| 00:32 | <Philip`> | Hmm, IE seems to support some named entities that Mozilla (and, I think, HTML5) don't |
| 00:32 | <Philip`> | in particular, 8203: zwsp; 8234: lre; 8235: rle; 8236: pdf; 8237: lro; 8238: rlo; 8298: iss; 8299: ass; 8300: iafs; 8301: aafs; 8302: nads; 8303: nods |
| 00:32 | Philip` | wonders if anyone cares about those |
| 00:39 | <zcorpan_> | Philip`: how did you find those? |
| 00:40 | <Philip`> | http://www.bbc.co.uk/dna/h2g2/A264548 has a list that includes them |
| 00:41 | <zcorpan_> | ok |
| 00:42 | <zcorpan_> | ie seems to be alone in supporting those |
| 00:43 | <zcorpan_> | but i'm not against adding them to html5 |
| 00:45 | <Philip`> | They don't appear to be the most highly documented entity names |
| 00:45 | <Philip`> | (like, that one list is the only place on the whole web) |
| 00:46 | <zcorpan_> | do they appear in your html research sample? |
| 00:46 | <zcorpan_> | (probably not) |
| 00:52 | <Philip`> | I didn't collect any data about unrecognised entities when I was looking at the ~8K pages |
| 00:53 | <Philip`> | (The tokeniser did actually report unrecognised entities but I just discarded all that data instead of saving it anyhwere) |
| 00:53 | <Philip`> | s/hw/wh/ |
| 00:53 | <Philip`> | In the other ~2.5K pages that I still have a copy of, none of those entities turn up at all |
| 00:56 | <zcorpan_> | yeah, we'd probably need a billion documents research to find the relevance of these entities |
| 00:59 | <Philip`> | http://www.tools.ietf.org/html/draft-duerst-iri-bidi-00 talks about lre, lro, rle, rlo, pdf |
| 01:01 | <zcorpan_> | Philip`: will you post this to the list? |
| 01:01 | <Philip`> | but that looks like it's just a coincidental use of the same names |
| 01:01 | <Philip`> | I will do |
| 01:01 | <Philip`> | (But which list? :-) ) |
| 01:02 | <zcorpan_> | (doesn't matter) |
| 01:02 | <zcorpan_> | i'll cover btw |
| 01:11 | <Philip`> | Opera 9.2 is missing ® |
| 01:48 | <Philip`> | http://software.hixie.ch/utilities/js/live-dom-viewer/?%3Cp%3E%3C/p%3E%3Cp%3E%26ampxyz123%3C/p%3E%0D%0A%3Cscript%3Edocument.getElementsByTagName%28%27p%27%29%5B0%5D.innerHTML%3D%27%26ampxyz123%27%3C/script%3E%0D%0A |
| 01:48 | <Philip`> | in IE |
| 01:50 | <Philip`> | When you set innerHTML, it appears to discard all alphanumeric characters after a "&" (or whatever) if there are no non-alphanumeric characters before the end of the string |
| 01:58 | <Philip`> | http://canvex.lazyilluminati.com/misc/entities.html |
| 01:58 | <Philip`> | That appears to reliably crash Firefox trunk |
| 02:03 | zcorpan_ | points at http://simon.html5.org/test/html/parsing/entities/trailing-semicolon/real/ |
| 02:05 | <Philip`> | Aha |
| 02:05 | <Philip`> | But mine makes IE fail more ;-) |
| 02:06 | <Philip`> | (and works in html5lib/etc, though unfortunately all the implementations get everything right which is no fun) |
| 02:08 | <zcorpan_> | :) |
| 02:18 | <Philip`> | http://yy28.60.kg/test/read.cgi/maido3/1096370177/l50 - there's someone using &zwsp; |
| 02:18 | <Philip`> | http://www.tasb.com/services/field/staff/index.aspx?print=true - there too |
| 02:18 | <Philip`> | It's lucky that Google doesn't understand these entities, so it lets you search for them by name |
| 02:19 | <zcorpan_> | there you go |
| 02:21 | <Philip`> | Most of the references seem to be in Chinese ASCII-art comment spam (or is that not spam?), largely on 2ch.net |
| 02:25 | <zcorpan_> | it's Shift_JIS-art (*ﾟーﾟ) |
| 02:41 | <deltab> | no zwsp in http://www.w3.org/TR/html4/sgml/entities.html |
| 02:41 | <deltab> | ... but you probably already knew that |
| 02:42 | <zcorpan_> | yeah... all entities in html4 are already in html5 |
| 02:42 | <zcorpan_> | (i hope!) |
| 02:43 | <Philip`> | At least all the ones in Mozilla and Opera are |
| 02:43 | <Philip`> | Only IE is 'special' :-) |
| 10:24 | <annevk> | CRLF was already raised once on whatwg⊙wo |
| 10:24 | <annevk> | g |
| 10:28 | <hsivonen> | implementing a check for NCName makes me wonder if the original XML WG stopped to write down the code for checking it before making it part of the spec... |
| 10:29 | <annevk> | In XML5 it's easy :) |
| 10:44 | <hsivonen> | speaking of NCNames and stuff: are there known test cases that test the detection of various XML spec lawyering NCName subtleties? |
| 10:47 | <hsivonen> | I wonder what back end software http://www.weblogs.com/feedvalidator.html runs |
| 10:48 | annevk | hopes his comparison between Harry Potter and specs won't be seen as offensive |
| 10:48 | <hsivonen> | annevk: on list or on blog? |
| 10:48 | <annevk> | list |
| 10:49 | <hsivonen> | annevk: my guess is it will be seen as offensive |
| 10:50 | <annevk> | bah |
| 10:53 | <hendry> | damn, i had the badware pcpro popup whilst using my bank's website... |
| 10:53 | <hsivonen> | I wrote a regexp for NCName suitable for sticking into a Java String literal and it is 7875 characters long! |
| 10:55 | <annevk> | ouch |
| 10:56 | <annevk> | write your own tokenizer? :) |
| 10:57 | <hsivonen> | annevk: this is part of the feature set for parsing HTML into XML 1.0 infosets |
| 10:58 | <hsivonen> | annevk: no I need to be able to detect non-NCNames either to treat them as fatal or to drop them |
| 10:58 | <hsivonen> | anyway, the regexp is so long that Eclipse refuses to scroll enough to the right to show it all |
| 11:00 | <hsivonen> | using ICU4J for this would be cleaner and might even perform better, but I want to avoid the dependency |
| 11:03 | <annevk> | omg, this quote/unquot has to stop |
| 11:12 | <hsivonen> | at least HTML 5 is much more author-friendly reading than any ISO spec |
| 11:13 | <hsivonen> | (I wonder how much work it takes to obfuscate something like ODF or OOXML to comply with ISO drafting rules.) |
| 11:13 | <hsivonen> | (or might ISO make an exception on those?) |
| 11:13 | <hsivonen> | (not to say that being more author-friendly than an ISO spec is enough) |
| 11:26 | <annevk> | he took it as insult |
| 11:28 | <hsivonen> | annevk: not surprising. if it wasn't meant as one, it was a badly formulated non-insult |
| 11:30 | <annevk> | I guess |
| 13:24 | <hsivonen> | canvas at work: http://westciv.com/xray/ |
| 13:26 | <annevk> | in Opera I always get back <canvas> |
| 13:26 | <annevk> | I wonder if that's a bug |
| 13:26 | <hsivonen> | annevk: documentation says it only works in Gecko and WebKit. not working in Presto and Trident is known |
| 13:27 | <annevk> | of course, but what's the issue? |
| 13:27 | <annevk> | and is that an issue in Opera or Firefox? |
| 13:28 | <hsivonen> | I don't know. Based on what you said, I guess the issue is making the canvas event-neutral so that clicks go through it |
| 13:31 | <annevk> | jgraham, simpleson has no attribute "loads" causes tons of errors |
| 13:32 | <annevk> | the rests of the tests are about some silly space after <!DOCTYPE ... |
| 14:00 | <G0k> | hey all |
| 14:04 | <annevk> | hi |
| 14:05 | <hsivonen> | hi |
| 14:14 | <G0k> | so this TCP/UDP connection stuff in HTML5 |
| 14:14 | <G0k> | has anyone actually tried implementing that? |
| 14:14 | <annevk> | I don't think so |
| 14:14 | <annevk> | That section is very much unstable, afaict |
| 14:15 | <G0k> | yeah i mean |
| 14:15 | <G0k> | i guess i'm not really even clear on its...purpose |
| 14:15 | <G0k> | i mean it seems kinda misnamed |
| 14:15 | <annevk> | communicating without the overhead of HTTP |
| 14:16 | <annevk> | and I believe it was also intended to address P2P at some point |
| 14:16 | <G0k> | it's not really a TCPConnection so much as a NewProtocolWeJustInventedThatHappensToUseTCPConnection |
| 14:16 | <hsivonen> | e.g. for games, IM clients, continuous visualizations of data, etc. |
| 14:16 | <hsivonen> | G0k: it isn't a pure TCP connection for security reasons |
| 14:17 | <G0k> | yeah that's surely legitimate, but calling it a TCP connection isn't really accurate |
| 14:17 | <G0k> | i mean there's also no reason that protocol couldn't be made to run over SCTP, for instance |
| 14:18 | <annevk> | you could point it out on the list |
| 14:19 | <G0k> | yeah that would require typing in complete sentences and stuff, it's much easier to complain to you people |
| 14:20 | <hsivonen> | G0k: the connection initialization if coupled with TCP ports isn't it? |
| 14:21 | <annevk> | there's also http://dev.w3.org/cvsweb/~checkout~/2006/webapi/network-api/network-api.html?rev=1.3 |
| 14:21 | <annevk> | G0k, hah |
| 14:21 | <G0k> | well yeah i mean the DOM interface is TCP-specific but the protocol itself needn't be |
| 14:23 | <met_> | annevk, you remind me Don LaFontaine in your spot http://annevankesteren.nl/2007/07/web it's like I hear his common sentence "I a strict world where..." (see http://www.youtube.com/watch?v=ZJMGS7l0wT8 ) |
| 14:25 | annevk | doesn't have flash atm |
| 14:27 | hsivonen | didn't know that "Don" is a real one guy |
| 14:28 | <met_> | hsivonen see http://www.youtube.com/watch?v=Wkhdy6bavuk |
| 14:30 | <hsivonen> | met_: just watched it |
| 14:33 | <met_> | Don's rewriting of annevk's spot: In a strict world, where pages using those new features completely break, new hero will rise. HTML5 from producers of Safari, Opera and Firefox. In December 2010 ! |
| 14:36 | <annevk> | with 10% discount if you order now |
| 14:44 | <Lachy> | hey, I'm looking for some example flash sites that provide alternative HTML versions of the site. In particular, ones that link to the HTML-only versions of the site below the flash. Anyone know of any? |
| 14:45 | <Lachy> | or even sites that use video and provide alternative content |
| 14:45 | <annevk> | http://www.google.com/search?q=%22skip+intro%22 ? |
| 14:45 | <Philip`> | http://www.jkrowling.com/ has a text-only version |
| 14:46 | <Lachy> | yeah, I already got jkrowling |
| 14:46 | <annevk> | http://www.google.com/search?q=%22text+only+version%22 |
| 14:47 | <hsivonen> | Lachy: I'd start by looking at recent movie promo sites |
| 14:47 | <hsivonen> | Lachy: with this method, starting from the Apple Trailer page, I found http://www.rushhourmovie.com/ |
| 14:47 | <Lachy> | oh, good idea |
| 14:47 | <hsivonen> | it puts the textual alternative in noscript! |
| 14:48 | <Lachy> | oh, and it uses black text on black background when you disable JS |
| 14:49 | <zcorpan_> | all i get is "A flash player upgrade is required to view this website, click here to continue." |
| 14:49 | <hsivonen> | http://www.noendinsightmovie.com/ also has a short writeup in the source |
| 14:49 | <hsivonen> | hmm. perhaps the alternative content is for google--not browser users |
| 14:53 | <hsivonen> | the interesting thing is that even though movie promos most likely aren't trying to take blind users into account, they still bother putting text in there. this suggests that pitching textual alternatives as a SEO method might be attractive to site makers |
| 14:55 | <hsivonen> | hmm. http://cleanishappy.com/ contains a really long text writeup aimed at seach engines |
| 14:55 | <hsivonen> | but useless for users who can't browse Flash |
| 14:58 | <Lachy> | wow, that's got to be the longest meta description I've ever seen! |
| 15:04 | <hsivonen> | Lachy: http://del.icio.us/tag/flash%2Bmarketing |
| 15:07 | <Lachy> | thanks |
| 15:37 | annevk | wonders why SVGSVGElement is not named SVGSvgElement |
| 15:48 | <annevk> | <dfn title="GDO">Garage Door Opener (<abbr>GDO</abbr>)</dfn> ... |
| 15:54 | <zcorpan_> | in typography, is it common to have the whole thing in italics, or just the "Garage Door Opener" part, or perhaps just the "GDO" part? |
| 15:54 | <annevk> | dunno |
| 17:01 | <gsnedders> | only 373 unread emails on public-html having been away |
| 17:19 | <Philip`> | Is this month going to beat May? |
| 17:19 | <Philip`> | It's only 15 behind at the moment |
| 17:23 | <zcorpan_> | we need something controversial |
| 17:24 | <annevk> | that should be trivial |
| 18:11 | <Lachy> | why on earth does Firefox return true for if(document.createNodeIterator); but it returns an NS_ERROR_NOT_IMPLEMENTED whenever I try to invoke it!? |
| 18:12 | <hasather> | Lachy: when are you coming to Norway? |
| 18:13 | <Lachy> | don't know yet. |
| 18:13 | zcorpan_ | might come to norway sometime soon also |
| 18:13 | <hasather> | zcorpan_: ahh, cool |
| 18:13 | <zcorpan_> | perhaps next week |
| 18:14 | <Lachy> | cause I want to go to Web Directions at the end of September, and if I go over in August, then I'd be coming back a few weeks later, so it was suggested that I go over in October |
| 18:14 | <hasather> | Lachy: so you're not coming in August? |
| 18:15 | <Lachy> | probably not, but I haven't heard back since I last emailed Annchen |
| 18:15 | <hasather> | ok |
| 18:16 | <annevk> | zcorpan_, you have to come next week :) |
| 18:16 | <Lachy> | is the difference between a NodeIterator and a TreeWalker that the NodeIterator only looks at child nodes, and the TreeWalker looks at all descendants? |
| 18:16 | <zcorpan_> | annevk: when are you leaving again? |
| 18:17 | <annevk> | zcorpan_, I'm going to Spain the 11th of August and will be back briefly the 17th to attend Friday beer :) The 18th I'll go to the Netherlands for some more vacation afer which university starts again... |
| 18:17 | <zcorpan_> | alrighty |
| 18:19 | <Philip`> | Lachy: createNodeIterator was added in https://bugzilla.mozilla.org/show_bug.cgi?id=82625 |
| 18:20 | <Philip`> | and it looks like those functions were just added for completeness |
| 18:21 | <Lachy> | Philip`: assuming you mean createTreeWalker was added in that bug, ok. |
| 18:21 | <Lachy> | but it looks like I should be able to use a TreeWalker and nextSibling for my needs anyway |
| 18:26 | <Philip`> | I meant createNodeIterator, or at least the createNodeIterator stub that just returns NS_ERROR_NOT_IMPLEMENTED |
| 18:49 | <Lachy> | Robert Burns' new quoting technique is so much harder to read :-( |
| 18:51 | <Lachy> | in fact, I just can't read it, cause I can't keep track of where I'm up to! |
| 18:52 | <Lachy> | I sent him off list mail telling him about RFC 2646 and asking him to comply with the existing convention |
| 18:53 | <annevk> | I gave up on reading that e-mail |
| 18:54 | annevk | thought Hixie would be back this week |
| 18:56 | <annevk> | given http://krijnhoetmer.nl/irc-logs/whatwg/20070705#l-669 he should be, maybe he died! |
| 19:13 | <Lachy> | I hope he hasn't |
| 19:14 | <Lachy> | I wonder who would know where he is? |
| 19:15 | <Lachy> | some weird is happening with my script. I'm getting a non-existent COL element returned from the tree walker |
| 19:22 | <Lachy> | oh, it's not weird, it's my mistake :-) |
| 21:21 | <jgraham> | Astronomy spam is the best |
| 21:33 | <hasather> | annevk: he's alive: http://twitter.com/Hixie/statuses/178895922 |
| 22:00 | <Hixie> | i have returned |
| 22:02 | <jgraham> | Nice holiday? |
| 22:02 | <Hixie> | certainly it was nice to have a break from public-html |
| 22:02 | <jgraham> | :) |
| 22:03 | <hasather> | Hixie: welcome back |
| 22:04 | <Hixie> | so, anything bad happen while i was away? any servers die? |
| 22:04 | <Hixie> | configurations go mysteriously awry? |
| 22:04 | <hsivonen> | Hixie: there have been complaints about whatwg archives |
| 22:04 | <hsivonen> | Hixie: URIs broke or something |
| 22:04 | Hixie | looks |
| 22:04 | <jgraham> | The archives are private |
| 22:06 | <Hixie> | fixed |
| 22:06 | <Hixie> | nothing i can do about the uri changes |
| 22:19 | <hsivonen> | hmm. 1300 messages on public-html this month |
| 22:19 | <hsivonen> | July, that is |
| 23:26 | <Philip`> | Three more messages to go... |