| 00:00 | <Philip`> | It'd be nice if it linked to the relevant part of WebIDL :-) |
| 00:00 | <Hixie> | i'm waiting for gsnedders to give us cross-spec auto-xreffing :-) |
| 00:00 | Philip` | kind of assumed that NameGetter would be called for any name that wasn't already part of the interface |
| 00:00 | <Philip`> | gsnedders: It'd be nice if it linked to the relevant part of WebIDL :-) |
| 00:00 | <Hixie> | i've already set up the cross-references for this, it just isn't being generated automatically yet |
| 00:01 | <Hixie> | Philip`: if you had it defined that way, you couldn't enumerate over the object |
| 00:02 | <Philip`> | Hixie: Enumeration could be defined over some finite list of keys, but NameGetter could still be called for all name gets regardless of whether they're in that list |
| 00:02 | <Philip`> | but I guess it's not done that way |
| 00:02 | <Philip`> | in which case that's fine |
| 00:03 | Philip` | prefers it when he doesn't have to read specs to work out what behaviour is specified, but that might be asking for too much |
| 00:04 | <Hixie> | i'm just doign what heycam and the browser vendors tell me to do here |
| 00:04 | <Hixie> | if you have any suggestiosn on making the spec clearer, let me know :-) |
| 00:04 | <Dashiva> | Is this behavior described in WebIDL? |
| 00:05 | <Hixie> | yes |
| 00:05 | <Hixie> | web storage invokes all the webidl terms |
| 00:05 | <Hixie> | they re underlined in green |
| 00:06 | <Philip`> | By green, do you mean blue? |
| 00:06 | <Dashiva> | Yes, but I don't see anything about the undefined/null duality there |
| 00:06 | <Hixie> | er, blue in the w3c version, yes |
| 00:06 | <Philip`> | Is there a non-W3C version? |
| 00:06 | Philip` | has a hard time finding a copy of the spec via Google whenever he wants it |
| 00:06 | <Hixie> | i don't think i'm currently generating a whatwg version of the web storage spec, no |
| 00:07 | <Philip`> | Oh, neat, Google's 'promote' thing lets me make it the top search result for "web storage" |
| 00:07 | <Hixie> | the web storage spec is the 6th hit for [web storage] on google |
| 00:08 | <Dashiva> | Hmm |
| 00:08 | <Dashiva> | No NameCreator |
| 00:08 | <Philip`> | I'm afraid I don't acknowledge the existence of any search results beyond 2 from the top, except in exceptional circumstances |
| 00:08 | <Hixie> | heh |
| 00:08 | Hixie | just uses the links on http://☺.damowmow.com/ |
| 00:09 | <Dashiva> | Okay, so reading WebIDL the "supported named properties" thing supports the null/undefined duality thing |
| 00:09 | <Dashiva> | But that also means you can't create new properties with .x = whatnot, since there's no NameCreator |
| 00:10 | <Hixie> | er yeah, i should put in a NameCreator |
| 00:11 | <Hixie> | file a bug or drop me a mail or something? |
| 00:11 | <Hixie> | i'm in the middle of massive edits for datagrid |
| 00:11 | <Philip`> | IE8 is buggy because it allows you to create new properties, in violation of the spec! |
| 00:11 | <Dashiva> | k |
| 00:11 | <Dashiva> | Philip`: Or just prescient! |
| 00:12 | <Dashiva> | Hixie: Can I send to whatwg list even though it's w3c spec? :) |
| 00:12 | <Hixie> | yes |
| 00:12 | <Hixie> | the web storage spec is a whatwg spec (it's in the same source document as html5), i'm just not currently generating a whatwg copy of the output |
| 00:13 | <Hixie> | i wonder why i don't have the NameCreator annotations |
| 00:13 | <Hixie> | that's odd |
| 00:15 | <Dashiva> | Maybe you wrote the IDL before the supported named properties thing was added to WebIDL, and forgot to update |
| 00:16 | <Hixie> | no, the rest of the spec carefully talks about the named properties stuff |
| 00:16 | <Hixie> | i had a big update day |
| 00:16 | <Hixie> | where i went through and redid all the interfaces to match WebIDL |
| 00:16 | <Hixie> | i just forgot, i guess |
| 00:16 | <Hixie> | i wonder how many more are missing NameCreator |
| 00:17 | <Hixie> | i guess it's pretty rare that you can add keys |
| 01:16 | <annevk2> | Hixie, that's ugly, they should do the same, imo |
| 03:13 | <myakura> | annevk2: html5-diff sec 3.1 says "ruby, rt and rb allow for marking up ruby annotations. " but i think it's not <rb> but <rp> |
| 07:44 | <MikeSmith> | are there any significant number of pages that use <basefont>? |
| 07:44 | <MikeSmith> | is it worth having a validator report it as obsolete? |
| 07:45 | <MikeSmith> | also <isindex> |
| 07:46 | <MikeSmith> | hmm, http://dev.opera.com/articles/view/mama-phrase-block-list/#misc |
| 07:46 | <MikeSmith> | so basefont appears in more pages than ins or del |
| 08:04 | <MikeSmith> | http://dev.opera.com/articles/view/mama-head-structure/#head |
| 08:05 | <MikeSmith> | "ISINDEX element was only found in 63 URLs" [out of 3,000,000+] |
| 08:06 | <MikeSmith> | so definitely not worth reporting isindex |
| 08:06 | <MikeSmith> | but <basefont> is also used in more pages than <acronym>, <s> and <strike> are -- and in around the same order of magnitude as many pages as <tt> is |
| 08:14 | <zcorpan> | what is the implementation status of <script async> and <script defer>? |
| 08:16 | <jwalden> | http://☺.damowmow.com/ hurts my eyes to read |
| 08:18 | <zcorpan> | ugh |
| 08:18 | <zcorpan> | Hixie: hire a designer :) |
| 08:25 | <jgraham> | Philip`: I couldn't really get anyone to do the binding for me since it has to be hard bound with cloth covers and all. I could maybe have got someone to do the printing for me but it seems like a big ask since there is a large volume and it would have to be taken to the binders and so on |
| 08:29 | <jgraham> | Hixie: What is the precedent for having localStorage.getItem("foo") != localStorage.foo when !('foo' in localStorage) |
| 08:40 | <zcorpan> | is spellcheck="" implemented in webkit and/or chrome? |
| 08:54 | <MikeSmith> | zcorpan: <script async> and <script defer> not implemented in WebKit yet |
| 08:54 | <MikeSmith> | https://bugs.webkit.org/show_bug.cgi?id=20710 |
| 08:55 | <MikeSmith> | but that comment there says, "Firefox finished this. It's in the 3.1 betas now." .. with some caveats |
| 08:55 | <MikeSmith> | oh, resolution : "Firefox decided to fire DCL after deferred scripts to faciliate the load pattern mentioned in this bug" |
| 09:01 | <zcorpan> | MikeSmith: so firefox supports async and defer? |
| 09:02 | <Rik`> | zcorpan: I believe it only supports defer |
| 09:02 | <zcorpan> | Rik`: ok |
| 09:14 | <hsivonen> | http://groups.google.com/group/mozilla.dev.platform/browse_thread/thread/89eea1cb2d87999/83b917bfaba2b597 |
| 09:31 | <zcorpan> | wow i didn't know about javascript "foo".sub() and .link() etc |
| 09:37 | <hsivonen> | does DanC's URL spec have a permalink yet? |
| 09:40 | <zcorpan> | http://software.hixie.ch/utilities/js/live-dom-viewer/saved/80 |
| 09:41 | <hsivonen> | zcorpan: from Netscape 2.0, IIRC |
| 09:42 | <hsivonen> | zcorpan: great companions to document.write() |
| 09:43 | <hsivonen> | http://homer.w3.org/~connolly/projects/urlp/raw-file/008373680cae/wah5/draft.html is 403 for me |
| 09:44 | <zcorpan> | the need for Web EcmaScript increases |
| 09:47 | <olliej> | zcorpan: ? |
| 09:52 | <zcorpan> | olliej: i doubt the ecmascript committee are interested in specifying "foo".fontcolor("red") |
| 09:52 | <olliej> | zcorpan: yes because that would belong (if anywhere) in html5 sec |
| 09:52 | <olliej> | spec |
| 09:53 | <zcorpan> | maybe |
| 09:54 | <hsivonen> | do those methods on strings exist in all script contexts in browsers? do they exist in Rhino? |
| 09:54 | <zcorpan> | they exist in the futhark jsshell |
| 09:55 | <olliej> | zcorpan: but it also doesn't make sense for them to really belong on string anyway, as hsivonen said they hark back to netscape2 |
| 09:55 | <olliej> | zcorpan: where i'm not even sure the dom would have actually existed |
| 09:56 | <roc> | it didn't |
| 09:56 | <hsivonen> | olliej: but if interop requires string objects to have those methods, should the spec that specs string object methods spec those methods? |
| 09:57 | <zcorpan> | are they required for web compat or are they just leftovers that noone use? |
| 09:57 | <olliej> | hsivonen: no, because a) the DOM spec could easily say "these additional functions must be present on String.prototype" |
| 09:57 | <olliej> | hsivonen: and b) i didn't even know they existed till now |
| 09:58 | <olliej> | hsivonen: so i doubt they're widely used |
| 09:58 | <jgraham> | They are implemeted in all the standalone shells I have |
| 09:58 | <olliej> | hsivonen: we could ask hixie to check for us :D |
| 09:58 | <jgraham> | So they are really part of ECMAsScript implementation wise |
| 09:58 | hsivonen | gets my first and only JS book from 1996 from the shelf |
| 09:59 | <annevk2> | does IE have them? |
| 10:00 | jgraham | is very happy to drop them if they are shown to not be needed for web-compat |
| 10:00 | <jgraham> | But I suspect that will not be the case unfortunately |
| 10:02 | <hsivonen> | the JS book from 1996 has a section on string methods and it lists stuff like big() and fixed() |
| 10:02 | <hsivonen> | aside: the hello world equivalent in the book used document.write() |
| 10:03 | <olliej> | hsivonen: that's still stupid |
| 10:03 | <hsivonen> | my guess is that existing content from the Netscape 2.0 era might use them |
| 10:03 | <olliej> | hsivonen: and they can still be made into dom defined functions |
| 10:03 | <olliej> | hsivonen: yeah, but i don't think they realistically matter anymore |
| 10:04 | <olliej> | we can even have "4" in our user agent now |
| 10:05 | <olliej> | hsivonen: and back when safari 2.0.4 was release we discovered that 4 in the user agent string was bad |
| 10:05 | <hsivonen> | olliej: good times |
| 10:06 | <olliej> | because we all know that is your user agent includes "4" you support layers, right? |
| 10:06 | <jgraham> | "Javascript: The Definitive Guide" (O'Reilly, 1996) lists them (is that the same book that hsivonen looked at?) |
| 10:06 | <hsivonen> | jgraham: mine is teach yourself JavaScript in a week by Arman Danesh |
| 10:08 | <olliej> | mine is JavaScriptCore.xcodeproj ;) |
| 10:09 | <hsivonen> | olliej: clearly, you need more accurate docs :-) |
| 10:09 | <olliej> | hsivonen: heheh |
| 10:12 | <zcorpan> | so what's the complete list of these methods? |
| 10:13 | <hsivonen> | whoa. a .otf (postscript outlines) version of a font can be one third smaller than a .ttf (truetype outlines) version |
| 10:14 | <hsivonen> | I wonder how SVG fonts fare on file size |
| 10:14 | <annevk2> | has anyone checked if IE supports them? |
| 10:15 | <hsivonen> | zcorpan: anchor, big, blink, bold, fixed, fontcolor, fontsize, italics, link, small, strike, sub, sup |
| 10:15 | <zcorpan> | http://www.google.com/codesearch?hl=en&lr=&q=%28%22%7C%27%29%5Cs%2A%5C.%5Cs%2Afontcolor%5Cs%2A%5C%28+lang%3Ajs&sbtn=Search |
| 10:16 | <hsivonen> | annevk2: it does |
| 10:17 | <annevk2> | nasty |
| 10:17 | <hsivonen> | it seems to me that purging the Web of these methods isn't worth the effort compared to just writing them down in the corner of a spec |
| 10:17 | <annevk2> | true |
| 10:18 | <zcorpan> | http://www.google.com/codesearch?hl=en&lr=&q=%5C.%5Cs%2Abig%5Cs%2A%5C%28+lang%3Ajs&sbtn=Search |
| 10:18 | <jgraham> | Things with "Demo" in the title don't count :) |
| 10:19 | <jgraham> | Where is Philip` when you need him? |
| 10:19 | <zcorpan> | http://www.google.com/codesearch?hl=en&lr=&q=%5C.%5Cs%2Afontsize%5Cs%2A%5C%28+lang%3Ajs&sbtn=Search |
| 10:21 | <annevk2> | putting them in HTML5 makes some sense I suppose if the ECMAScript guys don't do it |
| 10:21 | <annevk2> | it's HTML stuff after all |
| 10:22 | <jgraham> | Shall I start a ECMAScript iki page? |
| 10:22 | <jgraham> | *wiki |
| 10:23 | <annevk2> | yeah, something similar to http://wiki.whatwg.org/wiki/HTTP_Issues |
| 10:29 | <olliej> | !! |
| 10:30 | <olliej> | "foo".blink() exists!! |
| 10:30 | <olliej> | noooooo |
| 10:33 | <hsivonen> | zcorpan: IIRC, simple XLink is supported on XML element nodes whose namespace isn't well-known to Gecko |
| 10:34 | <annevk2> | I wonder why they don't just kill XLink support |
| 10:35 | <hsivonen> | annevk2: my understanding is that killing it is the plan |
| 10:35 | <hsivonen> | annevk2: and restoring special support for MathML is the plan also |
| 10:36 | <annevk2> | might make more sense to get the MathML guys to introduce an href attribute |
| 10:36 | <annevk2> | they seemed willing |
| 10:37 | <hsivonen> | perhaps I should switch my site to OTF fonts and let the theoretical Price users upgrade to Price 7 |
| 10:38 | <hsivonen> | it seems that the OTF rasterizer on Windows sucks less than the TTF rasterizer |
| 10:38 | <hsivonen> | and the OTF fonts are smaller |
| 10:40 | <hsivonen> | Grr. Opera 10 on Windows doesn't support OTF |
| 10:41 | <hsivonen> | oops. my test case is wrong |
| 10:48 | <zcorpan> | still wonder if it's practical to use html <a href> for linking in mathml |
| 11:03 | <annevk2> | so Safari/IE conform to the spec for nonexistent keys on localStorage and Firefox has a bug for localStorage.missing |
| 11:03 | <annevk2> | I guess it's not worth changing anymore then :/ |
| 11:09 | <hsivonen> | zcorpan: I think I've now fixed the answers in the Mozilla Web Author FAQ to be up-to-date |
| 11:09 | <hsivonen> | zcorpan: however, the set of questions might no longer be the most frequent set of questions |
| 11:16 | <jgraham> | Interestingly although the enumeration order for user-defined properties on JS Objects seems to be well defined, the relative order between between host-object properties and user defined properties seems to vary |
| 11:17 | <jgraham> | Although perhaps "interestingly" is the wrong word |
| 11:17 | <olliej> | jgraham: hehehe |
| 11:19 | jgraham | wwould have expected user-defined properties to enumerate last |
| 11:19 | hsivonen | points to topic |
| 11:20 | jgraham | has been n00b'd |
| 11:26 | <jgraham> | http://wiki.whatwg.org/wiki/Web_ECMAScript |
| 11:26 | <jgraham> | Still lots to add of course :0 |
| 11:26 | <jgraham> | :) even |
| 11:27 | <annevk2> | e.g. escaping of attribute values and such maybe? |
| 11:28 | <annevk2> | hmm, maybe not |
| 11:29 | <jgraham> | annevk2: I don't think there is anything clever at all |
| 11:29 | <annevk2> | this stuff is not safe at all |
| 11:29 | <jgraham> | I more meant other things where the ES5 spec and web browsers have disagreements e.g. things that are formally implementation defined but actually rather tightly constrained |
| 11:30 | <jgraham> | annevk2: No |
| 11:31 | <annevk2> | note btw that arguments are optional and infinite |
| 11:32 | <jgraham> | annevk2: What do you mean optional and infinite? |
| 11:32 | <annevk2> | e.g. .link("a","x","xd") gives "a" and .link() gives undefined |
| 11:32 | <annevk2> | "undefined"* |
| 11:32 | <jgraham> | annevk2: Yeah ToString will do that |
| 11:33 | <annevk2> | doesn't extra arguments normally throw? |
| 11:33 | <annevk2> | maybe not |
| 11:33 | <jgraham> | annevk2: I don't think so |
| 11:34 | <annevk2> | I guess I'm confusing ECMAScript and WebiDL |
| 11:34 | <annevk2> | though maybe in WebIDL they don't throw either |
| 11:53 | <Philip`> | jgraham: In bed |
| 12:03 | <jgraham> | Philip`: That's what you said last time I asked the same question |
| 12:04 | <Philip`> | jgraham: I sleep a lot |
| 12:04 | <Philip`> | Actually I don't, but I sleep a lot in mornings |
| 12:06 | <jgraham> | Whereas I actually do sleep a lot but am much less useful so people rarely notice |
| 12:11 | <Philip`> | <script language="javascript">var p1="mai";var p2="lto:";var r="johnny";var d="superbikeschool.co.uk";document.write(("Johnny Haynes").link(p1+p2+r+String.fromCharCode(64)+d));</script> |
| 12:11 | <Philip`> | this.value=this.value.fontcolor(c); |
| 12:11 | <Philip`> | strsearch +=' <div class="bodytext"><strong>The Keyword(s) you searched : </strong></div>'.big() ; |
| 12:12 | <Philip`> | People seem to use these things quite a bit |
| 14:06 | <annevk2> | hsivonen, I took the initiative of replying to adactio's Ariability post |
| 14:24 | <hsivonen> | annevk2: cool |
| 15:25 | <hsivonen> | annevk2: your comment system rejected http://pastebin.mozilla.org/643135 upon Post but not upon Preview |
| 15:27 | <hsivonen> | annevk2: in case you put it in via back door, typo fix: http://pastebin.mozilla.org/643136 |
| 15:34 | <annevk2> | hsivonen, added |
| 15:34 | <hsivonen> | annevk2: thanks |
| 15:46 | myakura | wonders if any browser vendors would support "grid" media features |
| 17:05 | <zcorpan> | hsivonen: http://annevankesteren.nl/2009/04/html5-wai-aria#comment-6756 - the architectural forms link is broken |
| 17:06 | <zcorpan> | um or at least it was broken for me. but i could search for the url in waybackmachine and then it worked |
| 19:47 | <jwalden> | olliej, annevk2, hsivonen, et al: the HTML String.prototype methods are broken, at least in SpiderMonkey, wrt escaping; I don't know whether this is common across implementations or not |
| 19:53 | <jgraham> | jwalden: The behaviour is the same in Squirrelfish and Futhark |
| 19:54 | <jwalden> | win! |
| 19:54 | <jwalden> | :-\ |
| 19:54 | <jgraham> | jwalden: At least it makes the spec and implementation simple :) |
| 20:02 | <annevk2> | jwalden, lets consider it a feature |
| 20:03 | mpilgrim_ | wonders where the hell i'm logged in already |
| 20:04 | <mpilgrim> | aha |
| 20:05 | <annevk2> | we call this hell #whatwg |
| 20:05 | <mpilgrim> | found this while digging for link rel discussions |
| 20:05 | <mpilgrim> | http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2006-September/007301.html |
| 20:07 | <mpilgrim> | "I wonder how open the HTML WG will be with regards to working with the WHATWG and HTML 5, especially now that the 2 specs will share the same namespace. If we don't resolve the incompatibilities, one of the specs will simply be doomed to failure." |
| 20:08 | <mpilgrim> | fast forward to today: http://www.w3.org/2009/04/16-html-wg-minutes.html#item05 |
| 20:09 | <mpilgrim> | and http://www.w3.org/html/wg/tracker/actions/105 |
| 20:09 | <mpilgrim> | "Current status: awaiting input from plh and Steve Pemberton as to next steps, will update the working group with status as it becomes available." |
| 20:11 | <jwalden> | I agree with "one of the specs will simply be doomed to failure" |
| 20:11 | <jwalden> | give you one guess which I mean |
| 20:11 | <jgraham> | Yeah well there is reality and there is W3C poliics |
| 20:12 | <jgraham> | The goal is to not let politics affect reality too much |
| 20:12 | <annevk2> | And there's me buying a round to the #whatwg channel if we give up now and let them have it :p |
| 20:13 | <annevk2> | Though that might not help as before #whatwg it didn't gain much traction either... |
| 20:15 | <Hixie> | if w3c politics could affect reality it wouldn't be such a big deal |
| 20:17 | <mpilgrim> | anyone know when rel=feed was added to the spec? |
| 20:18 | <mpilgrim> | earlier discussion i can find is http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2006-November/008080.html which references text already in the spec |
| 20:18 | <Hixie> | try an svn blame and then do spec archeology :-) |
| 20:22 | <annevk2> | mpilgrim, http://html5.org/tools/web-apps-tracker?from=334&to=335 |
| 20:22 | <mpilgrim> | bless you |
| 20:22 | <annevk2> | mpilgrim, the frontpage of the tracker takes a hidden argument limit with a hidden value -1 |
| 20:22 | <mpilgrim> | oooooh |
| 20:23 | <annevk2> | mpilgrim, it takes a while to complete the request but then you can do text search on the svn log |
| 20:25 | <mpilgrim> | what's the query parameter for that? |
| 20:26 | <mpilgrim> | oh |
| 20:26 | <mpilgrim> | is it "limit"? |
| 20:27 | <mpilgrim> | yes, yes it is |
| 20:27 | mpilgrim | enjoys talking to himself |
| 20:28 | <mpilgrim> | wow, my article just got a whole lot better |
| 20:28 | <mpilgrim> | thanks annevk2 |
| 20:28 | <mpilgrim> | i found checkins for virtually every rel value i want to talk about |
| 20:29 | <annevk2> | :) |
| 20:52 | <tantek> | rel="feed" - interesting, first I've heard of this. mpilgrim, do you think it's a good idea? do you agree with Mark Baker's (theoretical) points he makes in http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2006-November/008080.html |
| 20:54 | <tantek> | mpilgrim, also if you're writing an article about many rel values, perhaps this resource may be of assistance: http://microformats.org/wiki/existing-rel-values |
| 20:57 | <Hixie> | rel=feed was added because of implementations, iirc |
| 20:57 | <Hixie> | (and mark's e-mail i believe got an answer, mostly along the lines of "i'm just doing what browsers do", i imagine) |
| 21:00 | <annevk2> | I thought we added rel=feed because the existing solution became a mess |
| 21:01 | <annevk2> | afaik only Firefox supports it |
| 22:22 | <mpilgrim> | test cases: http://diveintomark.org/tmp/relalternate.html and http://diveintomark.org/tmp/relfeed.html |
| 22:22 | <mpilgrim> | all modern browsers support the former (except google chrome, which has no feed autodiscovery at all) |
| 22:23 | <mpilgrim> | firefox 3 supports rel=feed |
| 22:23 | <Hixie> | sounds right |
| 22:23 | <mpilgrim> | firefox 2 does not support rel=feed |
| 22:24 | <mpilgrim> | opera 9.62 does not support rel=feed |
| 22:24 | <mpilgrim> | safari 4 beta does not support rel=feed |
| 22:24 | <mpilgrim> | IE8 final does not support rel=feed |
| 22:37 | <mpilgrim> | rel=feed originated in the Atom working group in 2005 |
| 22:37 | <mpilgrim> | http://www.intertwingly.net/wiki/pie/PaceDifferentRelValue |
| 22:37 | <mpilgrim> | this appears to be the origin: http://www.imc.org/atom-syntax/mail-archive/msg15042.html |
| 22:38 | <mpilgrim> | my reaction to it at the time: http://www.imc.org/atom-syntax/mail-archive/msg15069.html "Part of my newfound personal definition of a life well-lived is to never again argue about semantics, markup, or the "correct" way to use them. This Pace will break every aggregator on the planet, but then again, so will Atom 1.0 feeds, so... +0." |
| 22:40 | <mpilgrim> | here's annevk2 explaining the benefits of rel=feed to sayrer: http://www.imc.org/atom-syntax/mail-archive/msg15074.html |
| 22:41 | <mpilgrim> | to answer tantek's original question, i think rel=feed is an interesting idea that hasn't panned out |
| 22:41 | <mpilgrim> | only one browser vendor has implemented it in 4 years |
| 22:42 | <mpilgrim> | author awareness is zero |
| 22:42 | <mpilgrim> | AFAIK, rel=feed links are not included in any major CMS's default templates |
| 22:43 | <mpilgrim> | though perhaps Philip` has some concrete data on that |
| 22:45 | <annevk2> | the benefits I mention are very ahum compelling |
| 22:46 | mpilgrim | doesn't know whether "ahum" is a typo or a new word he should look up |
| 22:46 | <mpilgrim> | firefox 3 support appears to be here: http://mxr.mozilla.org/seamonkey/source/browser/base/content/browser.js#2818 |
| 22:46 | <mpilgrim> | it looks like removing support for rel=feed would be a matter of deleting a single case statement in browser.js |
| 22:46 | <annevk2> | apparently it's a word in the language Twi?! -- http://www.websters-online-dictionary.org/translation/Twi/ahum |
| 22:46 | <mpilgrim> | and removing the associated tests cases |
| 22:47 | <Philip`> | I see four pages with rel=feed |
| 22:47 | <annevk2> | http://www.urbandictionary.com/define.php?term=ahum is also not that useful |
| 22:47 | <Philip`> | (based on token matches) |
| 22:47 | <mpilgrim> | 4 out of ...? |
| 22:47 | <Philip`> | and two of them are bogus |
| 22:47 | <Philip`> | Out of 130K from dmoz.org |
| 22:48 | <Philip`> | Oh, actually, 7 pages, but 3 are from the same site |
| 22:48 | <mpilgrim> | in the 2 valid cases, does the page also specify rel=alternate links? |
| 22:48 | <Philip`> | s/7/6/ |
| 22:49 | <sayrer> | I dunno |
| 22:49 | <sayrer> | I'm pretty lazy |
| 22:49 | <sayrer> | I think someone made me do rel=feed |
| 22:49 | <sayrer> | and I can at least find some ESPN feeds that use it |
| 22:50 | <mpilgrim> | my question stands: are there a statistically significant number of pages that use rel=feed to the exclusion of other feed autodiscovery mechanisms? |
| 22:50 | <Philip`> | http://www.dip-badajoz.es/municipios/municipio_dinamico/inicio/index_inicio.php?codigo=43 - <link rel="feed alternate" type="application/atom+xml" href="/canales/atom_xml_bop.php?c=1&u=1" title="Último B.O.P."/> |
| 22:50 | <Philip`> | http://www.volkswagen-stiftung.de/ - <link title="News und Termine der VolkswagenStiftung (RSS 2)" href="http://volkswagenstiftung.de/index.php?id=129&type=100"; type="application/rss+xml" rel="alternate feed" /> |
| 22:50 | <Philip`> | http://www.tourismcapetown.co.za/ - <a href="rss20.xml" rel="RSS 2.0 Feed" type="application/rss+xml" title="RSS Feed"><img border="0" src="img/template/rss20.gif" alt="RSS 20 Feed"></a> |
| 22:50 | <Philip`> | http://www.sevillagrande.com/ - <a href="rss.php" rel="rss feed" target="_blank"><img src="images/rss.gif" align="middle" alt="RSS" /></a> |
| 22:51 | <Philip`> | (plus two other pages from that first site) |
| 22:51 | <sayrer> | mpilgrim, it's a good question. I don't know. |
| 22:51 | <mpilgrim> | dip-badajoz.es would match on existing rel=alternate autodiscovery |
| 22:51 | <Philip`> | (This is from searching from parsed 'rel' attributes whose value matches (.*\s)?feed(\s|") (case-insensitive) |
| 22:51 | <Philip`> | ) |
| 22:51 | <mpilgrim> | so would volkswagen-stiflung.de |
| 22:51 | <Philip`> | (Uh, where the " is a delimiter) |
| 22:52 | <Philip`> | (Anyway I think the regexp is about right) |
| 22:52 | <mpilgrim> | tourismcapetown appears to be a typo / misunderstanding that happens to match rel=feed by accident |
| 22:52 | <mpilgrim> | same with sevillagrande.com |
| 22:52 | <sayrer> | oh that's weird |
| 22:52 | <sayrer> | the ESPN ones are Apple <pcast> documents |
| 22:53 | <mpilgrim> | honk. snort. |
| 22:53 | <mpilgrim> | i don't think those are within our scope |
| 22:53 | <mpilgrim> | ...yet |
| 22:53 | <sayrer> | yeah... |
| 22:54 | <mpilgrim> | if this is all the data we have, i'm going to send an email to the list recommending dropping rel=feed |
| 22:54 | <Philip`> | You could recommend getting more data |
| 22:55 | <sayrer> | I would be willing to turn it off in a beta |
| 22:55 | <mpilgrim> | i could |
| 22:55 | <Philip`> | There's non-zero usage of feed, so maybe there's still non-zero usage of feed without alternate, but my sample is far too small to find it :-p |
| 22:55 | <sayrer> | it doesn't seem to generate bug reports, so the cost seems low |
| 22:55 | <mpilgrim> | i could dust off some of hixie's mapreduce code and query against a larger sample |
| 22:56 | <mpilgrim> | sayrer: ok |
| 22:56 | Philip` | would like it if dotnetdotcom.org's data was more accessible (e.g. via EC2) so he could search more pages easily |
| 22:56 | <annevk2> | it looks a whole lot neater than all the cruft you need now, but I guess that's not sufficient |
| 22:57 | <mpilgrim> | annevk2: i'd say that was sufficient in 2005 to propose it |
| 22:57 | <sayrer> | yeah, worth a try, but IE and Firefox match on most other things |
| 22:57 | <mpilgrim> | but implementation has been... minimal |
| 22:57 | <sayrer> | so I can see how authors arrived somewhere else |
| 22:58 | <sayrer> | especially in a world with lower mac market share and no google browser |
| 22:58 | <mpilgrim> | we've cut larger features ;) |
| 22:58 | <mpilgrim> | based on non-implementation |
| 22:58 | <mpilgrim> | or near-zero implementation |
| 22:58 | <annevk2> | some of those we retain as well |
| 22:58 | <mpilgrim> | (no offense to sayrer's work) |
| 22:58 | <sayrer> | i think maybe philor heckled me |
| 22:58 | <mpilgrim> | lol |
| 22:59 | <sayrer> | or it was a way to short circuit autodiscovery |
| 22:59 | <sayrer> | er, sniffing |
| 22:59 | <annevk2> | i.e. <datagrid>, <menu> |
| 22:59 | <sayrer> | did datagrid get cut? |
| 22:59 | <mpilgrim> | annevk2: but it's not a new feature, it's a reformulation of an existing feature |
| 22:59 | <annevk2> | true |
| 23:00 | <mpilgrim> | into a form that virtually no one supports |
| 23:00 | <annevk2> | sayrer, no(t yet) |
| 23:00 | <mpilgrim> | when there's an existing form that virtually everyone supports |
| 23:01 | <mpilgrim> | IIRC, rel=alternate type=application(atom|rss)+xml links were present in 17% of all pages on the internet |
| 23:01 | <mpilgrim> | in 2005 |
| 23:01 | <mpilgrim> | but i'd have to check with hixie on that |
| 23:02 | <sayrer> | things that generate feeds are probably generate lots of pages |
| 23:02 | <sayrer> | er, no "are" needed |
| 23:02 | <mpilgrim> | (actually in 2005, it was probably all rss ;) ) |
| 23:02 | <mpilgrim> | true |
| 23:03 | <sayrer> | mpilgrim, did you see what your site did to AT&T? |
| 23:03 | <mpilgrim> | i wonder if i can do a mapreduce against the (pick a large number) most popular home pages |
| 23:03 | <mpilgrim> | sayrer: no |
| 23:03 | <mpilgrim> | reference? |
| 23:03 | <sayrer> | I had a chuckle, because that was a one of the old examples of broken RSS, and I thought it meant "Accessibility Technology" until I clicked |
| 23:04 | <Philip`> | I see type="application/(atom|rss)+xml" on 9942 out of ~130K pages on dmoz.org |
| 23:04 | <sayrer> | mpilgrim, you have that page that shows all of your internet activity |
| 23:04 | <sayrer> | it was on there |
| 23:04 | <mpilgrim> | ah |
| 23:04 | <mpilgrim> | yeah |
| 23:04 | <mpilgrim> | that's a bug |
| 23:04 | <mpilgrim> | i should fix that |
| 23:04 | <sayrer> | yeah, it's just funny that we still have it |
| 23:05 | <Philip`> | (2674 with atom, 9822 with rss) |
| 23:05 | <sayrer> | clearly ampersands are broken tech |
| 23:05 | <mpilgrim> | philip`: that's about 7% |
| 23:05 | <Philip`> | (But this data is a bit rubbish because it's all skewed by things like Wikipedia) |
| 23:05 | <mpilgrim> | damn, that whole firehose page is a complete hack |
| 23:05 | <mpilgrim> | yeah |
| 23:05 | <mpilgrim> | that's why i want to query just home pages |
| 23:05 | <sayrer> | the microformats religion supports that theory |
| 23:05 | <mpilgrim> | surely we can do that |
| 23:05 | <mpilgrim> | we're google |
| 23:06 | <sayrer> | if IE implements an HTML5ish parser, maybe that will work |
| 23:07 | mpilgrim | wanders off in search of food |
| 23:48 | <Rik`> | in what element should meta information about a blog post go (dates, tags, author name, number of comments, stuff like that) ? nav, aside, anything else ? |
| 23:49 | <Hixie> | <p> |
| 23:51 | <sayrer> | surely dates should go in <time> |
| 23:52 | <Hixie> | should? no. may, sure. |
| 23:52 | <Rik`> | sayrer: of course, I was just talking about the "meta block", not the individual elements inside |
| 23:52 | <Hixie> | <footer> is probably the right element for most of the stuff that nobody cares about |
| 23:52 | <sayrer> | oh, good luck with your meta block |
| 23:53 | <Rik`> | Hixie: so <aside> is better suited for footnotes maybe ? |
| 23:53 | <Hixie> | <aside> is for, well, asides. sidebars. side notes. footnotes too, though they wouldn't be actual footnotes if they were in an <aside>. |
| 23:54 | <Rik`> | ok thanks |
| 23:56 | <Rik`> | i bet will see a lot of different use for those new sectioning elements |
| 23:56 | <sayrer> | I wonder if they are meta crap |