| 18:23 | <nickshanks> | is anyone likely to be doing a survey of current element usage on the web? |
| 18:39 | <MikeSmith> | nickshanks - you've seen http://code.google.com/webstats/ ? |
| 18:40 | <nickshanks> | mike: yes, i am asking if ian or someone like him is going to do an update |
| 18:42 | <nickshanks> | i want to add <image></image> to HTML, but people have said that you can't do that, as sites already use that when they mean <img>. But no-one's ever provided any stats to back this up |
| 18:42 | <nickshanks> | and ian said he was too busy last time i asked (about mid-2006) |
| 18:43 | <MikeSmith> | nickshanks - I see. I think he's since run some one-off reports for specific things. maybe if you remind him |
| 18:43 | <MikeSmith> | Philip` runs some now and then |
| 18:43 | <MikeSmith> | but on a smaller set of pages |
| 18:44 | <nickshanks> | how big a sample? where does he get the sample from? |
| 18:44 | <MikeSmith> | At least some of them from dmoz.org I think |
| 18:44 | <MikeSmith> | not sure how big a sample |
| 18:45 | <MikeSmith> | Shawn Medero (smedero) from the HTML WG has access to a very big index and has talked about setting up some reporting mechanism for that |
| 18:45 | <nickshanks> | anyway, i think adding <image></image> and deprecating <img alt=""> would be a step forward, even if only in standards mode |
| 18:46 | <nickshanks> | hmm, interesting. i'm not familiar with him |
| 18:46 | <MikeSmith> | smedero over on #html-wg on irc.freenode.net |
| 18:46 | <MikeSmith> | err |
| 18:46 | <MikeSmith> | I mean |
| 18:46 | <MikeSmith> | irc.w3.org |
| 18:46 | <nickshanks> | yep :) i guessed |
| 18:47 | <MikeSmith> | but definitely worth asking Philip` about it too if you can catch him |
| 18:48 | <nickshanks> | do you think this would be a good idea? i draw parallels with the resently introduced <video> and <audio> elelments. previously the idea was shot down by people saying "why not just use <object> ?" but now i have a counterpoint to that |
| 19:05 | <Lachy> | nickshanks, <image> is already handled by the browsers exactly the same as <img>. It can't be changed without breaking back compat |
| 19:05 | <nickshanks> | yeah, i know. i still want it changed in standards mode. it can stay the same in tag soup mode |
| 19:10 | <nickshanks> | anyway, will have to come back later |
| 19:12 | <annevk> | yeah, we're not going to do that |
| 19:16 | <annevk> | 2 promille of the page out there are using <image> and introducing more standards mode / quirks mode differences is silly |
| 19:17 | <annevk> | s/page/pages/ ... |
| 20:06 | <Philip`> | nickshanks (if you see this message): See http://www.imdb.com/ for an example of <image> on a site which browsers probably don't want to break |
| 20:09 | <Philip`> | http://canvex.lazyilluminati.com/survey/2007-07-17/analyse.cgi/pages/tag/image lists where I found <image> in ~8K pages |
| 20:10 | <annevk> | interesting, that matches the 2 promille figure |
| 20:10 | <annevk> | even though this sample is much smaller, hmm |
| 20:20 | Lachy | wonders if "promille" is supposed to mean "permille" |
| 20:26 | annevk | wonders why promille is Dutch-specific |
| 20:27 | <annevk> | anyway: ‰ |
| 20:27 | <annevk> | oh, seems Finnish and German have it too |
| 20:28 | <webben> | I thought "standards mode" and "tag sop mode" where the same thing in HTML5? |
| 20:28 | <webben> | *soup |
| 20:28 | <webben> | (in terms of HTML parsing anyhow) |
| 20:28 | <annevk> | there's quirks mode, limited quirks mode and no quirks mode in HTML5 |
| 20:29 | <annevk> | parsing currently only deals with no quirks mode, although it does do DOCTYPE processing already |
| 20:29 | <webben> | oh okay |
| 20:31 | <Philip`> | annevk: Why should the smaller sample size make it not match? :-) |
| 20:32 | <webben> | <picture> is probably a more promising line of inquiry, given it has no legacy |
| 20:33 | <annevk> | doing away with <img> would be silly imo |
| 20:33 | <annevk> | and there's always <object> |
| 20:33 | <annevk> | otherwise |
| 20:33 | <webben> | Does object for images actually work well enough to be used? |
| 20:33 | <kig> | promille is from latin, permille is the english bastardization (i guess) |
| 20:33 | <webben> | or does it require conditional comments to make it work? |
| 20:34 | <Philip`> | (I think 15 out of 7739 gives standard deviation around 4, so it's too small to get an especially accurate figure out) |
| 20:34 | <Philip`> | (*too small a sample) |
| 20:35 | <webben> | also: could one specify an API for images included via OBJECT? |
| 20:35 | <Philip`> | webben: See http://canvex.lazyilluminati.com/survey/2007-07-17/analyse.cgi/pages/tag/picture for <picture> |
| 20:35 | <webben> | Philip`: I'm more worried about what UAs do or don't do with it. |
| 20:35 | <Philip`> | <!--webbot bot="PhotoAlbum" U-Include="photogallery/photo18322/real.htm" clientside TAG="BODY" startspan --> <picture file-href="photogallery/photo18322/real_p.htm" /> |
| 20:36 | <webben> | Philip`: but thanks for the link :) |
| 20:36 | <Philip`> | Argh, second link has Java and crashed Opera :-( |
| 20:36 | <annevk> | apart from getImageData() I'm not sure what API you'd want |
| 20:37 | <annevk> | and even getImageData() is not really needed given that there's drawImage() |
| 20:37 | <Philip`> | <!--webbot bot="PhotoAlbum" U-Include="photogallery/photo25215/real.htm" clientside TAG="BODY" startspan --> <picture file-href="photogallery/photo25215/real_p.htm" /> |
| 20:38 | <Philip`> | Looks like both pages using <picture> get it via the same PhotoAlbum software |
| 20:39 | <webben> | annevk: Can getImageData() be used with OBJECT? |
| 20:39 | <webben> | (how large the API is less of a question than whether having an image-specific API is even possible for OBJECT) |
| 20:39 | <webben> | IIRC one of the rationale's for forking OBJECT to create VIDEO was that it made providing a dedicated API easier. |
| 20:40 | <Philip`> | http://support.microsoft.com/kb/813736 - ah, those <picture>s are from Microsoft FrontPage |
| 20:40 | <annevk> | for images, a dedicated API on <object> is less of an image, but this is all highly theoretical anyway |
| 20:40 | <annevk> | less of a problem, even :) |
| 20:40 | <webben> | annevk: It's not theoretical. |
| 20:40 | <annevk> | it is, there's no proposed API for images other than <canvas> |
| 20:41 | <annevk> | and <img> already works pretty damn well |
| 20:41 | <annevk> | compared to using <object> for video that is |
| 20:54 | <kig> | annevk, Philip`: re: canvas spec, i'll mail some comments on the ml, based on writing an svg renderer on canvas |
| 20:55 | <annevk> | cool |
| 20:55 | <annevk> | hopefully hixie will take another look at again as well |
| 21:33 | <G0k> | hey all |
| 21:34 | <G0k> | how's the party? |
| 21:35 | <takkaria> | "not on IRC" |
| 21:36 | <G0k> | ouch, cut my heart out why don't you |
| 21:39 | <takkaria> | :P |
| 21:39 | <takkaria> | hey, I'm on IRC too |