't understand why aria is so popular |
| 10:15 | <Philip`> | A bug / missing feature in an early implementation of a new technology? It's terrible :-( |
| 10:15 | <Hixie> | it's a bandaid |
| 10:16 | <annevk> | some accessibility people seem to love bolton stuff |
| 10:16 | <Philip`> | There's probably a billion-dollar industry in bandaids |
| 10:16 | <jgraham> | Hixie: There seems to be something of a culture of liking bandaids because they are nicely visible |
| 10:16 | <annevk> | Philip`, consultancy certainly benefits from it |
| 10:17 | <jgraham> | Whereas just fixing stuff properly doen't look like you're trying so much |
| 10:17 | <annevk> | jgraham, so you can pat yourself on the back? |
| 10:17 | <zcorpan> | for anyone interested in aria (or at least the moz ua impl guide document), there's a new googlegroup: aria-ua-impl |
| 10:18 | <jgraham> | annevk: Well I don' know about the authors point of view I guess but that seems to be the approach of a number of a11y types |
| 10:18 | <Philip`> | Hmph, I can't find any data on the size of the bandaid industry, since searches keep turning up bandaid metaphors instead |
| 10:18 | <annevk> | so the issue with workers-tracker is that in order to do it properly you actually want to abstract out the tracker stuff and just import it or something |
| 10:18 | <zcorpan> | also: http://monotonous.org/2008/08/25/lets-all-do-it-right-the-first-time/ should be nice for testing aria |
| 10:18 | <zcorpan> | haven't had time to look into it properly yet though |
| 10:19 | jgraham | should go do something useful |
| 10:20 | <hsivonen> | is there a simple test case that shows if IE8 is in IE7 standards mode or in IE8 mode? |
| 10:20 | <annevk> | "Let’s all do it right the first time" and ARIA in one post is sort of ironic :) |
| 10:21 | <Hixie> | in the last checkin comment, i meant DOS when i wrote DOM. before anyone points it out. |
| 10:21 | <annevk> | (how do you paste in a terminal?) |
| 10:21 | <Hixie> | annevk: :-/ |
| 10:22 | <Hixie> | annevk: unix? |
| 10:22 | <Hixie> | annevk: middle click pastes selection if you're in X Windows |
| 10:22 | <annevk> | Ubuntu? :) |
| 10:22 | annevk | is such a noob |
| 10:23 | <hsivonen> | IIRC, Gnome Terminal has a context menu |
| 10:24 | <Hixie> | ew |
| 10:24 | <Hixie> | gnome terminal |
| 10:24 | <Philip`> | hsivonen: document.compatMode |
| 10:24 | <Hixie> | who the hell is 64.111.120.16 and why are they DOSsing me |
| 10:25 | <Hixie> | alameda.dreamhost.com |
| 10:25 | <Hixie> | is that one of you :-P |
| 10:25 | <zcorpan> | :roll: |
| 10:25 | <zcorpan> | j/k |
| 10:25 | <Hixie> | oh it's stopped now |
| 10:25 | <Philip`> | That's where html5.org is |
| 10:26 | <Hixie> | thought it might be :-) |
| 10:26 | <Philip`> | But I don't run any interesting scripts there so it's not my fault :-) |
| 10:26 | <Hixie> | a lot of hits to svn.whatwg.org, made the load on hixie go up to 24 |
| 10:26 | <zcorpan> | perhaps it's the hidden feature of webapps-tracker? |
| 10:26 | <Hixie> | i blame anne :-P |
| 10:29 | <annevk> | hsivonen, it does, but using that is annoying |
| 10:30 | Philip` | always uses shift+insert to paste |
| 10:31 | hsivonen | exhausted a hard drive while storing validation messages |
| 10:31 | Philip` | always turns off middle-click paste when possible, because it's far too likely he'll click accidentally and send random clipboard content to random people |
| 10:32 | <hsivonen> | I guess I should have used shorter markers. like Q instead of QUIRKS |
| 10:33 | <Hixie> | heh |
| 10:33 | <hsivonen> | to fit more data onto the disk |
| 10:33 | <Hixie> | Philip`: i do the same with the clipboard in general |
| 10:33 | <Philip`> | If you're compressing, that shouldn't make a difference |
| 10:33 | <Philip`> | s/a/much/ |
| 10:33 | <hsivonen> | or perhaps I should have omitted the file system |
| 10:34 | <Philip`> | Perhaps you should have done a small sample and measured how much space it used, and extrapolated that to the whole dataset, rather than doing as much as possible until you discover you've run out of space? :-) |
| 10:34 | <Hixie> | hah |
| 10:35 | <Hixie> | that's a pessimistic approach |
| 10:35 | <hsivonen> | Philip`: the dataset *almost* fit on the drive |
| 10:35 | <hsivonen> | missed last 6000 pages or so |
| 10:35 | <Hixie> | a better approach would have been to use a drobo, and as the capacity was reached, rushed to the store and slammed in a bigger capacity drive at the last minute :-) |
| 10:35 | <Philip`> | hsivonen: Did it get that close through planning, or through luck? :-) |
| 10:36 | <hsivonen> | Philip`: luck |
| 10:36 | <hsivonen> | Philip`: my planning said 60 GB |
| 10:36 | <hsivonen> | reality was closer to 150 GB |
| 10:56 | <Philip`> | Hixie: I think the text-coalescence change looks kind of reasonable (though I reserve the right to change my mind in the future) |
| 10:56 | <Hixie> | k |
| 10:56 | <Hixie> | that was pretty much my reaction |
| 10:57 | <annevk> | http://html5.org/tools/web-forms-tracker and http://html5.org/tools/web-workers-tracker now work |
| 10:57 | <Hixie> | woo |
| 10:57 | <Hixie> | thanks! |
| 10:58 | <annevk> | it was apparently quite easy to extract the functions into a trackerlib file and then invoke one function with a few arguments from the individual files |
| 10:58 | <Philip`> | I think it means a</li>b</li>c... will still be one text node, which is possible to optimise by being adequately clever, and a<table>b<td></td>c<td></td>d... will now be split into lots of text nodes, which is good since I don't think it was possible to optimise when you could have scripts inside the <td>s |
| 10:58 | <Hixie> | yeah the examples you gave in your original e-mail all work as you wanted them to definitely |
| 10:58 | <Hixie> | i'm just worried that there's some edge case where i have it defined in some ass-backwards way that conflicts with this |
| 10:59 | <annevk> | I should probably commit this back to the html5 SVN now |
| 10:59 | <Hixie> | is namespace-well-formed a subset of well-formed, or an orthogonal concept? hmm |
| 11:01 | <annevk> | done |
| 11:08 | <annevk> | shirt+insert works Philip`, but they are rather far apart |
| 11:08 | <annevk> | (on my keyboard) |
| 11:09 | <annevk> | prolly better than rightclicking and all though |
| 11:15 | <hsivonen> | Philip`: in IE8, what values for document.compatMode should I look for? |
| 11:17 | <Philip`> | hsivonen: Whoops |
| 11:17 | <Philip`> | I meant document.documentMode |
| 11:17 | <hsivonen> | Philip`: thanks |
| 11:17 | <Philip`> | which is 8 in IE8-mode |
| 11:17 | <Philip`> | and 5 in quirks mode |
| 11:18 | <hsivonen> | Philip`: thanks |
| 11:19 | <Philip`> | and 7 in IE7-mode |
| 11:19 | <Hixie> | hsivonen: i think i found a bug in the live html5 dom viewer (and in the spec) |
| 11:19 | <Hixie> | http://livedom.validator.nu/?%3C!DOCTYPE%20html%3E%0A...%3Cdiv%3E%3Coption%3E%3C%2Fdiv%3Ex |
| 11:19 | <Philip`> | I don't think it can ever be 6 |
| 11:19 | <Hixie> | <option> is not defined as phrasing or formatting |
| 11:19 | <Hixie> | so the </div> shouldn't see the <div> |
| 11:20 | <Philip`> | http://msdn.microsoft.com/en-us/library/cc196988(VS.85).aspx but don't believe where it says "In Internet Explorer 8 and later, the compatibility mode of the document is determined wholly by the X-UA-Compatible meta tag." because it's lying |
| 11:20 | <annevk> | it might help to rename the title of http://livedom.validator.nu/ |
| 11:21 | <annevk> | it's now identical to the service hixie provides while the two are clearly distinct |
| 11:21 | <Hixie> | weird, html5lib does the same thing |
| 11:21 | <Hixie> | am i misreading the spec or something? |
| 11:21 | <Hixie> | http://parsetree.validator.nu/?parser=html5&content=%3C%21DOCTYPE+html%3E%0D%0A...%3Cdiv%3E%3Coption%3Ez%3C%2Fdiv%3Ex&submit=Print+Tree |
| 11:21 | <annevk> | wasn't <option> automatically closed? |
| 11:22 | <annevk> | see implied end tags |
| 11:22 | <Hixie> | why would an end tag be implied? |
| 11:22 | <Hixie> | it wouldn't be if you replaced the </div> with </foo> |
| 11:22 | <Hixie> | and you should bail before you find out that you know about a <div> |
| 11:22 | <Hixie> | because <option> isn't phrasing or formatting |
| 11:23 | <annevk> | huh? |
| 11:23 | <Hixie> | (which is the bug in the spec, and will be changed shortly) |
| 11:23 | <annevk> | when you see the end tag </div> |
| 11:23 | <annevk> | you generate implied end tags |
| 11:23 | <Hixie> | why? |
| 11:23 | <annevk> | because that's what the spec says |
| 11:23 | <Hixie> | ohhh |
| 11:23 | <Hixie> | my bad |
| 11:23 | <Hixie> | i'm being dumb |
| 11:23 | <hsivonen> | Hixie: so I'm implementing the spec, right? then it's not a live dom bug :-) |
| 11:24 | <hsivonen> | annevk: I'll change the title |
| 11:24 | <Hixie> | yeah |
| 11:24 | <Hixie> | my bad |
| 11:24 | <Hixie> | forgot </div> had its own entry, duh |
| 11:24 | <Hixie> | anyway |
| 11:24 | <hsivonen> | annevk: title fixed |
| 11:25 | <Hixie> | <optgroup> and <option> are becoming phrasing-level |
| 11:25 | <hsivonen> | Hixie: why? |
| 11:25 | <Hixie> | so that <datalist><option>x</datalist>y works |
| 11:26 | <Hixie> | (also <option> is getting its own entry in the "in body" state that implies </option>) |
| 11:26 | <Hixie> | (so that <datalist><option>x<option>y</datalist> works) |
| 11:26 | <hsivonen> | did the chairs decide that datalist is in? |
| 11:27 | <Hixie> | i've been told to go ahead and merge wf2 in, yes |
| 11:27 | <hsivonen> | ok |
| 11:27 | <Hixie> | but even if that hadn't happened, we'd still want to fix the parsing model |
| 11:27 | <Hixie> | so that we have the option later |
| 11:28 | <Philip`> | It's odd that IE8 devotes equal screen space to the "render with IE7 bug compatibility" button as it does to "refresh page" |
| 11:28 | <Philip`> | Prediction: People will click the compatibility button, see that it causes the page to reload, and deduce that that must be the refresh button |
| 11:29 | <Philip`> | and so they'll unknowingly toggle between IE7-bugs mode and IE8-bugs mode each time they refresh a page |
| 11:29 | <Hixie> | i guess <optgroup> should also imply </option> |
| 11:30 | <Hixie> | just in case we ever introduce <optgroup> to something |
| 11:30 | <Hixie> | like datalist |
| 11:34 | <Hixie> | the html5 live dom viewer rocks. |
| 11:40 | <annevk> | how the hell can the headers issue be resolved? |
| 11:41 | <hsivonen> | hmm. did the HTML WG make a decision on a telecon? |
| 11:41 | <hsivonen> | fwiw, I'd have resolved it by letting td[headers] to point to th and letting th[headers] to point to another th |
| 11:42 | <annevk> | for people who care so much about process they are doing a bad job at following the charter |
| 11:42 | <hsivonen> | pointing it to td seems like semantic abuse to me |
| 11:42 | <Philip`> | It sounded like a decision in the "agreement between some group of people" sense, not the "formal binding WG decision" sense |
| 11:42 | <hsivonen> | and to make a migration from bolt-on to built-in, authors should get in the habit of using <th> where appropriate |
| 11:44 | <annevk> | Philip`, not the way steven announced it |
| 11:44 | <Lachy> | Hixie, in case you missed this bug I noticed yesterday, see http://krijnhoetmer.nl/irc-logs/whatwg/20080828#l-569 |
| 11:45 | <annevk> | fwiw: http://lists.w3.org/Archives/Public/www-archive/2008Aug/0129.html |
| 11:45 | <Hixie> | Lachy: i think it was just that the server side of the script failed to update the file with the latest version in it |
| 11:45 | <Lachy> | ok |
| 11:47 | <annevk> | Hixie, <optgroup> should close <optgroup> |
| 11:47 | <Hixie> | that would prevent us from ever nesting optgroups |
| 11:48 | <annevk> | <select> already prevents that |
| 11:48 | <Hixie> | this is for outside select |
| 11:48 | <annevk> | do we want them to be inconsistent? hmm |
| 11:48 | <Hixie> | <optgroup> is currently never allowed outside <select> |
| 11:48 | <Hixie> | <datalist> only accepts <option>s |
| 11:49 | <Hixie> | iirc |
| 11:49 | <Hixie> | i just want to leave the door open, just in case |
| 11:49 | <annevk> | okay |
| 11:50 | <Philip`> | annevk: Ah, I meant during the telecon - I agree his email sounds more like it was a Process-style decision |
| 11:56 | <annevk> | ah, Mark Birbeck replied to http://www.ibm.com/developerworks/blogs/page/JohnBoyer?entry=the_ubiquity_strategy_for_promoting |
| 11:57 | <annevk> | seems to me people actually want new functionality from browsers, eg Offline Web Applications, multi-core CPU usage, etc. |
| 12:05 | zcorpan | updates html5-elements with bb and hidden |
| 12:05 | <zcorpan> | is there something else that has changed or been added recently? |
| 12:06 | <zcorpan> | where recently is after event-source was renamed |
| 12:07 | <annevk> | hidden should be before id |
| 12:08 | <Hixie> | <action> sounds like <command> |
| 12:08 | <Hixie> | and <form action> |
| 12:08 | <annevk> | yeah, <browserbutton> |
| 12:08 | <zcorpan> | annevk: thanks |
| 12:11 | <Hixie> | hsivonen: <!DOCTYPE html PUBLIC ""> fails the "make it look worse" test to avoid cargo cultists thinking more is more |
| 12:11 | <annevk> | does anyone recall when I tried to avoid accepting what the HTML charter said? |
| 12:11 | <annevk> | was that ever the case? |
| 12:11 | <Hixie> | has anyone ever suggested ignoring the charter? |
| 12:12 | <annevk> | stevef just told me "btw It is good to see you recognising the HTML WG charter in this case." |
| 12:12 | <Hixie> | o_O |
| 12:12 | <annevk> | yeah, wtf |
| 12:16 | <Hixie> | i wonder when people will come to realise that when an e-mail just says "+1" or "I agree" i delete it without even seeing who wrote it |
| 12:16 | <Hixie> | and never actually take that e-mail into account when making the related edits |
| 12:16 | <annevk> | http://www.devx.com/webdev/Article/38983 cool |
| 12:18 | <annevk> | hmm, code doesn't do much |
| 12:21 | <annevk> | (moving is s-x and j-l) |
| 12:23 | <hsivonen> | Hixie: At least I mentioned testing browsers instead of just saying that I agree. |
| 12:24 | <hsivonen> | I wonder what it would take to make a compiler that takes Flex stuff and outputs SVG+JS instead of Flash |
| 12:25 | <Hixie> | hsivonen: your e-mail ended up in the pile of e-mails that i will consider |
| 12:25 | <Hixie> | because it had actual information in it |
| 12:25 | <Hixie> | like almost all your e-mails :-) |
| 12:25 | <Hixie> | btw, i don't understand what ui you would have for hierarchical headers |
| 12:26 | <Hixie> | just have the header cell itself read out its own headers? or propagate them to the data cells too? |
| 12:26 | <hsivonen> | both, I suppose |
| 12:26 | <Hixie> | the latter is pretty annoying |
| 12:27 | <Hixie> | especially for the kind of tables where you see it |
| 12:27 | <jgraham> | Hixie: I would like to see user tests to prove that |
| 12:27 | <annevk> | incidentally, the minutes for that particular discussion are a mess, with Lachy asking several questions not being answered and Chris Wilson stating that people should show up to telcons if they had wanted to advance counter arguments... |
| 12:27 | <jgraham> | But the a11y people hate me for saying that |
| 12:27 | <Hixie> | jgraham: yeah, user tests would help a lot here |
| 12:27 | <jgraham> | Shall I raise the process issue with SF's email? |
| 12:28 | <annevk> | jgraham, I raised something on www-archive |
| 12:28 | <annevk> | jgraham, not sure if more is required, but maybe |
| 12:28 | <jgraham> | annevk: OK, well I will do it then |
| 12:32 | <Hixie> | svn blame on the html5 spec takes an ungodly amount of time |
| 12:32 | <hsivonen> | even though you are to blame for each check-in |
| 12:32 | <Dashiva> | Doesn't every line say hixie? |
| 12:33 | <Hixie> | Dashiva: yeah, trying to pin down which rev i added <rt> and <rp> to the "optional tags" section since i have a note here saying i have to do something with them, but i can't tell if the something is to add them (and the note predates the add) or to fix them (and the note postdates the add) |
| 12:33 | <Dashiva> | How about a binary reach instead? :) |
| 12:33 | <Dashiva> | *search |
| 12:33 | <Lachy> | annevk, the minutes for telcons are always a mess and freqently omit information |
| 12:34 | <Hixie> | Dashiva: would prolly have been quicker |
| 12:35 | <Hixie> | Lachy: that's one reason i avoid telecons |
| 12:35 | <Lachy> | I would have avoided it, but annevk asked me to call in for him, so I did |
| 12:35 | <Hixie> | heh |
| 12:35 | <Hixie> | why would he ask you something like that |
| 12:35 | <Hixie> | that sounds like work! |
| 12:36 | <Dashiva> | Telcons sound kinda like the spec equivalent of 5-second soundbites |
| 12:37 | <Lachy> | Hixie, it's fair enough. Since I don't do any other work, 30 minutes on a telcon isn't too bad :-) |
| 12:37 | <Hixie> | hah |
| 12:38 | <Lachy> | Hixie, what can I say to convince you that you're wrong about not allowing <area alt="..."> without href=""? http://lists.w3.org/Archives/Public/public-html/2008Aug/0860.html |
| 12:40 | <Hixie> | you can show how that text would be used |
| 12:40 | <Lachy> | since you rejected the valid use case I presented, finding more evidence means more work. |
| 12:40 | <Hixie> | right now the spec says it must be ignored |
| 12:40 | <annevk> | Hixie, web-apps-tracker takes a limit parameter which you can set to -1 or 1000 or something |
| 12:40 | <Hixie> | and we generally want conformance checkers to report things that aren't useful |
| 12:40 | <annevk> | Hixie, that's probably easier than using svn blame |
| 12:40 | <Hixie> | annevk: quicker, certainly |
| 12:40 | <Hixie> | annevk: easier, no :-) |
| 12:40 | <annevk> | ok |
| 12:40 | <Hixie> | annevk: if blame was quick, i could just look up the number and be done with it :-) |
| 12:41 | <hsivonen> | Hixie: I hope you missed a "don't" in the statement about conformance checkers above |
| 12:42 | <Hixie> | i mean, report things in the markup that aren't going to do anything |
| 12:42 | <Hixie> | since they are likely an indication of an error |
| 12:42 | <Hixie> | since authors don't generally go out of their way to do something that isn't goign to do anything |
| 12:42 | <hsivonen> | actually, they do it all the time |
| 12:42 | <hsivonen> | <style type=text/css> |
| 12:42 | <Hixie> | let me rephrase |
| 12:43 | <hsivonen> | type=text/css is a no-op |
| 12:43 | <Hixie> | authors don't intentionally and knowingly do it |
| 12:43 | <Lachy> | Hixie, using <area> to annotate regions of images with additional info, even if they aren't links, seems like a perfectly reasonable thing to do. I guess I will have to look at some more real world examples of image maps |
| 12:44 | <Hixie> | Lachy: an image map is a list of links. using it for something else is an abuse of the semantics. |
| 12:44 | <Hixie> | Lachy: now if we want to handle those other use cases, maybe we should repurpose <area> in some way, e.g. with some element other than <map> |
| 12:44 | <Hixie> | Lachy: or with some new ua-level requirements |
| 12:44 | <Hixie> | Lachy: but we'd have to justify doing so |
| 12:55 | <Lachy> | Hixie, HTML4 defined an image map as: "Image maps allow authors to specify regions of an image or object and assign a specific action to each region" |
| 12:55 | <Lachy> | The action could be an event handler like onclick instead of just a link |
| 12:56 | <Hixie> | then alt="" would be allowed with onclick or href="" |
| 12:56 | <Hixie> | we could do that i guess |
| 12:56 | <Hixie> | seems a bit weird |
| 12:57 | <Hixie> | we don't make <a onclick=""> into a link |
| 12:58 | <Lachy> | the event handler could be attached dynamically, so there wouldn't need to be onclick in the markup |
| 12:58 | <hsivonen> | 'null' is null or not an object |
| 12:59 | <hsivonen> | says IE8 |
| 12:59 | <hsivonen> | if I had a stringified 'null', it would be an object, wouldn't it? |
| 13:00 | <Lachy> | hsivonen, for what code does it say that? |
| 13:00 | <Dashiva> | It would act as one, at least |
| 13:00 | <hsivonen> | Lachy: the script on validator.nu |
| 13:01 | <Lachy> | grrr. Why does IE8 make the error dialog modal?! |
| 13:01 | <Dashiva> | hsivonen: You're living pretty dangerously with all those implied semicolons |
| 13:02 | <hsivonen> | Dashiva: what's a language feature good for if living with it is dangerous? |
| 13:03 | <annevk> | living on the edge can give you a kick? |
| 13:04 | <Dashiva> | One could imagine JS was made for small scripts with small chances of breaking horribly :) |
| 13:04 | <Lachy> | I never remember when it's ok to omit semicolons, and it always seems to complain when I do. So I do it anyway |
| 13:04 | <Lachy> | I mean, I always use them |
| 13:05 | <Lachy> | it seems to be complaining about this line: newStyle.styleSheet.cssText = rule |
| 13:05 | <Dashiva> | Lachy: I'm pretty sure you're leaving some out, you just don't know it :) |
| 13:05 | <Lachy> | so it means either newStyle or newStyle.styleSheet is returning null |
| 13:07 | <Lachy> | Dashiva, if I do, it's only by accident; But I do always try to end every line with a semicolon; |
| 13:07 | <hsivonen> | boohoo. the data I recorded is corrupt. It's corrupt in such a crazy way that I suspect gzipping gone wrong at some point |
| 13:08 | Philip` | wonders if "some point" was 4GB in |
| 13:09 | <annevk> | heh, Opera shipped with APNG support before Firefox; I did not know |
| 13:09 | <hsivonen> | let's find out |
| 13:09 | <Philip`> | annevk: Only by a few days :-p |
| 13:10 | <Philip`> | and since there are approximately zero APNG images in the world, it doesn't matter much in practice |
| 13:12 | <Philip`> | Opera renders some perfectly valid APNGs incorrectly on some platforms anyway |
| 13:13 | <Philip`> | whereas Firefox seems to work fine, except in some obscure cases like 8-bit greyscale or 1-bit APNGs |
| 13:13 | <Hixie> | wow, IE8 finally adds the inline find toolbar, and they manage to make the one mistake that every other browser has agonised over finding a way to avoid |
| 13:13 | <Hixie> | they move the whole content down when you hit find |
| 13:13 | <Hixie> | ok with that, bed time |
| 13:13 | <Hixie> | nn |
| 13:16 | <hsivonen> | Philip`: is there a known problem with GZIPOutputStream with stream longer than 4 GB? |
| 13:18 | <Philip`> | hsivonen: None that I know of; I'm just assuming it's the most obvious point of failure when dealing with unusually large files |
| 13:18 | <hsivonen> | quick googling shows problems with .zip files over 4GB and java class lib |
| 13:18 | <hsivonen> | nothing about gzip, though |
| 13:20 | <Philip`> | It's also possible that the compression is fine but the decompression isn't working |
| 13:20 | <hsivonen> | good point |
| 13:22 | <hsivonen> | could be the python side http://bugs.python.org/issue1672853 |
| 13:23 | <hsivonen> | one would think that these were solved problems by now |
| 13:23 | <hsivonen> | all handled in the infrastructure |
| 13:25 | <hsivonen> | interestingly, I can't use zless to traverse the whole file. it complains about mmap failing at some point |
| 13:25 | <Philip`> | Use 'gzip -cd | less' ? |
| 13:26 | <hsivonen> | isn't that what zless is? |
| 13:26 | <Philip`> | I would assume zless has to do effectively that anyway, since I didn't think you could decompress zlib without starting from the beginning |
| 13:27 | <Philip`> | I would assume zless is a program that does some stuff itself, so it'd have different bugs to gzip/less |
| 13:28 | <hsivonen> | Philip`: zless is a shell script that runs gzip and less |
| 13:28 | <Philip`> | Oh, okay |
| 13:29 | <hsivonen> | anyway, I made a copy of the first 4GB with dd. now trying if that part makes Python fail |
| 13:31 | <hsivonen> | how can this be so hard? |
| 13:33 | <Philip`> | Alternatively: How can it be so easy for you to manipulate a trillion bits of data in a few hours from the comfort of your chair? :-) |
| 13:34 | <Philip`> | It's not the kind of thing you could contemplate doing five or ten years ago |
| 13:34 | <hsivonen> | well, the failure happens before 4 GB of compressed data |
| 13:34 | <hsivonen> | dunno what the uncompressed counter was |
| 13:37 | <Philip`> | Easy to find out - just do 32 bisections :-) |
| 13:48 | <annevk> | hmm, I need to fix cookies for the various trackers |
| 13:55 | <annevk> | fixed |
| 13:59 | <hsivonen> | the decompressed data is at 5.9 GB when the failure happens |
| 13:59 | <hsivonen> | weird |
| 14:03 | <Philip`> | Does anyone happen to have a copy of IE8b2's mshtml.dll they could send me? (I want to look at some stuff in it, but left my copy at home) |
| 14:21 | Philip` | sees that http://blogs.msdn.com/ie/archive/2008/08/27/internet-explorer-8-beta-2-now-available.aspx has at least three requests for <canvas> support, and can't see any for any other HTML5 features |
| 14:30 | <Philip`> | gsnedders: Are you aware that http://google.com/search?q=cache:simplepie.org/ is still full of spam? |
| 14:32 | <Philip`> | (and now it has a "Buy Phentermine" link - when I last looked, I think all these hacked pages only had links to each other, so maybe that was just in preparation for inserting the actual target link) |
| 14:33 | <virtuelv> | speaking of spam, is anyone able to get hold of Blake Ross? Turn off JS and look at http://www.blakeross.com/ |
| 14:39 | <Philip`> | shepazu: http://google.com/search?q=cache:http://schepers.cc still has the same spam too |
| 14:40 | <shepazu> | Philip`: what should I do about this? |
| 14:40 | <Philip`> | shepazu: I have no idea |
| 14:40 | <shepazu> | the real page seems fine... |
| 14:41 | <Philip`> | Live Search and Yahoo have the same spam in their cached versions |
| 14:41 | shepazu | decides not to care :) |
| 14:41 | <Philip`> | (I assume it's only sending spam to certain IPs or something) |
| 14:41 | <annevk> | check the source code of the PHP pages? |
| 14:42 | <Philip`> | shepazu: I suppose that's the easiest solution :-p |
| 14:43 | <svl> | upgrade to the latest wordpress version? (which is 2.6.1 rather than 2.5.1) |
| 14:44 | svl | assumes there's a vulnerability in 2.5.1 which is being exploited to have the spam inserted. |
| 14:44 | <shepazu> | bbiab |
| 14:46 | <Philip`> | http://philip.html5.org/misc/spammy-sites.txt is my notes from when I last looked at this - it seemed to be specific to Dreamhost |
| 14:52 | <csarven> | Is krijnh human or bot? |
| 14:52 | <svl> | human |
| 14:53 | <svl> | Though not able to access IRC atm. |
| 14:54 | <csarven> | Do you know if I /msg him (or her), they'll be able to check their logs when they return? |
| 14:55 | <svl> | He will be, yes. And he's also reading log files, so even saying it here in the channel should reach him. *waves to krijn* ;) |
| 14:56 | <csarven> | Oh alright then. In that case, I've already done that yesterday on #microformats :) Thanks. Figured he was more of a regular here. |
| 14:58 | <svl> | Not quite certain if he'd have read it in there, but now that the pointer exists here, he should. |
| 14:59 | <csarven> | krijnh re: http://krijnhoetmer.nl/irc-logs/microformats/20080829#l-66 :) |
| 15:01 | <annevk> | _a should probably be data-author or something |
| 15:01 | <annevk> | and <abbr> should be <time> |
| 15:02 | <annevk> | though it seems like a lot of extra bandwidth |
| 15:02 | <csarven> | _a was already in place. I wanted to add minimal changes and add hCard and hAtom |
| 15:02 | <csarven> | I don't know why there is _a and if it is required. |
| 15:03 | <Philip`> | It's there so you can make your own lines get highlighted, I think |
| 15:03 | <csarven> | You mean flagging? |
| 15:03 | <Philip`> | (via the "who are you?" box in http://krijnhoetmer.nl/irc-logs/ ) |
| 15:04 | <csarven> | That's what the original <span> </span> is supposed to do |
| 15:04 | <Philip`> | Not the flagging - you can do client-side nick-based styling |
| 16:00 | <Lachy> | http://www.theregister.co.uk/2008/08/29/hakon_lie_ie8_interoperability/ |
| 16:03 | <Lachy> | since a large majority of intranet pages are built for IE only, it's understandable would make IE7 mode the default for intranets. Though, it is sad because companies that make intranet pages could still continue targetting IE7 mode |
| 16:06 | <annevk> | they would, unless they use some MS-specific workaround |
| 16:09 | <Philip`> | hsivonen: Is there a way I can use something like getElementById on the Document returned by your parser, that doesn't appear to just return null all the time? |
| 16:11 | <hsivonen> | Philip`: are you using the DOM in Java? |
| 16:12 | <Philip`> | hsivonen: Yes |
| 16:12 | <hsivonen> | Philip`: does DOM have a public API for setting IDness? |
| 16:14 | <Philip`> | hsivonen: Uh... |
| 16:15 | <hsivonen> | Philip`: meanwhile, you can grab this code http://pastebin.ca/1188548 |
| 16:15 | <hsivonen> | ooh. Element.setIdAttribute() |
| 16:16 | <Philip`> | It says "The DOM implementation is expected to use the attribute Attr.isId to determine if an attribute is of type ID." |
| 16:17 | <Philip`> | hsivonen: Aha, that code is handy and seems to work - thanks! |
| 16:32 | <annevk> | hsivonen, we want to remove that crap |
| 16:32 | <annevk> | (from the DOM) |
| 16:33 | <annevk> | (though that shouldn't affect Java I suppose) |
| 16:34 | <hsivonen> | Philip`: fixed on trunk |
| 16:34 | <hsivonen> | Philip`: thanks |
| 16:37 | Philip` | sees 3 pages that have @headers pointing at td, and 18 pointing at th |
| 16:37 | <Philip`> | and 16 that have totally bogus @headers |
| 16:38 | <Philip`> | (headers="24", headers="Symbol information: ", headers="100%", etc) |
| 16:39 | <Philip`> | Only one has <th headers>, vs 35 with <td headers> |
| 16:39 | <Philip`> | (out of ~130K pages) |
| 16:58 | <gsnedders> | Philip`: I see no spam there |
| 16:59 | <Philip`> | gsnedders: Oh |
| 16:59 | <Philip`> | gsnedders: I do |
| 17:00 | <gsnedders> | I do in the text only version, actally |
| 17:00 | <Philip`> | e.g. it says "online pharmacy" just before the "Newer. Better. Faster." button |
| 17:00 | <Philip`> | and then it says "I must be only to be vizier. "Holla, you can't unclog me", winged the excise. "I central nervous system, spongy degeneration of the delicately nib", he recognized." |
| 17:00 | <gsnedders> | Philip`: How are you finding this? :P |
| 17:01 | <Philip`> | gsnedders: I'm looking at the Google cache page, and then reading it with my eyes |
| 17:01 | <gsnedders> | Philip`: Yes, but why are you looking at the cache page anyway? :P |
| 17:01 | <Philip`> | gsnedders: Because I was wondering whether it had been fixed in the several months since you first noticed the problem |
| 17:01 | <gsnedders> | Philip`: It was |
| 17:02 | <Philip`> | Oh, okay |
| 17:02 | <Philip`> | If so, it broke again :-p |
| 17:02 | <Philip`> | (How was it fixed?) |
| 17:02 | gsnedders | emails the guy from Yahoo! Search (who originally found it) and Ryan Parman (who deals with the website mainly) about it again |
| 17:03 | <gsnedders> | "I still don't know how they got in, but the guys at Dreamhost found an obscure file that was being added through a couple layers of includes. In our case, we were using the PHP4-compatible nusoap class for our search results (accessing Google's SOAP API that isn't even supported anymore, but I hadn't updated the code yet). |
| 17:03 | <gsnedders> | The offending code has been removed and the search box now searches Google directly. |
| 17:03 | <gsnedders> | " |
| 17:04 | <Philip`> | If the guys at Dreamhost care, they could try following all the spam links on that page, since I expect they're all pointing to similarly-hacked Dreamhost sites |
| 17:06 | <krijnh> | *waves back at svl* |
| 17:06 | <krijnh> | csarven: I'll take a look at it, gotta cook some diner now :) (As you see, I'm not a bot) |
| 17:07 | <gsnedders> | Philip`: Oh, Ryan's email went on to say that Dreamhost basically didn't care |
| 17:07 | <Philip`> | krijnh: You might just be a particularly clever bot |
| 17:08 | <svl> | Round 1 of the Turing Test: *ding* |
| 17:08 | <krijnh> | Philip`: Thank you! In a way :) |
| 17:10 | <Philip`> | krijnh: It wasn't necessarily a compliment, given how dumb most existing bots are ;-) |
| 17:11 | <krijnh> | Hence the "in a way" |
| 17:14 | <csarven> | krijnh Awesome :) |
| 17:21 | <gsnedders> | Hixie: I'd probably rather you didn't use a web service for the same reason you don't want to run it yourself: the load :P |
| 17:22 | <Philip`> | But web services don't cause any more load than non-web services, and the service has to be available via some mechanism, so why not via the web? |
| 17:22 | <gsnedders> | Oh, sure, it will be. But I'm not sure I want my website to have to cope with the load of Hixie generating HTML 5 :P |
| 17:25 | <krijnh> | csarven: any particular reason why I should double my bandwidth with those fluffy microformats? :) |
| 17:26 | <Philip`> | krijnh: You have compression, so it wouldn't double :-) |
| 17:27 | <krijnh> | http://www.csarven.ca/labs/microformats/krijnhoetme.nl_irc-logs/code-change.txt |
| 17:27 | <krijnh> | Anybody else think this is useful? |
| 17:27 | <krijnh> | (It would mean some extra overhead in parsing each line, to get a timestamp) |
| 17:28 | <Philip`> | You should at least s/_a/data-author/ to be conforming HTML5 :-) |
| 17:28 | <krijnh> | But I want it to be accessible! :) |
| 17:29 | <krijnh> | I've read on the internets that HTML5 doesn't care about accessibility |
| 17:29 | <krijnh> | *changes doctype :X* |
| 17:30 | Philip` | doesn't personally have a use for any of the proposed changes |
| 17:33 | <krijnh> | csarven: do you? |
| 17:34 | gsnedders | realizes it's basically the weekend now |
| 17:35 | gsnedders | is slow |
| 17:37 | <Philip`> | jgraham: http://james.html5.org/cgi-bin/tables/table_inspector.py?input_type=type_uri&uri=http%3A%2F%2Fquickfacts.census.gov%2Fqfd%2Fstates%2F34%2F34003.html&source=&algorithm=html5 :-( |
| 17:39 | <csarven> | krijnh From #laconica: [10:18:10] <csarven> So, what I was thinking last night with hAtom->Atom->XMPP was to have IRC logs available as notices |
| 17:39 | <csarven> | People can also subscribe to the logs (for whatever reason) |
| 17:39 | <krijnh> | Would it be okay if I only made the fluffy format available for #microformats ? |
| 17:40 | <krijnh> | As a test :) |
| 17:40 | <csarven> | Of course, it is your call at the end :) |
| 17:40 | <krijnh> | Yeah, I'm the evil IRC logs overlord here :) |
| 17:40 | <Philip`> | They're not evil IRC logs |
| 17:41 | <csarven> | SO, the creator is evil but no the product |
| 17:41 | <krijnh> | Okay, you lost me :) |
| 17:41 | <Philip`> | By "subscribe to logs", you mean "put greatly increased load on the logs server"? :-) |
| 17:41 | <csarven> | :S |
| 17:42 | <csarven> | This is experimental. I highly doubt it is be impacted by load |
| 17:42 | <krijnh> | Okay, I'll fix it for #microformats, as an experiment |
| 17:42 | <csarven> | s/it is be/it would be |
| 17:42 | <krijnh> | Just like the [off] thing in #webapps :) (which was a FAIL btw) |
| 17:43 | <csarven> | Thanks. |
| 17:44 | <krijnh> | When I've got some time for it (probably after September 13th) |
| 17:44 | <csarven> | Cool. |
| 17:44 | <csarven> | I'll bug you on September 14 ;) |
| 17:45 | csarven | goes to hunt lunch |
| 17:45 | <krijnh> | csarven: Oki :) |
| 17:48 | annevk | will appear on http://openwebpodcast.com/ later today / tomorrow |
| 17:49 | <gsnedders> | Lachy: The prophesy in /topic will likely be fulfilled tomorrow |
| 17:50 | <gsnedders> | Philip`: more info from guy at Yahoo! Search: "BTW: when Wustl.edu got hit with this, they said that it was a modrewrite in Apache that was doing all this." |
| 18:01 | <hsivonen> | sigh. it seems that my data is corrupt. |
| 18:02 | <hsivonen> | at least Python's gzip and Java's gzip see the same problem |
| 18:02 | <hsivonen> | and zlib itself is supposed to get large files right |
| 18:03 | Philip` | wonders what's the probability of getting a 1-bit error on a full hard disk |
| 18:04 | <annevk> | geez, this RDF thread sure creates a lot of e-mail |
| 18:05 | <annevk> | the last six months we had around 300 messages, now it's almost 600 |
| 18:05 | <annevk> | (300 messages on average) |
| 18:14 | <annevk> | http://lists.w3.org/Archives/Public/www-archive/2008Aug/0145.html o_O |
| 18:27 | <Philip`> | http://factfinder.census.gov/servlet/QTTable?_bm=n&_lang=en&qr_name=DEC_2000_SF1_U_DP1&ds_name=DEC_2000_SF1_U&geo_id=05000US48487 has interestingly nested headers |
| 18:37 | <annevk> | http://juicystudio.com/wcag/tables/complexdatatable.html uses <td class=header> uh? |
| 18:39 | <Philip`> | The cells contain data, so they should be <td> |
| 18:40 | <Philip`> | where you could define "data" as being "any values that get returned by the SQL query I'm using to generate this table" (as opposed to the headers which are strings that are typed in somewhere else) |
| 18:41 | <sicking> | annevk, does the XHR spec say what to do if the supplied username/password fails? |
| 18:41 | <sicking> | annevk, appears we put up a dialog asking the user, other browsers don't |
| 18:48 | <annevk> | sicking, it does |
| 18:49 | <annevk> | sicking, http://dev.w3.org/2006/webapi/XMLHttpRequest/#send "If authentication fails, and stored user and stored password are not provided, user agents should prompt the users for credentials." |
| 18:51 | <sicking> | annevk, cool |
| 18:52 | <sicking> | annevk, though |
| 18:53 | <sicking> | annevk, that doesn't really say what to do if the script provides a username/password |
| 18:54 | <annevk> | sicking, I could make it more explicit, the intent is that you don't do any UI interaction |
| 18:55 | <annevk> | (the idea is that the spec states what you need to do and everything else is "forbidden") |
| 18:55 | <sicking> | annevk, "user agents should prompt the users for credentials" sounds like UI interaction, no? |
| 18:55 | <annevk> | well yes, in that case everyone does UI interaction so it made sense to keep that |
| 19:04 | <sicking> | annevk, ah, so the idea is if the script provides credentials, use those, if they fail just fail. If the script doesn't provide, ask the user if needed? |
| 19:06 | <annevk> | sicking, yes |
| 19:06 | <sicking> | annevk, makes sense |
| 19:11 | <annevk> | good :) |
| 20:58 | <Hixie> | gsnedders: the main reason i like using web services is it means i can kick off several jobs without having to worry about the load they put on the machine, since they're all on different machines |
| 20:59 | <gsnedders> | Hixie: On the whole I expect it'll be all right… We'll see how it goes :) |
| 21:14 | <annevk> | did anyone put up a web service yet? |
| 21:14 | <annevk> | otherwise I might give it a try |
| 21:14 | <gsnedders> | No |
| 21:18 | <gsnedders> | annevk: As I said before, I can put it up on my website |
| 21:40 | gsnedders | wonders whether to bother to support py2.3 |
| 21:41 | <BenMillard> | gsnedders, shall I make the hotel booking then? |
| 21:41 | <gsnedders> | BenMillard: If you want to. :) |
| 21:41 | <BenMillard> | gsnedders, it's not something which excites me but I guess I should do some part of the logistics. :P |
| 21:42 | gsnedders | tries to understand that column beside the form |
| 21:42 | <BenMillard> | gsnedders, my parents are taking my trip as an excuse to visit Cannes, so much of my travel arrangements will be identical to theirs |
| 21:42 | <gsnedders> | Heh. |
| 21:43 | <gsnedders> | The booking has to be confirmed by phone or by fax |
| 21:43 | gsnedders | looks at the French, and instantly it makes more sense |
| 21:45 | <Hixie> | i sent an e-mail and they confirmed it no problem |
| 21:45 | <Hixie> | still have to get my flights though |
| 21:45 | gsnedders | goes downstairs to con a card out of his parents to buy train tickets |
| 21:46 | <gsnedders> | And I'm gonna have to find somewhere to stay in Cannes for two nights, unless I go straight from my uncle's to the TOAC |
| 21:46 | <gsnedders> | *TPAC |
| 21:46 | <gsnedders> | Which on the whole I wasn't going to do |
| 21:47 | <gsnedders> | Hmm… |
| 21:48 | <gsnedders> | Maybe I should stay in Antibes… |
| 21:50 | <gsnedders> | Trains going from Lyon south are remarkably expensive compared with those further north |
| 21:58 | <annevk> | gsnedders, if you can do it that's fine too |
| 21:59 | <Hixie> | opening a new tab in IE8 takes takes an ungodly amount of time |
| 22:00 | <Hixie> | hitting ctrl+f in IE8 on the html5 spec is a bad idea |
| 22:00 | <annevk> | fwiw: http://openwebpodcast.com/episode-3-anne-van-kesteren-on-css-xhr-and-other-web-standards |
| 22:02 | <Hixie> | wow this find in page sucks ass on the spec |
| 22:06 | <Hixie> | aah, you mention me |
| 22:06 | Hixie | hides |
| 22:06 | <Hixie> | aaah, you did it again! |
| 22:11 | <Hixie> | hahaha |
| 22:11 | <Hixie> | that was funny |
| 22:11 | <Hixie> | about acid3 |
| 22:11 | <Hixie> | "so ian can add anything to acid3, is that a bad thing?" "well actually it was the bits that other people added that were the problem..." |
| 22:12 | <Hixie> | (personally i think the range/traversal stuff i added to acid3 was a bigger problem than the smil stuff) |
| 22:34 | Lachy | disobeyed Hixe's instructions and pressed Ctrl+F on the spec in IE8... It's *really* not a good idea! |
| 22:34 | <Hixie> | BenMillard: your e-mail looks fine to me (indeed i'm saving it because it's got so much useful data and logical arguments!) |
| 22:34 | <jruderman> | Lachy: what happens? |
| 22:35 | <BenMillard> | Hixie, yay. :) |
| 22:35 | <Lachy> | jruderman, it takes ages to find anything |
| 22:35 | <Lachy> | become unresponsive |
| 22:37 | <Hixie> | fwiw my plan with this tables stuff is to do exactly what i was going to do anyway, which is to get to it when i get to it, and when i get to it, read all the arguments and research, and come to a conclusion based on that |
| 22:37 | <BenMillard> | Lachy, I've just forwarded the message Hixie is talking about to you, since you were at that telecon |
| 22:38 | <Hixie> | and unless some browser or AT vendor contacts me to say that this is urgent for them, i don't plan on getting to it any time soon |
| 22:38 | <BenMillard> | Hixie, yeah that's what I expect. You said after the changes earlier this year it'd be a while before returning to it, which works out well for me since I get time to do more research. |
| 22:38 | <Hixie> | i'm amused that chris wilson's e-mail suggests that we're not making progress |
| 22:38 | <Hixie> | and that the table headers thing is important for making progress |
| 22:39 | <Hixie> | especially given that the whole point of the table headers proposal, as i understand it, is to make it possible to mark up pages that work in existing ATs |
| 22:39 | <Hixie> | in a way that HTML4 still allows |
| 22:39 | <Hixie> | and is thus pretty much the exact opposite of "required for progress" |
| 22:40 | <BenMillard> | for the records, we're talking about this Public-HTML thread: http://lists.w3.org/Archives/Public/public-html/2008Aug/thread.html#msg518 |
| 22:43 | <annevk> | I'm confused by CWs e-mail. Not sure whether to reply |
| 22:43 | <BenMillard> | (I'm gathering some peer review of my message before adding it to that topic.) |
| 22:43 | <annevk> | ah, I see Ben will reply, good :) |
| 22:46 | <annevk> | BenMillard, skimmed/read through your e-mail, looks good to me |
| 22:49 | <BenMillard> | annevk, cheers |
| 22:50 | <BenMillard> | I don't anticipate major changes unless reviewers find mistakes (there are a sprinkling of typos I'll address, though) |
| 22:55 | <Hixie> | oh great, anne's told everyone that the real reason for websocket is my model railway :-P |
| 22:57 | <annevk> | anecdotical evidence to feed the flames :p |
| 22:58 | <Lachy> | BenMillard, ok. reading it now |
| 22:58 | <Hixie> | annevk: good interview |
| 22:59 | <annevk> | ta |
| 22:59 | <annevk> | gsnedders, so are you putting up a Web service tonight? |
| 22:59 | <BenMillard> | smedero has sent me a review of the message. it's enlightening but I think it doesn't change what my message says |
| 22:59 | <smedero> | no, it shouldn't |
| 23:00 | <smedero> | it is just backstory, since you probably haven't been following the telecons.... |
| 23:00 | <BenMillard> | yes, thanks for providing me with that |
| 23:07 | <jacobolus> | are unescaped ampersands allowed in html? |
| 23:08 | <annevk> | sometimes |
| 23:10 | <jacobolus> | the validator seems happy to accept “<!DOCTYPE html><title></title>&” |
| 23:11 | <annevk> | that's correct (not sure if the syntax section accurately reflects it currently) |
| 23:12 | <jacobolus> | I didn't look super carefully, but I only noticed a mention w.r.t. xml |
| 23:12 | <annevk> | jacobolus, it does not need to be escaped when followed by a space, EOF, another ampersand, start tag, end tag, comment |
| 23:12 | <jacobolus> | ah, okay |
| 23:13 | <jacobolus> | “<!DOCTYPE html><title></title>AT&T” properly fails then |
| 23:13 | <annevk> | yup |
| 23:14 | <annevk> | hsivonen, "Probable cause: & should have been escaped as &D." what is the "D" doing there? |
| 23:15 | <BenMillard> | annevk, nothing much. :P |
| 23:16 | <jacobolus> | ah, it is described in detail, but should maybe be more explicit what authors should do? |
| 23:16 | <jacobolus> | (i.e. what implementors should do is described in detail) |
| 23:17 | <annevk> | section 8.1 describes in detail what authors should do |
| 23:17 | <jacobolus> | oh, nevermind |
| 23:17 | <annevk> | though it seems it's not entirely clear on things |
| 23:17 | <jacobolus> | the text must not contain the character U+003C LESS-THAN SIGN (<) or an ambiguous ampersand. |
| 23:18 | <jacobolus> | okay, that's reasonable :) |
| 23:18 | <jacobolus> | http://www.whatwg.org/specs/web-apps/current-work/#ambiguous doesn't mention EOF :) |
| 23:19 | <BenMillard> | Lachy, are you double-checking my numbers? if so, thanks! |
| 23:20 | <annevk> | jacobolus, yeah, whatwg⊙wo ;) |
| 23:21 | <Lachy> | BenMillard, I suppose I can go through it again and do that. But let me finish it at least once first... I'm in the middle of doing some other things too |
| 23:21 | <BenMillard> | Lachy, sure thing...I don't wish to monopolise your time |
| 23:25 | <Lachy> | BenMillard, btw, I found a rather interesting table might be interesting for you to study. But it requires membership to see it on this site http://www.newzbin.com/ |
| 23:26 | <BenMillard> | Lachy, would it be permitted for you to extract the table and send it to me for analysis and publishing in my collection? my work is non-commercial research so I imagine that would qualify as "fair use" |
| 23:26 | annevk | summons zcorpan |
| 23:27 | BenMillard | hopes zcorpan will arrive with a witty /me line... |
| 23:27 | <Lachy> | BenMillard, sure. I'll do it after I'm done with your email |
| 23:28 | <BenMillard> | Lachy, cool. I'll go start some dinner :) |
| 23:31 | <gsnedders> | annevk: As I said, if anyone writes it :P |
| 23:32 | <gsnedders> | BenMillard: Peh! Silly logn emails you want me to read! |
| 23:32 | <gsnedders> | *long |
| 23:34 | <Lachy> | gsnedders, you could just pretend to read it, skim it and post random comments about things you see to make him think you're really reading it. ;-) |
| 23:34 | <gsnedders> | Lachy: True. |
| 23:34 | gsnedders | hopes all this crazy insane ECMAScript optimization work shows through in other interpreted languages |
| 23:37 | <annevk> | gsnedders, oh I see |
| 23:37 | <annevk> | well, now it's too late |
| 23:37 | <annevk> | I guess tomorrow I can download the relevant packages and try making it work, shouldn't be too hard |
| 23:37 | <gsnedders> | annevk: It's probably worth working from what I'll put in hg tomorrow |
| 23:38 | <gsnedders> | Actually, if you want to work on it in the morning, I could do that now |
| 23:40 | jgraham | wonders if there is any point in arguing process with CW |
| 23:40 | <BenMillard> | gsnedders, there's no obligation and no salesperson will call. :D |
| 23:41 | <jgraham> | I think I'll fix html5lib first |
| 23:41 | <gsnedders> | Hixie: PostScript is a turing-complete language |
| 23:41 | <annevk> | gsnedders, that doesn't mean it will cause a risk |
| 23:41 | gsnedders | doesn't know enough to know that |
| 23:41 | jgraham | saw a ~3 line postscript file that did a ray tracing image of a ball on a chessboard or something once |
| 23:41 | <gsnedders> | I'm just pointing out what I know :P |
| 23:41 | <Philip`> | Turing machines with no IO devices are a bit rubbish in practice |
| 23:42 | gsnedders | passes Philip` an IO device |
| 23:42 | <Philip`> | I suppose you could try to make them jump backwards and forwards really fast so their tape catches fire |
| 23:44 | <gsnedders> | Philip`: Problem with that is next to no-one actually implements a turing-machine using tape |
| 23:45 | <Philip`> | Partly because it'd be physically impossible to implement one at all |
| 23:45 | <Lachy> | BenMillard, your byte counts for the sizes of the tables seem to be a little off |
| 23:45 | <BenMillard> | Lachy, perhaps there was a bug during the copying... |
| 23:46 | <Lachy> | I saved the both files using wget, stripped out all markup before and after the <table>...</table> and resaved the files |
| 23:46 | <BenMillard> | Lachy, what numbers do you get and by which method? I expect the error is on my part |
| 23:46 | <Lachy> | For noscope.html, I get 1659, and for complexdatatable.html, I get 2591 |
| 23:48 | <BenMillard> | hmm, what types of line endings do you see? I see 2 characters per newline, which might be my error |
| 23:48 | <Lachy> | what method did you use to get the sizes? |
| 23:49 | <BenMillard> | I viewed source in Firefox 2, copied and pasted into a plain text editor, then selected the text from the start of "<table" to the end of "</table>" and read off what the statusbar said |
| 23:49 | <BenMillard> | your method sounds better to me :) |
| 23:49 | gsnedders | reads the start of BenMillard's email and turns off |
| 23:49 | <Lachy> | line endings are CRLF |
| 23:50 | <BenMillard> | Lachy, same here |
| 23:50 | <Lachy> | the problem with copying and pasting from firefox's view source is that it doesn't copy properly |
| 23:51 | <Lachy> | it frequently adds in extra lines in random places |
| 23:51 | <BenMillard> | Lachy, I now measure 2,625 for the complexdatatable.html |
| 23:51 | <BenMillard> | Lachy, yes I removed the empty lines |
| 23:51 | <BenMillard> | I'm happy to steal your numbers if that's OK by you :) |
| 23:52 | <Lachy> | no, those numbers are copyrighted to me! :-) |
| 23:53 | <BenMillard> | Lachy, perhaps you could e-mail me the files you downloaded and the cropped versions, then I can see where the difference is |
| 23:53 | <BenMillard> | thanks for going to the trouble to do this, btw |
| 23:55 | <Lachy> | http://lachy.id.au/temp/tables.zip |
| 23:57 | <BenMillard> | Lachy, both the text editor's selection and Windows Explorer agree with your numbers |
| 23:58 | <Lachy> | did you find the diffs with your files? |
| 23:59 | <BenMillard> | Lachy, in noscope.html I see class="header" on the <td> for each of the 3 dates |
| 23:59 | <jgraham> | argh |