13:17 | <gsnedders> | http://geoffers.uni.cc/archives/2007/04/07/why-xhtml-as-texthtml-is-evil/ – anyone want to check I haven't made some really stupid mistake in that? |
13:21 | <hsivonen> | s/as HTML is a SGML application/as HTML 4 purports to be an SGML application/ |
13:22 | <hsivonen> | "This is rendered as such in a browser that the first <img> and the final “>” is displayed." umm. not in the real non-SGML world |
13:23 | <hsivonen> | s/reference engineering/reverse engineering/ |
13:23 | <hsivonen> | "a few UAs that still use SGML parsers, such as feed readers implementing autodiscovery" Huh? |
13:24 | <Dashiva> | Might want to point a reference on that one |
13:24 | <ajnewbold> | s/(*.)/xhtml sucks/ |
13:24 | <hsivonen> | "thereby not being a delta specification, based upon SGML" the comma seems to give the sentence exactly the wrong meaning |
13:24 | <Dashiva> | I thought the / in the tags would appear as text, but that could be my mistake |
13:24 | <Philip`> | s/\(\*\.\)/(.*)/ |
13:24 | <ajnewbold> | Philip`: drat |
13:24 | <ajnewbold> | I fail :) |
13:25 | <moeffju[Headway]> | Dashiva, no, the slash is either an invalid attribute name or a NET-enabling marker |
13:25 | <hsivonen> | typo: guaranty |
13:25 | <ajnewbold> | hsivonen: he's british |
13:25 | <ajnewbold> | so it's ok |
13:25 | <ajnewbold> | I, too, prefer "guarantee" |
13:26 | <hsivonen> | gsnedders: well-formed as a concept does not apply to HTML |
13:26 | <hsivonen> | ajnewbold: ok. I wasn't familiar with the alternative spelling |
13:28 | <ajnewbold> | I wish he found an excuse to use "bloody" in there |
13:28 | <gsnedders> | maybe I wrote it a bit too quickly :P |
13:29 | <ajnewbold> | I also wish Safari would properly match the background-color to the color used in the background image of the main section :( |
13:30 | <gsnedders> | ajnewbold: I never got around to doing anything about that |
13:30 | <hsivonen> | my crystal ball tells me you are talking about this: http://hsivonen.iki.fi/png-gamma/ |
13:30 | <ajnewbold> | yeah, that's the issue |
13:30 | <Dashiva> | I sometimes fear that if too many people go writing "xhtml sucks" articles, the weaker ones will be used as an excuse to derail the real issue |
13:31 | <ajnewbold> | Dashiva: yeah, but it would sure save on bandwidth |
13:31 | <ajnewbold> | ;P |
13:32 | <gsnedders> | hsivonen: surely the "this is rendered" in that context refers to the DOM tree? |
13:32 | <Philip`> | The background colours don't match in Opera either, at least if you zoom in/out |
13:32 | <hsivonen> | gsnedders: well, you might want to emphasize that it is a hypothetical situation |
13:32 | <hsivonen> | gsnedders: you do mention browsers |
13:35 | <gsnedders> | hsivonen: "How then do real world browsers create the same DOM tree in both parsing modes?…" doesn't that paragraph cover that? |
13:35 | <hsivonen> | does anyone remember off-hand how <legend>foo <ul><li>bar</li><ul></legend> parses as HTML? |
13:36 | <hsivonen> | gsnedders: yes, but I considering how easy it is to start bogus rumors, I suggest emphasizing the hypothetical nature of the example |
13:41 | <Dashiva> | Heh |
13:42 | <Dashiva> | Someone is blasting anne's mail about comments because "they are not compatible with current or previous UAs" |
13:47 | <gsnedders> | hsivonen: emphasising it how? |
13:49 | <gsnedders> | hsivonen: "(thereby in the real world making it irrelevant)"? |
13:53 | <hsivonen> | "if brewsers used real SGML parsers" or something to that effect |
13:53 | <hsivonen> | browsers |
15:44 | <annevk> | hsivonen, that becomes html > [head, body > legend > ["foo", ul > [li > "bar", ul]]] |
15:44 | <annevk> | if you meant to write </ul> the syntax would match the tree |
15:45 | <annevk> | (per html5lib anyway) |
15:54 | <hsivonen> | annevk: excellent. thanks |
15:54 | <hsivonen> | annevk: the result means that struct-inline works in <legend> as text/html |
15:55 | <hsivonen> | annevk: and yes, I meant </ul> |
15:56 | <annevk> | someone should just make some simple applets and document <applet> |
15:58 | <annevk> | I thought of doing it, but I'm not sure I want to promote client-side Java usage |
16:01 | <hsivonen> | annevk: I don't want to promote applet usage, either, but I don't want conformance checking to implicitly promote <object> contortions, either. |
16:01 | <webben> | What's our reason for wishing to discourage applet usage? |
16:02 | <webben> | s/applet/client-side Java/ |
16:03 | <hsivonen> | webben: client-side Java is fine as far as I am concerned, but applets suck really badly |
16:03 | <webben> | i thought all client-side Java was applets? |
16:03 | <webben> | or do you mean <applet> sucks? |
16:04 | <hsivonen> | webben: there's Web Start and there are permanently installed apps |
16:04 | <hsivonen> | webben: I meant Java applet suck no matter how you embed them |
16:04 | <webben> | ah okay. i thought <applet> relied on Web Start |
16:05 | <hsivonen> | it doesn't |
16:05 | <hsivonen> | predates Web Start by several years |
16:05 | <webben> | I see. |
17:20 | <annevk> | hsivonen, right, <object> contortians? |
17:20 | <annevk> | <embed type=application/java code=foo.class> |
18:03 | <hsivonen> | annevk: the Sun <object> examples are rather ugly compared to <applet> |
19:58 | <annevk> | hsivonen, eventually one of those <object> cases will have to work cross browser |
20:05 | <annevk> | What's IPTC? |
20:05 | <hsivonen> | annevk: isn't that the pre-XMP way of putting metadata in bitmaps in Photoshop? |
20:05 | <annevk> | see http://www.w3.org/mid/4617ECB7.4050308⊙ac |
20:06 | <annevk> | International Press Telecommunications Counsel |
20:07 | <hsivonen> | seems to be the organization that defined the metadata field that Photoshop uses |
20:07 | <hsivonen> | so I wasn't totally off-mark |