10:31 | <krijnh> | Pint |
11:39 | <met_> | whow, Gareth negative vote was removed |
14:11 | hsivonen | notes that that the current spec doesn't specify expansions for HTML5 or XHTML5. |
14:11 | hsivonen | notes that RELAX NG isn't defined to be an abbreviation |
14:12 | <annevk> | https://bugzilla.mozilla.org/show_bug.cgi?id=147856 shows that authors are confused with how CSS works in XHTML |
14:12 | <hsivonen> | so far, I've marked HTML5 and XHTML5 as <abbr> but not RELAX NG |
14:13 | <annevk> | for the former, me too |
14:13 | zcorpan_ | thinks the expansion of XHTML5 should be "XML serialization of HyperText Markup Language 5" :) |
14:13 | <annevk> | for the latter, dunno |
14:13 | <annevk> | XHTML = XML serialization of HTML |
14:14 | <annevk> | HTML = HTML language and / or HTML syntax |
14:14 | hsivonen | just wrote Extensible HyperText Markup Language 5 in his glossary |
14:14 | <hsivonen> | Hixie: should I revise? |
14:15 | <Dashiva> | Then there's the battle over (XHTML)5 vs X(HTML5) |
14:15 | <hsivonen> | Dashiva: my glossary has "(X)HTML5" ;-) |
14:16 | <zcorpan_> | "(XML serialization of) HTML5" |
14:16 | <zcorpan_> | :) |
14:47 | <Philip`> | Hixie: http://forum.doom9.org/showthread.php?t=122363 seems a useful explanation - there is a secret fixed set of possible processing keys, and each can be 'reached' (in some mathematical way I don't quite understand) from a different subset of all device keys; so you can revoke compromised devices by choosing a (small) set of processing keys such that none can be reached from any of the device keys known to any compromised device. |
14:47 | <Philip`> | Once you know a processing key that's used on a disc (which is typically the same for all discs), you can combine that with some other disc-specific data to end up with the keys that decrypt the actual data; at least until new discs are manufactured with new processing keys, in which case you have to start all over again and try to discover the new keys. |
14:48 | <Philip`> | (...or at least that's my understanding - it looks like a not entirely trivial system) |
14:56 | <Dashiva> | Security through really many keys |
15:13 | <Philip`> | At least it's a readable published specification, and the security depends entirely on the keys being kept secret, which is how cryptography ought to work - the problem is that it's impossible to keep the keys secret, and it has statements like "A device shall keep the value s_0 confidential, as defined in the license agreement" which doesn't work quite so well when millions of customers have your devices and want to extract the secret values |
15:14 | <Dashiva> | As the aacs people say, the system hasn't been breached, it's key leakage. They just leave out the 'inevitable'. |
15:14 | <Philip`> | So apart from being fundamentally flawed, it seems quite good :-) |
15:55 | <hsivonen> | Philip`: like HTML5?-) |
16:12 | <hsivonen> | any interest in reviewing my glossary? http://hsivonen.iki.fi/thesis/html5-conformance-checker.xhtml#glossary |
16:14 | <hasather> | hsivonen: American Standard Code of Information Interchange, s/of/for/? |
16:14 | <Dashiva> | You don't define conforming, but you define both valid and validation |
16:14 | <hsivonen> | hasather: right. thanks |
16:15 | <zcorpan_> | Hypertext vs. HyperText |
16:15 | <hasather> | hsivonen: IANA: "... maintains regisries of Internet media types" |
16:16 | <hasather> | regisries is wrong |
16:16 | <zcorpan_> | not sure which it should be (for the expansion of HTML) |
16:21 | <hsivonen> | Dashiva: defined. |
16:21 | <hsivonen> | zcorpan_: fixed. HyperText for HTML and XHTML. Hypertext for HTTP and WHATWG. |
16:21 | <hsivonen> | hasather: typo fixed |
16:21 | <hsivonen> | thanks |
16:22 | <hsivonen> | oops. the explanation for GML is missing |
16:25 | <hasather> | hsivonen: is RELAX NG ever expanded as "... New Generation". I haven't seen that in any official docs |
16:26 | <hsivonen> | hasather: neither the OASIS nor the ISO spec defines RELAX NG as an abbreviation |
16:26 | <hsivonen> | of either New Generation or Next Generation |
16:27 | <hasather> | ok, so is that bogus? Did someone just make it up? Or did it disappear sometime? |
16:27 | <hsivonen> | hasather: Wikipedia says Next |
16:27 | <hsivonen> | hasather: it is pretty obvious that NG comes from Next Generation even if unofficially |
16:28 | <hsivonen> | or perhaps New |
16:31 | <hasather> | hsivonen: yea, it's pretty obvious what it stands for, but kinda weird that it's not used in any of the official docs |
16:32 | <hsivonen> | hasather: yeah. it's weird that HTML5 isn't expanded, either |
16:33 | <hasather> | it is in your glossary though :) |
16:34 | <hsivonen> | hasather: which is why I asked Hixie if I need to revise :-) |
16:35 | <hsivonen> | btw, my generator works in the opposite direction than usual |
16:35 | <hsivonen> | it doesn't collect title attributes |
16:35 | <hsivonen> | but it collects dl definitions and assigns titles based on that |
16:36 | <hasather> | hsivonen: ah, cool |
16:38 | <hsivonen> | hasather: hence the seemingly extra spans |
16:38 | <hsivonen> | in the definitions, that is |
16:45 | <Philip`> | "UTF-32: ... a one-to-one mapping between characters and code units" - I think "code point" would be more accurate than "character" |
16:48 | <hsivonen> | Philip`: ok. though characters and code points are one-to-one |
16:50 | <othermaciej> | I don't think "code point" is accurate |
16:51 | <othermaciej> | because there is a one-to-one mapping between characters and code points even in UTF-8 |
16:51 | <othermaciej> | the code point is the unique numeric identifier of the character |
16:51 | <othermaciej> | what's different about UTF-32 is that it always takes exactly one unit of the encoding to identify a code point |
16:52 | <othermaciej> | whereas for UTF-8 it's 1-5 and for UTF-16 it's 1-2 (I think) |
16:59 | <Philip`> | hsivonen: I can't find any precise definition of "character" in Unicode at all - their glossary gives four different meanings, including "abstract character" which is defined as being possibly multiple code points (it "can often be represented by the use of combining character sequences") |
16:59 | <Philip`> | othermaciej: I intended "between code points and code units" (rather than "between characters and code points") |
17:07 | <hsivonen> | hmm. I need to add entries for document tree and infoset |
17:08 | <hsivonen> | Added character-related entries |
17:25 | <Philip`> | "UTF-32: ... a character encoding a single ..." -> "a character encoding in which a single ..." |
17:26 | <hsivonen> | Philip`: uploading a fix right now |
17:28 | <Philip`> | "XML parsing failed: syntax error" - hmm, I'm unsure whether it's good or bad that that happens with partially-uploaded files |
17:28 | <Philip`> | I suppose it's good that it makes it obvious there's a problem |
17:28 | <Philip`> | (given that the problem will be fixed shortly) |
17:28 | <hsivonen> | now fully uploaded |
17:33 | <Philip`> | Looks good :-) |
17:33 | <Philip`> | Unrelated to the glossary, in the section headings for 5.10 and 8.10, there seems to be insufficient space between the 0 and the words |
17:34 | <hsivonen> | Philip`: which format? which UA? |
17:35 | <Philip`> | XHTML and HTML, in Opera and Firefox |
17:35 | <hsivonen> | Philip`: do you have Palatino or Palladio available? |
17:36 | <hsivonen> | Philip`: firefox 2 or 3 nightly? |
17:36 | <Philip`> | Ah, neither - I guess the different font would cause it to mess up |
17:37 | <hsivonen> | works for me in both Firefox and Opera on Mac with Palatino available |
17:37 | <hsivonen> | Philip`: does your serif font have unusually wide digit glyphs? |
17:37 | <Philip`> | It breaks in FF3, works in FF2, but that's just because they have different fonts selected |
17:38 | <Philip`> | Looks like it's DejaVu Serif that causes problems |
17:38 | <hsivonen> | Philip`: Firefox 2.0 gets different layout because it doesn't support inline-block |
17:38 | <hsivonen> | Philip`: I'm going to resolve as WONTFIX |
17:38 | <hsivonen> | Philip`: but thanks |
17:39 | hsivonen | is optimizing for Prince |
17:40 | <Philip`> | Okay, sounds sensible :-) |
19:43 | <Dashiva> | +1 (ducks) |
19:43 | <Dashiva> | I didn't know we had ducks |
19:44 | <bewest> | yeah, they're cute and will follow you around. also they are yellow. |
19:44 | <Philip`> | Most ducks just run away from me :-( |
20:22 | <met_> | reading http://www.evardsson.com/blog/2007/05/02/looks-like-whatwgs-html-5-is-a-go/ it's true only 110 members of html wg can vote? or all 384 members? |
20:23 | <annevk> | companies can vote once |
20:24 | <met_> | looking at http://www.w3.org/2000/09/dbwg/details?group=40318&public=1 there are >300 invited experts so >300 votes, isn't? |
20:24 | <Philip`> | He's probably counting the 90 voters plus the 20 in the non-responders list |
20:24 | <Philip`> | (which won't quite work since the non-responders list is broken) |
21:30 | <Hixie> | hsivonen: i personally prefer "XML serialisation of...", but I don't believe I or the WHATWG has ever taken an official stance on the matter before now |
21:31 | <hsivonen> | Hixie: ok. |
21:32 | <hsivonen> | Hixie: does HTML5 stand for HyperText Markup Language 5? that is, is it still an abbreviation? (I hope it is) |
21:33 | <Hixie> | sure |
21:34 | Hixie | doesn't really have an opinion either way :-P |
21:36 | <hsivonen> | the fix to the glossary should be live now |
21:36 | <hsivonen> | Hixie: thanks |
23:30 | <Hixie> | annevk: btw, i don't even remotely intend to go through and read every piece of feedback sent to public-html. if you see specific things that should be taken into account, please feel free to forward them to me (or the whatwg list, or forward a pointer to the whatwg list) so i can make sure to take them into account. |
23:42 | <Hixie> | wow, anne actually caused a spec to not go through the CR transition call |
23:42 | <Hixie> | i'm impressed |
23:43 | <othermaciej> | really? how? |
23:43 | <hober> | pointer? |
23:43 | <othermaciej> | and which spec? |
23:44 | <Hixie> | http://lists.w3.org/Archives/Member/w3c-css-wg/2007AprJun/0146.html |
23:44 | <hober> | behind the member-wall :( |
23:45 | <hsivonen> | cool! |