| 00:00 | <zcorpan_> | running together with subsequent attributes seems fairly common |
| 00:01 | <Philip`> | Ooh, it finished |
| 00:01 | <Philip`> | http://philip.html5.org/data/attribute-numbers-followed-by-af.txt - now with all the matches on 130K pages |
| 00:02 | <zcorpan_> | thanks a lot |
| 00:04 | Philip` | notes that he made it case-insensitive, and added \s at the beginning because otherwise it picked up lots of lines of JS code like thing.style.width='10px' etc |
| 00:10 | <gsnedders> | takkaria: "Decriminalise incest and prostitution." — trying to be controversial? :P |
| 00:26 | <takkaria> | no, I just don't think the state should be moralising |
| 00:28 | <Hixie> | (incest marriage or incest conception? the latter isn't a moral matter but a health matter.) |
| 00:30 | <takkaria> | the chance of having congenital defects if born from a incestuous couple is only a tiny bit larger than non-incestous couples, AIUI |
| 00:31 | webben | thought the point of the state was to be demoralizing. ;) |
| 00:36 | <Philip`> | takkaria: Maybe it's only a tiny bit larger for the first generation, but it compounds if you carry on several times? |
| 00:37 | <webben> | Philip`: I wonder if that's not unlikely in the absence of socio-political structures actively encouraging incestuous unions? |
| 00:37 | <webben> | (e.g. structures like inheritance regimes where you're trying to keep property or power within a given kin-group) |
| 00:38 | zcorpan_ | -> bed |
| 00:40 | <VeXocide> | hi, besides the html5lib python and ruby implementation, i was wondering if there was one in C++ |
| 00:40 | <Hixie> | takkaria has one, i believe, and hsivonen is working on converting his java implementation to C++ |
| 00:41 | <VeXocide> | nice, and are they open source, because if so, it'll mean a hell of a lot less work for me :) |
| 00:41 | <Hixie> | i believe so |
| 00:41 | <Hixie> | takkaria, hsivonen? |
| 00:41 | <jmb> | yes, and yes |
| 00:42 | <Hixie> | takkaria: the chances of defects go up significantly with each incestuous generation, as i understand it (just look at the european monarchies of the last millenia) |
| 00:43 | <VeXocide> | any links, and this means i can stop trying to do it myself, which is probably a good thing |
| 00:43 | <jmb> | http://about.validator.nu/htmlparser/ |
| 00:43 | <Philip`> | VeXocide: http://www.netsurf-browser.org/projects/hubbub/ |
| 00:44 | <takkaria> | Hixie: right, but I find it unlikely that that will happen |
| 00:45 | <gsnedders> | Hixie: Right, but I think at the state of the first generation it is fairly low |
| 00:45 | <Hixie> | i'm not disagreeing with that, i was just saying that it wasn't a moral issue |
| 00:45 | <gsnedders> | takkaria: What about the age of consent as well, then :P |
| 00:46 | <gsnedders> | (Now that really is a stupid law. Any law which the vast majority break really should be reconsidered.) |
| 00:46 | <takkaria> | gsnedders: I would like to see a study on the possible effects of lowering age of consent by a couple of years before saying anything firm on that |
| 00:46 | <gsnedders> | takkaria: I expect it would make pretty much no difference whatsoever |
| 00:47 | <Philip`> | Hixie: It's still a moral issue, because you're balancing the health of future generations and the desires of current generations |
| 00:47 | gsnedders | is listening to Desire by U2 from Rattle And Hum |
| 00:47 | <Hixie> | Philip`: well, insofar as any law is a moral issue, sure |
| 00:48 | <gsnedders> | takkaria: I mean, currently here there are two different legal statuses. One is statutory rape (< 13), and the other is the lesser charge of underage sex. |
| 00:49 | <gsnedders> | takkaria: I wonder how many people are actually convicted of the latter charge, and what their mean age is |
| 00:49 | <VeXocide> | and there actually is some documentation, how lovely \o/ |
| 00:49 | gsnedders | has also totally dragged the channel off-topic, again |
| 00:50 | <Hixie> | yeah how dare you take us away from whining about the w3c process |
| 00:51 | <gsnedders> | There again, expecting the channel to be on-topic would be logical |
| 00:51 | <gsnedders> | Therefore, being off-topic is to be expected |
| 00:51 | <Hixie> | anyone want to volunteer to find out which properties should be [Replaceable] ? |
| 00:51 | <VeXocide> | haha, thanks a bunch Philip`, gonna have a serious look at hubbub |
| 00:52 | <takkaria> | gsnedders: freedom of information act could tell you that :) |
| 00:52 | <takkaria> | VeXocide: any questions direct at me or jmb, we wrote it :) |
| 00:53 | <takkaria> | probably me, since volunteering jmb probably isn't appreciated ^_^ |
| 00:53 | jmb | disowns all responsibility :P |
| 00:53 | <gsnedders> | takkaria: That means putting in a request in all likelihood, which means bothering |
| 00:53 | <gsnedders> | takkaria: Instead of just randomly wondering on IRC |
| 00:54 | <takkaria> | true. I suspect that child protection laws are probably strong enough to drop the age of consent a couple years without any ill effects |
| 00:55 | <gsnedders> | takkaria: Rape would still apply as it does now, so it would make little difference |
| 00:55 | <VeXocide> | takkaria, is there any simple implementation which reads a file, parses it, and outputs it in some form ? |
| 00:55 | <gsnedders> | takkaria: Other issue with getting data: Scotland is seperate |
| 00:55 | <takkaria> | VeXocide: yeah, the test programs do |
| 00:57 | <gsnedders> | takkaria: I also have no idea how to actually make such a request :P |
| 00:59 | <VeXocide> | great, then i'm going to bookmark this, and actually catch up on sleep |
| 01:00 | <jmb> | takkaria: I suggest we take NetSurf's hubbub_binding.c and turn it into an example driver for hubbub |
| 01:01 | <takkaria> | gsnedders: just send a request to the home office, or the equivalent in scotland |
| 01:02 | <takkaria> | jmb: yeah, not a bad plan. I could get onto that tomorrow morning |
| 01:02 | <gsnedders> | takkaria: Not the justice ministry? |
| 01:02 | <VeXocide> | well, if there is any way of taking a piece of possibly borked html input, and outputting a clean tree, i'd be very happy |
| 01:02 | <takkaria> | gsnedders: maybe them |
| 01:03 | <takkaria> | VeXocide: you mean, serialised? |
| 01:03 | <Philip`> | VeXocide: Do you need a real HTML5 parser, or could something like libxml2's HTML parser work too? |
| 01:04 | <takkaria> | we don't serialise, but you can get a tree easily enough |
| 01:04 | <VeXocide> | takkaria, preferably, but i can do serialisation myself |
| 01:04 | <VeXocide> | fair enough |
| 01:04 | <jmb> | implementing the serialisation algorithm is on the todo list. somewhere, anyway :) |
| 01:04 | <gsnedders> | http://www.scottish.parliament.uk/vli/language/scots/index.htm — awesome. |
| 01:05 | <gsnedders> | I have never seen anything so format in Scots. |
| 01:05 | <takkaria> | jmb: I guess we'd have to have a vtable for the tree access functions in the serialising code, since we can't rely on any particular implementation |
| 01:05 | <jmb> | takkaria: yeah |
| 01:06 | <jmb> | takkaria: basically the inverse of the treebuilder, I guess |
| 01:06 | <VeXocide> | Philip`, it's mostly the borked html which causes problem, we currently have a tokenizer and parser which works fine but there are seemingly infinately many corner cases |
| 01:06 | <gsnedders> | I'm not sure where in Scotland would be relevant |
| 01:07 | <VeXocide> | taking clean html or xml and turning it into a tree is relatively easy compared to parsing the mess out there on the interwebs |
| 01:08 | <Philip`> | VeXocide: libxml2 says it "should be able to parse "real world" HTML, even if severely broken from a specification point of view", though I have no idea how well it works in practice, and HTML5 is much more likely to be similar to what web browsers do |
| 01:08 | <VeXocide> | Philip`, the reason i found html5 interesting is because the spec describes a complete tokeniser |
| 01:08 | <gsnedders> | It doesn't that well |
| 01:08 | <jmb> | Philip`: badly |
| 01:08 | <gsnedders> | (libxml2) |
| 01:08 | <Philip`> | Ah |
| 01:08 | <jmb> | Philip`: if it worked, hubbub wouldn't exist |
| 01:08 | <jmb> | (NetSurf used to use libxml's parser) |
| 01:08 | <VeXocide> | that was basically my conclusion |
| 01:09 | <VeXocide> | libxml2 works fine if the input is nice, but it will mess up under pressure |
| 01:09 | <jmb> | takkaria: feel free to relicense the result under MIT or somesuch -- only you and I have done anything significant to that file |
| 01:10 | <VeXocide> | we're basically looking for a tree which describes the html as well as it can, but one we can rely upon to be good |
| 01:11 | <takkaria> | jmb: alright |
| 01:11 | gsnedders | wonders quite how to write the request |
| 01:13 | gsnedders | finds an issue with the Sexual Offences Act 1956 which surely goes against Human Rights legislation |
| 01:13 | <takkaria> | what's that? |
| 01:13 | <gsnedders> | "A man also commits rape if he induces a married woman to have sexual intercourse with him by impersonating her husband." |
| 01:13 | <gsnedders> | That doesn't take into account (and there is no amendment thereof) for same-sex civil partnerships |
| 01:14 | <Philip`> | Is the Act full of useful tips like that? |
| 01:14 | <gsnedders> | http://www.opsi.gov.uk/RevisedStatutes/Acts/ukpga/1956/cukpga_19560069_en_1 if you are curious |
| 01:14 | <gsnedders> | Good reading at 1:20am. |
| 01:15 | <takkaria> | VeXocide: well, libxml2 does alright, but not always and html5 is the better option. hubbub can give you a tree easily, so it sounds like it might be useful to you. it is C, though C++ bindings would be quite welcome |
| 01:15 | <gsnedders> | Section 6 of part one is what is relevant |
| 01:16 | <gsnedders> | Wow, the law is bizarre |
| 01:16 | <gsnedders> | "A man is not guilty of an offence under this section because he has unlawful sexual intercourse with a girl under the age of sixteen, if he is under the age of twenty-four and has not previously been charged with a like offence, and he believes her to be of the age of sixteen or over and has reasonable cause for the belief." |
| 01:16 | <gsnedders> | (it then goes on to define "a like offence") |
| 01:16 | <VeXocide> | takkaria, C is just fine, i basically used C++ because i had a tokeniser, scanner and parser framework for it |
| 01:17 | <VeXocide> | and if i can build and walk a tree in memory, i'd be delighted, serialization should be relatively easy |
| 01:17 | <gsnedders> | The act of buggery is illegal between two men if it happens when more than two person take part or are present, or in a public loo. |
| 01:18 | <gsnedders> | (WTF?) |
| 01:18 | <VeXocide> | actually, if i write a nice serializer i'd probably release it under mit |
| 01:21 | <takkaria> | gsnedders: it was considered that gay men would meet up in public toilets to have sex, tha law was put there to try and stop it |
| 01:21 | <takkaria> | gsnedders: are you sure there's no updated legislation? |
| 01:22 | <gsnedders> | takkaria: There's no modification noted |
| 01:22 | <webben> | Are modifications from subsequent statutes noted? |
| 01:22 | <takkaria> | it could have been superceded, maybe? |
| 01:22 | <gsnedders> | takkaria: And I'm aware of its origins. I just didn't think it still existed. And the first part still seems odd. |
| 01:22 | webben | would have that that such notes could be legally arguable. |
| 01:22 | <gsnedders> | takkaria: That would be noted as a modification. |
| 01:22 | <takkaria> | since they seem like anarchonisms that shouldn't be on the books anymore |
| 01:23 | <gsnedders> | The act in 2000 lowering the age to 16 doesn't touch it |
| 01:25 | <jmb> | VeXocide: hubbub is designed to allow you to use whatever in-memory format you like -- the client has to provide a bunch of callbacks used to build a tree. test/tree.c should give you some idea of how to drive it (though it's not remotely well commented). hopefully takkaria will have something more friendly soon :) |
| 01:25 | gsnedders | sighs |
| 01:25 | <gsnedders> | The entire issue of homosexuality is stupid. |
| 01:26 | <gsnedders> | (I mean in the sense of making it an issue whatsoever.) |
| 01:26 | <VeXocide> | jmb, well, consider me seriously interested and following this code |
| 01:26 | <gsnedders> | (Not homosexuality being an issue itself.) |
| 01:27 | <VeXocide> | and probably committing some stuff if it's usefull |
| 01:27 | <jmb> | VeXocide: cool |
| 01:31 | <VeXocide> | jmb, well, it's nice to see a browser which actually uses a seperate parser |
| 01:31 | <gsnedders> | This act is screwed up and hard to read. |
| 01:31 | <VeXocide> | webkit and gecko need you to initialise a rendering engine to parse html |
| 01:31 | <gsnedders> | The punishment for intercourse with a girl whose age is 13 ≥ age > 16 is two years, I think. |
| 01:32 | <Philip`> | gsnedders: Did you intend that to be a zero-sized range? |
| 01:32 | <gsnedders> | Oh, I've misread stuff. |
| 01:32 | <gsnedders> | I'm looking at the wrong act |
| 01:32 | <gsnedders> | Philip`: How is that zero-sized? |
| 01:34 | <Philip`> | gsnedders: The set of possible values for 'age' which are simultaneously less than or equal to 13, and greater than 16, is quite limited |
| 01:34 | <gsnedders> | Philip`: I typed ≥ and not ≤ |
| 01:34 | <gsnedders> | Or have I totally fucked up? |
| 01:34 | <VeXocide> | don't think so |
| 01:34 | <Philip`> | gsnedders: You said "13 >= age" which is equivalent to "age <= 13" |
| 01:34 | <VeXocide> | then again, i'm falling half asleep |
| 01:35 | <Philip`> | gsnedders: and also you said "age > 16" which means what it says, and is also wrong |
| 01:35 | <gsnedders> | Yes, I have |
| 01:35 | <gsnedders> | 13 <= age < 16. |
| 01:35 | gsnedders | was trying to evaluate from rtl and not ltr |
| 01:35 | <Philip`> | Too lazy to find the Unicode character this time? ;-) |
| 01:37 | heycam | would love to be able to use vim digraphs in os x / gtk input fields |
| 01:37 | <heycam> | or even to be able to escape to a vim to edit the contents of them |
| 01:38 | <gsnedders> | Philip`: No, I was just copying and pasting what I was checking in Python :) |
| 01:38 | <gsnedders> | Philip`: And Py doesn't like ≤ |
| 01:40 | <Hixie> | anyone here write chinese or japanese? |
| 01:40 | <gsnedders> | takkaria: I give up with the 2003 law, I'm just confused by it |
| 01:41 | <jmb> | gsnedders: that's the point of legislation, right? |
| 01:41 | <gsnedders> | jmb: The 1956 act is readable |
| 01:41 | <VeXocide> | i've had a pretty serious look at the python tokenizer from html5lib, and it looks pretty ok |
| 01:41 | <jmb> | gsnedders: provide a barrier to entry to its interpretation :) |
| 01:41 | <jmb> | gsnedders: ah, they knew how to write in those days |
| 01:42 | <VeXocide> | it's just a shame the inputstream is slow, mostly because there aren't any pointers in python ;) |
| 01:42 | <jmb> | gsnedders: these days, noone has the attention span |
| 01:43 | <gsnedders> | It seemingly makes it illegal for someone 18 or older to do any sexual touching (snogging?) of someone under 16 |
| 01:44 | <heycam> | Hixie, what level do you need? |
| 01:44 | <heycam> | i can read/write a very tiny amount of chinese |
| 01:44 | <Philip`> | VeXocide: The input stream doesn't seem to be a huge bottleneck now - it's the whole parser that's slow, which makes it hard to optimise :-( |
| 01:44 | <VeXocide> | Philip`, hmz ok, never got that far |
| 01:46 | <gsnedders> | takkaria: Oh, sex is a public loo is now illegal regardless of who it involves |
| 01:46 | <Philip`> | I found quite a few ways to get ~5% speedups in html5lib, but it's still orders of magnitude slower than the Java parser |
| 01:54 | <Hixie> | heycam: enough to make up a convincing example of <ruby> usage in japanese and/or chinese |
| 01:55 | <heycam> | the use of ruby isn't that common in chinese, a japanese example would be more convincing |
| 01:55 | <heycam> | and for that i suppose MikeSmith would be a good person to ask |
| 01:56 | <gsnedders> | jmb: Also, the proposed new Scottish law is readable |
| 02:02 | <gsnedders> | The amount of crime done by each age increases up to the age of 18 |
| 02:04 | <olliej> | Philip`: the best way to improve performance is to rewrite it in assembler. duh. :D |
| 02:04 | olliej | hides |
| 02:04 | <gsnedders> | :D |
| 02:05 | <gsnedders> | I'm still convinced of where to make a request under the freedom of information act |
| 02:07 | <Hixie> | heycam: i'd like both |
| 02:07 | <Hixie> | woo, MikeSmith is here now |
| 02:07 | <Hixie> | MikeSmith: would you be able to provide a more correct example of <ruby> than the one now in the spec? |
| 02:07 | <heycam> | Hixie, ok i'll try to come up with a chinese example |
| 02:07 | <Hixie> | MikeSmith: or more than one |
| 02:07 | <Hixie> | heycam: cool, that'd be really helpful |
| 02:07 | <Hixie> | heycam: if you can just lift one from some existing text that's even better, fwiw |
| 02:10 | <heycam> | oh i found an example of using bopomofo ruby in css3-ruby |
| 02:10 | <heycam> | http://www.w3.org/TR/css3-ruby/ figure 4.1.3 |
| 02:11 | <heycam> | that's nifty, i didn't know bopomofo goes vertically top-to-bottom besides each ltr chinese character |
| 02:12 | <Hixie> | i'm trying to introduce new examples rather than reuse the ones in those specs -- reusing the one that's in the html5 spec now is how i ended up needing a new example in the first place :-P |
| 02:14 | <heycam> | how big do you want the example? just a few characters? |
| 02:15 | <heycam> | and what do you want, a png/svg of some text using ruby? the html markup? |
| 02:16 | <MikeSmith> | Hixie, heycam: a simple example would be 漢字 |
| 02:16 | <MikeSmith> | かん じ |
| 02:16 | <jcranmer> | kanji :-) |
| 02:17 | <jcranmer> | that's really the kanji for kanji? |
| 02:17 | <heycam> | no, hanzi :) |
| 02:17 | <MikeSmith> | heh |
| 02:17 | <Hixie> | is that one base and one ruby? |
| 02:17 | <MikeSmith> | two base and two ruby |
| 02:17 | <Hixie> | awesome |
| 02:17 | <jcranmer> | I'm guessing the かん is ruby for the first, じ for the second |
| 02:17 | <MikeSmith> | yeah |
| 02:18 | <MikeSmith> | 字 is じ "ji" .. it means "character" |
| 02:19 | <jcranmer> | The only kanji I can recognize is nippon |
| 02:19 | <MikeSmith> | 文字列 also uses 字 |
| 02:19 | <MikeSmith> | 文字列 might be a good choice too |
| 02:19 | <heycam> | ㄏㄢˋ ㄗˋ |
| 02:19 | <MikeSmith> | it means "string" |
| 02:19 | <heycam> | i think is the bopmofo for han4zi4 |
| 02:19 | <heycam> | gotta check the second syllable tho |
| 02:20 | <MikeSmith> | string in the programming sense |
| 02:20 | <MikeSmith> | もじれつ |
| 02:20 | <MikeSmith> | mo-ji-retsu |
| 02:24 | <jcranmer> | drat, I seem to have lost japanese character support somewhere along the line |
| 02:24 | <Hixie> | MikeSmith: is the example at http://www.whatwg.org/specs/web-apps/current-work/#the-ruby-element right then? |
| 02:26 | <heycam> | ok i'm pretty sure "ㄗˋ" is right for the second character |
| 02:26 | <MikeSmith> | Hixie: yeah, it's right for Japanese |
| 02:27 | <heycam> | might be good to have an example using pinyin too |
| 02:30 | <heycam> | Hixie, http://paste.lisp.org/display/72812 |
| 02:30 | <Hixie> | heycam: sweet, thanks |
| 02:32 | <Hixie> | i wonder why Brian Smith said "Also note that Chinese uses Ruby markup differently than Japanese, so a Chinese-language example would be a good idea as well" |
| 02:32 | <Hixie> | http://www.w3.org/mid/003101c8bf64$f536e730$0202a8c0@T60 |
| 02:32 | <Hixie> | i'll have to ask him |
| 02:33 | <heycam> | and here's a pinyin one: http://paste.lisp.org/display/72814 |
| 02:33 | <heycam> | as would be used in mainland china |
| 02:33 | <heycam> | (the bopomofo would be for use in taiwan) |
| 02:34 | <heycam> | so with the pinyin one, the ruby would go on top of the characters (at least i've seen it like that in some children's books) |
| 02:34 | <heycam> | but bopomofo goes vertically on the right of each character |
| 02:34 | <heycam> | MikeSmith, in japanese you can write some string of ruby that corresponds to multiple kanji, can't you? |
| 02:35 | <heycam> | in chinese i don't think you would do that |
| 02:35 | <heycam> | Hixie, also note that in that pinyin example the 汉 is the simplified form of 漢, so it's really the same word |
| 02:37 | <MikeSmith> | heycam: yeah, that's one of the things about written Japanese that makes it challenging.. it's not a one-to-one relationship between kanji and reading.. most individual kanji have at least two readings, and some have more |
| 02:39 | <MikeSmith> | e.g., 山 is both さん and やま - "san" and "yama" |
| 02:39 | <heycam> | さん sounds like the chinese-derived reading |
| 02:40 | <MikeSmith> | yeah |
| 02:40 | <heycam> | but actually my question was more: for a given reading, would you map multiple kana to multiple kanji? |
| 02:40 | <heycam> | or can you always decompose that to multiple (or one) kana mapping to a single kanji? |
| 02:41 | <MikeSmith> | of the two base readings, one is a reading that orginally came from Chinese |
| 02:42 | <MikeSmith> | multiple kana to multiple kanji.. not sure what you mean |
| 02:42 | <MikeSmith> | you mean the same pair or string of kanji could have multiple readings? |
| 02:42 | <MikeSmith> | e.g., it would be ambiguous how to read them as a set? |
| 02:43 | <heycam> | i mean do you ever consider two kanji together, inseparably, and map those to a given string of kana? |
| 02:43 | <heycam> | or can you always split the kana up to the two kanji? |
| 02:44 | <MikeSmith> | you read the two kanji together, inseparably |
| 02:44 | <MikeSmith> | you can't know how to read them just by looking at them individually |
| 02:44 | <heycam> | ok |
| 02:44 | <MikeSmith> | or you can't determine how to read them by looking at them individually |
| 02:44 | <MikeSmith> | an example: |
| 02:44 | <heycam> | and then the ruby is centered over the middle of the two kanji |
| 02:45 | <MikeSmith> | hmm, not necessarily |
| 02:45 | <MikeSmith> | or not even usually |
| 02:45 | <heycam> | <ruby>kanji1kanji2<rt>blahblah</rt></ruby> ? |
| 02:46 | <MikeSmith> | no, they are still written over each kanji.. even if the reading depends on the combination |
| 02:46 | <heycam> | ah ok, understood |
| 02:46 | <MikeSmith> | 馬車道 is an example |
| 02:46 | <MikeSmith> | it's the name of a place in Yokohama |
| 02:46 | <MikeSmith> | and my 10-year-old daughter and I were going there a while back |
| 02:47 | <MikeSmith> | but neither of us know how to read that name |
| 02:48 | <MikeSmith> | although we knew all the characters |
| 02:48 | <heycam> | :) |
| 02:48 | <MikeSmith> | we said, うま くるま みち |
| 02:48 | <heycam> | horse chariot something? |
| 02:48 | <MikeSmith> | yeah, |
| 02:48 | <MikeSmith> | but the way we were reading it, we knew it was wrong |
| 02:48 | <heycam> | heh |
| 02:48 | <heycam> | could people understand what you meant if you said it? |
| 02:48 | <MikeSmith> | the real reading is ば しゃ みち |
| 02:49 | <MikeSmith> | heycam: yeah, they would |
| 02:49 | <MikeSmith> | though they would laugh |
| 02:49 | <MikeSmith> | they'd know what we were saying |
| 02:49 | <MikeSmith> | "basha" is the reading that comes from chinese |
| 02:50 | <MikeSmith> | "uma" and "kuruma" are the normal Japanese readings for those |
| 02:50 | <MikeSmith> | but "michi" is also the Japanese reading |
| 02:50 | <MikeSmith> | if that name were consistent in the way it used Chinese readings instead of Japanese ones, |
| 02:50 | <heycam> | so many options for the novice to choose! |
| 02:51 | <MikeSmith> | that place should be called "bashadou" |
| 02:51 | <MikeSmith> | heycam: yeah, it can make learning to read quite a challenge |
| 02:52 | <MikeSmith> | and complicated by the fact that it doesn't always seem consistent |
| 02:53 | <MikeSmith> | e.g., why "bashamichi" instead of "bashadou"? who knows... |
| 02:54 | <heycam> | sounds more idiomatic than english pronunciations |
| 02:54 | heycam | afk dry cleaning |
| 03:10 | <Hixie> | http://bookoutlines.pbwiki.com/Predictably-Irrational is interesting |
| 03:19 | <Hixie> | anyone know what font this is? http://www.whatwg.org/specs/web-apps/current-work/images/sample-ruby.png |
| 03:19 | <Hixie> | i'm having trouble reproducing the quality of that example |
| 03:23 | <olliej> | lucida grande? |
| 03:25 | <Hixie> | seems like it |
| 03:25 | <Hixie> | thanks |
| 03:26 | <Hixie> | next question |
| 03:26 | <Hixie> | why doesn't IE8 render Lucida Grande with antialiasing...? |
| 03:26 | Hixie | fakes it with safari |
| 03:27 | <Hixie> | MikeSmith: should a ruby text be centered above its base? or left-aligned? |
| 03:32 | <MikeSmith> | Hixie: for text that's written horizontally, it's centered |
| 03:33 | <Hixie> | k, thanks |
| 03:41 | <Hixie> | ok japanese ruby case done; now let's look at these chinese cases from heycam |
| 03:44 | <Hixie> | MikeSmith: btw is your example the kanji for hanzi? or did i get confused at some point with the conversation earlier. |
| 03:46 | <Hixie> | heycam: what are the magic words i should use to describe these two examples? |
| 03:48 | <MikeSmith> | Hixie: kanji |
| 03:49 | <Hixie> | the kanji for kanji? |
| 03:49 | <Hixie> | that is, the kanji for "kanji"? |
| 03:55 | <MikeSmith> | Hixie: yep |
| 03:55 | <MikeSmith> | the actual kanji for "kanji" |
| 03:55 | <Hixie> | cool |
| 03:56 | <Hixie> | and heycam's examples are the bopomofo and pinyin respectively for "hanzi"? |
| 03:56 | <Hixie> | (which happens to be the same word but in chinese?) |
| 03:56 | <MikeSmith> | Hixie: I think so, but I don't know.. I don't know Chinese at all |
| 03:58 | <Hixie> | i wonder what lang="" to use for his two examples |
| 03:59 | <Hixie> | MikeSmith: what is the relationship between all this and "furigana"? any idea? |
| 04:00 | <Hixie> | (or zhuyin?) |
| 04:00 | <Hixie> | ah, zhuyin = bopomofo |
| 04:01 | <heycam> | Hixie, correct |
| 04:02 | <heycam> | so that's the same word ("chinese character") in japanese (mikesmith's example), traditional chinese (the bopomofo example) and simplified chinese (the pinyin example) |
| 04:03 | <heycam> | and yes zhuyin is bopmofo |
| 04:03 | <heycam> | i'd use zh-TW for the first and zh-CN for the second |
| 04:04 | <Hixie> | for both the base and the ruby? |
| 04:04 | <MikeSmith> | Hixie: furigana is the normal/general term in Japanese for those annotations for readings, even outside of the online context |
| 04:04 | <MikeSmith> | or, as far as Japanese goes, ruby is a way to represent furigana |
| 04:05 | <Hixie> | ah ok so furigana=ruby, not one of the alphabets or languages used in the ruby |
| 04:05 | <heycam> | Hixie, yeah i'd use it for the whole thing |
| 04:06 | <Hixie> | cool, thanks |
| 04:07 | <Hixie> | i love hsivonen's validator |
| 04:07 | <Hixie> | it even checks language codes for correctness |
| 04:08 | <heycam> | you could even use zh-Hans-CN and zh-Hant-TW respectively, i suppose |
| 04:08 | <heycam> | dunno if there's an advantage in doing that though |
| 04:13 | <heycam> | Hixie, in the rendering of the pinyin example, you should put it over or under the characters, not next to like bopomofo |
| 04:13 | <heycam> | not sure which (above or below) is preferred, i've seen both |
| 04:14 | <Hixie> | i've done the pinyin already, take a look |
| 04:14 | <Hixie> | doing the middle one now |
| 04:14 | <Hixie> | (bopomofo) |
| 04:14 | <heycam> | ok cool, looks good |
| 04:14 | <Hixie> | two questions for bopomofo |
| 04:15 | <Hixie> | should the single bopomofo character be centered vertically next to the base glyph, or top-aligned? |
| 04:15 | <Hixie> | and how should I render the "..." bit? |
| 04:15 | <heycam> | vertically centered |
| 04:15 | <heycam> | dots just like in the pinyin example |
| 04:15 | <heycam> | but the tone mark needs to be placed specially |
| 04:16 | <heycam> | the ˋ should go just to the above and right of the previous bopomofo character (in this example, at least) |
| 04:16 | <heycam> | i found an image before showing this placement, hold on |
| 04:16 | <Hixie> | there's a ` ? :-) |
| 04:16 | <Hixie> | oh i see |
| 04:17 | <Hixie> | so there is |
| 04:17 | <heycam> | you can see it in http://en.wikipedia.org/wiki/Bopomofo#Writing |
| 04:18 | <heycam> | and in http://www.w3.org/TR/2001/WD-ruby-20010216/bopomofo.gif |
| 04:19 | <Hixie> | wouldn't the ` go to the above and right pretty much automatically by virtue of just being next to the character? |
| 04:19 | <Hixie> | for the dots, i mean, how are ellipsis done in vertical text? |
| 04:20 | <Hixie> | is there a vertical ellipsis character? |
| 04:20 | <heycam> | Hixie, if you place the ` in the same line, instead of breaking before it? |
| 04:20 | <heycam> | i suppose it would |
| 04:20 | <Hixie> | cool ok, that's what i did without thinking of it :-) |
| 04:20 | <heycam> | not sure if the spacing would look right, or what the actual requirements are |
| 04:22 | Hixie | sidesteps the ... issue by removing the dots in the example! |
| 04:23 | <Hixie> | ok, check it out now |
| 04:24 | <heycam> | if you include literal "." characters, then it would render like that |
| 04:24 | <heycam> | you'd need to include the actual other ellipsis character if you wanted it |
| 04:25 | <heycam> | oh you made the characters top-to-bottom |
| 04:26 | <Hixie> | is that wrong? |
| 04:26 | <heycam> | no, you can find vertical text in chinese |
| 04:26 | <heycam> | online it's more common to be ltr tho |
| 04:27 | <heycam> | the bopomofo looks good |
| 04:27 | <Hixie> | ok cool |
| 04:27 | <Hixie> | we'll try this and see who complains! :-D |
| 04:27 | <heycam> | heh ok |
| 04:28 | <heycam> | i sorta think the horizontal chinese characters looks cooler though, since then you have a combination of horizontal text with the vertical ruby |
| 04:28 | <heycam> | but it doesn't matter |
| 04:28 | <Hixie> | i had to fake all the typography using tables in safari |
| 04:28 | <heycam> | heh |
| 04:28 | <Hixie> | so we'll just go with this for now |
| 04:29 | heycam | lunch |
| 04:30 | <Hixie> | later |
| 04:30 | <Hixie> | heycam, MikeSmith: thanks for the help |
| 04:31 | <MikeSmith> | cheers |
| 05:57 | <heycam> | Hixie, should the "<pre class=zh-TW>" be "<pre lang=zh-TW>"? (and similarly for the zh-CN example) |
| 05:58 | <Hixie> | er yes |
| 09:55 | <annevk> | hmm, heycam, why is the callback definition so verbose? can't the interface and handleEvent method all be implicit? |
| 10:08 | Hixie | replies from a mail from 2005 |
| 10:08 | <Hixie> | man, i didn't realise i had any of those left |
| 10:09 | <Hixie> | (other than in the folders that i've intentionally not opened yet) |
| 10:16 | <olliej> | Hixie: slacker |
| 10:44 | <annevk> | heh, the Predictably-Irrational link is great |
| 10:46 | <zcorpan> | Hixie: s/next to/above/ in the pinyin ruby example |
| 10:47 | <Hixie> | thx |
| 10:54 | <olliej> | Hixie: have got your irritating test case to 10fps now :p |
| 10:55 | <Hixie> | :-) |
| 11:23 | <Hixie> | wow, we ain't popular on http://blogs.zdnet.com/open-source/?p=3234 |
| 11:34 | <Dashiva> | I like how that one guy puts CSS 3 as a success and CSS 2.1 as a failure |
| 11:37 | <jgraham> | I was more asonished that the original author had managed to get someone to publish a book they had written |
| 11:39 | <jgraham> | (the CSS3 guy has the cute belief that MS's motivation for Silverlight is mainly to create the best product it can technically) |
| 11:40 | <Philip`> | I'm just surprised by how much effort they went to to make blogs.zdnet.com's comment (uh, TalkBack) system as hard to read as possible by putting every comment on a separate page |
| 11:42 | <jgraham> | I wonder if I'm the only person who still associates the term "talkback" with that crash reporter tool that Netscape used to ship |
| 11:42 | <Philip`> | It wasn't just Netscape - Firefox used it pretty recently |
| 11:42 | <roc> | jgraham: certainly not! |
| 11:43 | <roc> | Yeah, Firefox 2 used it |
| 11:43 | <roc> | fortunately we got rid of it, with Google's help |
| 11:44 | <jgraham> | It is just amusing to have the commenting system named after something that recalls memories of browser crashes past... |
| 11:45 | <jgraham> | (I did also remember that Firefox used it but misguidedly thought that it wasn't too necessary to be accurate on IRC; obviously I should have known better) |
| 11:47 | <annevk> | Hixie, that post doesn't have a lot of research behind it... |
| 11:47 | <Philip`> | jgraham: (I only commented on the Firefox association because you made it sound like an ancient forgotten tool from the prehistory of the web when people used Netscape, whereas actually it was current only six months ago) |
| 11:47 | <Hixie> | the post itself is just wrong (we're actually bang on schedule according to what i predicted back in 2006) |
| 11:48 | <Hixie> | but the blog comments are quite negative to html5 as a whole |
| 11:48 | <Hixie> | i wonder what we should be doing about this |
| 11:48 | <Hixie> | if anything |
| 11:49 | <roc> | most of them just seem to want a clean-slate platform |
| 11:49 | <Philip`> | We should be amused by their misguided comments and assert our superiority here in IRC |
| 11:49 | <roc> | there's nothing really wrong with wanting a clean-slate platform, that's just not the Web |
| 11:50 | <Hixie> | Philip`: i was hoping for something that would work more effectively to convince them that html5 is the right thing to do :-) |
| 11:50 | <Hixie> | we probably need some sort of lightweight marketing to associate the name "html5" with something good |
| 11:50 | annevk | starts reading comments |
| 11:51 | <annevk> | we could call it BUBBLE programming, or some other weird term liking to Ajax, Comet, Clouds, etc. |
| 11:51 | <jgraham> | Are we suposed to backronym BUBBLE? |
| 11:52 | <annevk> | I'm not sure it matters, people will prolly refer to it as Bubble anyway :) |
| 11:53 | <jgraham> | Hixie: Only two or three of the commenters seemed to really dislike HTML5 |
| 11:53 | <annevk> | "The fail here lies with the browser developers who have misguidedly put their weight behind the red herring that is HTML5 instead of getting their **** together and implementing XHTML, XForms, and getting behind evolving the preexisting standards." |
| 11:54 | <olliej> | um |
| 11:54 | <olliej> | because xhtml is the future? |
| 11:54 | <Hixie> | but html5 _is_ evolving the preexisting standards... |
| 11:54 | <annevk> | "Why bother implementing something so completely worthless? The purpose of the HTML 5 working group isn't really to develop the next version of HTML, it's a case study on design by committee. Instead of Moore's law, why not talk about the HTML5 law: something about the inverse relation of the value of a product to the number of people who designed it." |
| 11:55 | <Hixie> | hey at least that guy isn't complaining about me being a dictator |
| 11:55 | <jgraham> | Hixie: And it is possible that paying too much attention to them is the fallacy whereby you mistake someone's percieved needs (what they say they want) for their actual needs |
| 11:55 | <roc> | the fact that these complaints are contradictory suggests things are somewhat on the right path |
| 11:55 | <Hixie> | heh |
| 11:56 | jgraham | wonders if there is a convenient name for asking what people want instead of observing their behaviour |
| 11:57 | <annevk> | then there is the longer comment from PB_z explaining a few things what's wrong with HTML5, but most of it seems not quite right |
| 11:57 | <webben> | jgraham: referendum? |
| 11:58 | <jgraham> | webben: heh |
| 11:58 | <Hixie> | annevk: seems pretty right to me (apart from the theory behind the motivations, though we might just be biased there) |
| 11:58 | <jgraham> | I guess "democracy" is close.. |
| 11:58 | <annevk> | and then there's sobri who thinks XHTML is substantially better than HTML... |
| 11:59 | <Dashiva> | It has the X, of course it is |
| 11:59 | <jgraham> | That's why thy called the TV show the X-Factor afterall |
| 12:00 | jgraham | wonders if that reference works anywhere outside GB |
| 12:01 | <annevk> | Hixie, well, 1) "Backward-looking rather than forward-looking." I don't believe that's true; we're doing both; 2) " It's one monolithic spec." I indeed disagree with the motivation here; 3) He forgets to mention that Microsoft wasn't doing anything at all at the time |
| 12:01 | <webben> | jgraham: I think XFactor has crossed borders. |
| 12:01 | <Philip`> | jgraham: By that reasoning, HTML5 should have been called Krypton-HTML |
| 12:01 | <webben> | jgraham: http://en.wikipedia.org/wiki/The_X_Factor |
| 12:02 | <annevk> | zcorpan, I'm right about spellchecking, correct? |
| 12:02 | annevk | should maybe have asked that just before posting... |
| 12:16 | <Hixie> | hmm |
| 12:16 | <Hixie> | tristate checkboxes |
| 12:16 | Hixie | ponders whether to support them and how |
| 12:17 | <Hixie> | btw i do think we should get more exposure in the media for html5, specifically the background reasoning, the timetable, and the features we've added |
| 12:17 | <annevk> | IE already has support for them, no? |
| 12:17 | <Hixie> | so that when we go to last call we have primed people to send comments |
| 12:17 | <Hixie> | IE has some weird support, yeah |
| 12:20 | <annevk> | you could write an article elaborating on those three things and then link it from the spec or some such |
| 12:22 | <Philip`> | "exposure in the media" probably involves things like press releases and emailing reporters, rather than just writing something and hoping people will find it |
| 12:22 | <annevk> | blogs would find it, at least |
| 12:23 | <Hixie> | we could also actually write an article for zdnet or some such |
| 12:23 | <Philip`> | Blogs that are outside of the existing HTML5 community? |
| 12:24 | <annevk> | our last news item was from 2006, nice |
| 12:24 | <annevk> | maybe http://www.whatwg.org/news/ should link to the blog |
| 12:25 | <annevk> | Philip`, I'm assuming there's some kind of ripple effect |
| 12:25 | <zcorpan> | annevk: well now we just have to implement it or we'll look silly for not being able to make up our minds :P |
| 12:25 | <annevk> | Hixie, that could work |
| 12:25 | <annevk> | zcorpan, great :) |
| 12:33 | <Hixie> | updated whatwg.org/news |
| 12:36 | <annevk> | all of W3C seems to be taking a break, most mailing lists I track have their last message posted Dec 19 or so |
| 12:49 | <Hixie> | .indeterminate in IE is a weird attribute |
| 12:49 | <Hixie> | i don't understand it |
| 12:51 | <annevk> | setting it to true puts the checkbox in the indeterminate state, quite simple, no? |
| 12:51 | <zcorpan> | only if the checkbox is checked iirc |
| 12:51 | annevk | just tested it, the checked and indeterminate are orthogonal, with indeterminate winning |
| 12:52 | <annevk> | states are orthogonal... |
| 12:52 | <zcorpan> | ok |
| 12:54 | <Hixie> | and what submits? |
| 12:57 | <annevk> | do you have some echo script somewhere? |
| 12:57 | annevk | is playing with the live dom viewer |
| 12:57 | <Philip`> | If you just submit to "" then it'll echo the query string |
| 12:58 | annevk | finds http://software.hixie.ch/utilities/cgi/test-tools/echo |
| 13:00 | <annevk> | Hixie, doens't affect submission |
| 13:00 | <annevk> | doesn't, even |
| 13:00 | <Hixie> | that's pretty useless then |
| 13:01 | <annevk> | I thought the main use case was some kind of UI used in tree views where submission is not important |
| 13:04 | <annevk> | Larry seems to be saying the same again... or I'm missing something |
| 13:05 | <annevk> | and not at all addressing points such as that the requirements for "web browsers" are not at all unique to them |
| 13:05 | <Hixie> | i considered just copy and pasting my last e-mail |
| 13:06 | <Hixie> | but that seemed rude |
| 13:06 | <Hixie> | oh he sent new mail |
| 13:17 | <annevk> | yeah, i.e. three times the same message without addressing points raised |
| 13:35 | <billyjackass> | we clearly need to use the word "interoperability" interoperably |
| 13:36 | <annevk> | all of the W3C uses it to mean what Larry describes as "uniform" |
| 13:36 | <karlcow> | do we have an interoperable definition for it |
| 13:36 | <karlcow> | one which would permit that used separatly in texts written by different people would have the same meaning |
| 13:37 | <annevk> | e.g. http://www.w3.org/QA/2008/05/open-standards-interoperability.html or http://www.w3.org/TR/CSS21/#crec |
| 13:37 | <annevk> | maybe we need triples! |
| 13:37 | <annevk> | larrry:interoperable and w3c:interopable |
| 13:37 | <Philip`> | Define "interoperability" so that it MUST be interpreted as whatever each reader wants it to mean, and then the word will interoperable because it's got a conformance requirement |
| 13:38 | <Philip`> | s//be/ |
| 13:38 | karlcow | remembers having written a mail about that in the past. |
| 13:39 | <annevk> | Philip`, that might be hard to test |
| 13:39 | <karlcow> | http://www.w3.org/QA/glossary it is not defined here |
| 13:39 | <billyjackass> | I didn't realize we were all of us consistently using the word completely wrong all this time. Clearly we need to do some mass re-education on ourselves |
| 13:39 | <annevk> | just disambiguate yourself by using a colon ;) |
| 13:39 | <billyjackass> | and probably we need some Maoist-style struggle sessions to drill the re-think into ourselves collectively |
| 13:41 | <Philip`> | annevk: Not at all - you just read the word, and if you think it means what you want it to mean then the test passes |
| 13:41 | <Hixie> | gsnedders: your complaint is ambiguous. Are you complaining about the attribute existing, or me not adding it to the spec? |
| 13:41 | <gsnedders> | Hixie: The attribute existing at all, and being supported at all |
| 13:42 | <Hixie> | gsnedders: the idea is to disable the feature when the input is known to be not spellcheckable, e.g. an address or some such |
| 13:42 | <Hixie> | or, say, a CAPTCHA |
| 13:42 | <Hixie> | (and the idea is to let the user override it to on anyway) |
| 13:42 | <karlcow> | ah no it was about conformance |
| 13:42 | gsnedders | expects it will probably be abused anyway |
| 13:43 | <gsnedders> | Hixie: Addresses are spell-checkable :P |
| 13:43 | <Hixie> | yours might be :-P |
| 13:43 | <karlcow> | but I came up with |
| 13:43 | Philip` | likes code editors that spell-check comments and strings |
| 13:43 | <karlcow> | >The main goal of conformance and normative requirements is to ensure a minimum of interoperability (aka two products are able to work in the same technological space even if developed totally independently.) |
| 13:43 | <Philip`> | (and ignore the rest of the code) |
| 13:43 | <gsnedders> | Hixie: Well, with a custom dictionary :P |
| 13:43 | <gsnedders> | (Heck, my surname isn't in most default dictionaries) |
| 13:44 | <Hixie> | the theory is that it's just a pain to see red underline every time one goes to google maps, i guess (or whatever site uses it) |
| 13:44 | gsnedders | would rather not |
| 13:44 | <gsnedders> | I can never spell foreign city names |
| 13:44 | <gsnedders> | Having spell checking is a net-gain, IMO |
| 13:44 | <annevk> | it can also be used to implement your own spell checking and turn built in off, if you have your own custom language |
| 13:44 | <karlcow> | 東京 |
| 13:45 | <gsnedders> | annevk: Why would you do that? |
| 13:45 | <gsnedders> | :P |
| 13:45 | <gsnedders> | Can't the spell checker just use whatever the language of the element is, and just have no spell-checking if the language is unknown |
| 13:45 | <gsnedders> | *? |
| 13:45 | <gsnedders> | *? |
| 13:45 | <annevk> | ask Tolkien |
| 13:46 | <Philip`> | I don't think he used a web browser much while developing languages |
| 13:49 | <annevk> | gsnedders, also, I could imagine a Web based spellchecker to be superior to the one offered by the browser because it can heuristic processing on some backend, sites might want to depend solely on such a service |
| 13:49 | <gsnedders> | annevk: I want my own damned spell-checker. If I want to use my spell-checker, let me. |
| 13:49 | <annevk> | as Hixie said, you can |
| 13:50 | <annevk> | that's part of the feature |
| 13:50 | <gsnedders> | But I have to manually re-enable it, no? |
| 13:50 | <gsnedders> | Which I think is stupid. |
| 13:50 | <annevk> | depends on your configuration |
| 13:51 | <jim_c> | Good morning all. For HTML5, is the correct content type(s) text/html, and application/xhtml+xml for the HTML and XHTML serializations, respectively? |
| 13:52 | <Philip`> | (Also, I don't think it's entirely realistic to use spell-checking on a language where the entire known lexicon is only a few thousand words) |
| 13:52 | <annevk> | yup |
| 13:52 | <gsnedders> | Philip`: why not? |
| 13:52 | <annevk> | Philip`, programming languages then? |
| 13:53 | <annevk> | (the yup was in reply to jim_c) |
| 13:53 | <jim_c> | Thanks. |
| 13:53 | <Philip`> | gsnedders: Because it means most words will be outside the known lexicon, and so spell-checkers will report far too many false positives |
| 13:54 | <gsnedders> | meh. |
| 13:54 | <annevk> | Philip`, if you implement your own spellchecker it might take care of that though or allow you to add new words easily |
| 13:56 | <annevk> | anyway, it seems useful for sites to be able to go beyond what the browser offers by default; after all, there's vastly more developers out there than those part of the browser teams |
| 13:56 | <Philip`> | annevk: Programming languages seem like basically the same case as e.g. blog comment forms where you can enter HTML markup, or forums where you use some custom markup language |
| 13:56 | <Philip`> | (since they're mixing some natural language and some special syntax) |
| 13:56 | <Philip`> | and in the latter cases you still want proper spell-checking |
| 13:57 | <Philip`> | so I guess browsers ought to somehow cope with checking text that's mixed with a markup/programming language |
| 13:57 | <gsnedders> | TextMate does, IIRC. |
| 14:21 | <annevk> | Hixie, you changed your mind about submission? |
| 14:22 | <Hixie> | i found that we already had interoperability |
| 14:22 | <Hixie> | lol, anne, you quoted text i wrote as a way to defend me |
| 14:22 | <Hixie> | :-P |
| 14:24 | <annevk> | heh |
| 14:25 | <Hixie> | right, bed time |
| 14:25 | <Hixie> | nn |
| 14:28 | <jgraham> | Philip`: Allowing for spellchecking that ignored markup was one of the usecases I had in mind for accept on textarea several years ago |
| 17:16 | <annevk> | guess I'm not going to reply to the "interopability is the wrong word" thread on www-tag (well, not more than I already did...) |
| 17:37 | <annevk> | section 5 is dull |
| 17:37 | <annevk> | try reading e.g. 5.1.5 out loud |
| 17:38 | <annevk> | and check if the person sitting next to you has a clue as to what is going on |
| 17:40 | <karlcow> | annevk: maybe it is just a difference in philosophy over developping technical specifications more than words |
| 17:40 | <gsnedders> | Can I please be less good at procrastinating? |
| 17:41 | <gsnedders> | annevk: I just burst out laughing trying to say browsing context an absurd number of times in succession |
| 17:41 | <annevk> | karlcow, I suppose, I wonder why you'd persue such a distraction then... |
| 17:42 | <annevk> | gsnedders, yeah, it's fun in a way :) |
| 17:42 | <gsnedders> | Now, I'm procrastinating again! |
| 17:42 | <gsnedders> | Damnit! |
| 17:42 | <annevk> | karlcow, sorry, meant "they would", not "you'd" |
| 17:43 | <gsnedders> | Oooo… reply from Cambridge should be sent out at the end of this week! |
| 17:46 | <karlcow> | annevk: like in any social structures, strong advocates of one camp tend to think that their system is right. In fact it is rarely the case. Nobody is really wrong or right. They just chose different ways. :) When I see such things, I try to not be fundamentalist on either sides. :) |
| 17:46 | Philip` | discovers the slightly-painful way that when EC2 has given your virtual machine a random initial login password, you ought to change it to something you know, otherwise you'll never be able to log in again |
| 17:59 | <karlcow> | gsnedders: more for your procrastination http://www.fbi.gov/page2/dec08/code_122908.html |
| 18:01 | <Philip`> | It's a rather pointlessly trivial code |
| 18:04 | <gsnedders> | On 43 Things, people who are doing "be joyful" are also doing "master CSS". |
| 18:04 | <gsnedders> | So, once you've mastered CSS, you are joyful? |
| 18:04 | <Philip`> | Yes, because then you can backups of all your DVDs |
| 18:04 | <Philip`> | s//make/ |
| 19:01 | <annevk> | karlcow, pacifist! :p |
| 19:03 | annevk | will comment anyway |
| 19:07 | <karlcow> | hehe |
| 20:34 | <doodlewarrior> | does anyone else think it's silly that boolean attributes don't accept true/false as values? |
| 20:34 | <Philip`> | Yes |
| 20:34 | <Philip`> | but it's too late to change it now |
| 20:35 | <doodlewarrior> | any idea why it's that way (or why it's too late to fix)? |
| 20:37 | <Philip`> | I would guess it's the way it currently is because HTML used to try to be SGML and tried to be concise, and SGML let you abbreviate attributes by skipping the attribute name, so <input disabled> would be equivalent to <input disabled=disabled> ... |
| 20:37 | <doodlewarrior> | fair enough |
| 20:38 | <Philip`> | (which worked because only the 'disabled' attribute could have the value 'disabled', and so the SGML parser could guess the attribute name given the value) |
| 20:38 | <doodlewarrior> | however required='true' should alias required='required' |
| 20:38 | <doodlewarrior> | whereas required='false' should be the same as omitting required altogether |
| 20:38 | <Philip`> | and so the boolean states were <input disabled=disabled> and <input>, and implementations just looked for the presence/absence of the attribute and ignored its value |
| 20:39 | <Philip`> | and so <input disabled=false> got treated like <input disabled=disabled> |
| 20:40 | <doodlewarrior> | do modern rendering engines actually treat disabled='false' as disabled='disabled' though? |
| 20:40 | <doodlewarrior> | or is this old-school 'we think we're SGML' stuff |
| 20:41 | <doodlewarrior> | -brb- |
| 20:41 | <Philip`> | They still do |
| 20:41 | <Philip`> | (They parse it into an attribute with value "false", but they process it as if it said "disabled") |
| 20:48 | <virtuelv> | implementors can't really change this either, given that the web has come to rely on such quirky behavior |
| 20:51 | <Philip`> | For the particular case of <input disabled="false">, I only find http://crafts.kaboose.com/holidays/chinese_new_year.html and that would actually work better if browsers did treat it as non-disabled |
| 20:51 | <Philip`> | but I would expect people do all sorts of crazy things with other attributes and would become very unhappy if the behaviour suddenly changed |
| 21:08 | <doodlewarrior> | i can't imagine there is any case where someone has disabled='false' when they mean disabled='disabled' |
| 21:09 | <doodlewarrior> | implementors could check if the value == false before they interpret it as true |
| 21:09 | <doodlewarrior> | it doesnt make any sense to have a boolean that doesn't accept true/false as attributes |
| 21:09 | <doodlewarrior> | and considering most pages are built with dynamic languages where booleans are typed to true/false, it makes way more sense |
| 21:10 | <gsnedders> | But we're stuck and constrained by legacy content |
| 21:10 | <doodlewarrior> | how is that constraining? |
| 21:11 | <Philip`> | doodlewarrior: Looking at real web pages reveals a lot of cases that one could never have imagined :-) |
| 21:11 | <doodlewarrior> | Philip`: have you ever seen someone assign a value of false when they meant to have a value of true? |
| 21:11 | <doodlewarrior> | the whole point of standards is to push the web forward |
| 21:12 | gsnedders | would be surprised if there wasn't a website that relied on it |
| 21:12 | <gsnedders> | Sure, but we can't break existing websites. |
| 21:12 | <doodlewarrior> | we shouldnt be shackled by previous nonsense |
| 21:12 | <doodlewarrior> | i dont imagine that would break many sites |
| 21:12 | <gsnedders> | Should we break 99% of the web because it is non-conforming? |
| 21:12 | <doodlewarrior> | youre not breaking 99% of the web |
| 21:13 | <gsnedders> | Why should we risk breaking anything making this change, though? |
| 21:13 | <gsnedders> | What do we gain by making the change? |
| 21:14 | <doodlewarrior> | logical coherence for one |
| 21:14 | <gsnedders> | There are far bigger things that aren't logical in HTML :) |
| 21:15 | <doodlewarrior> | to call something a boolean and have it accept false as a value for true is asking for trouble |
| 21:15 | <doodlewarrior> | (i only stumbled upon this because my HTML5 site stopped validating when i added attributes with true/false as values) |
| 21:15 | <gsnedders> | To call something "SGML-like" and not say what it is is asking for trouble :) |
| 21:15 | <Philip`> | doodlewarrior: What about accepting "FALSE", or "0", or "no", or ""? |
| 21:16 | <doodlewarrior> | Philip`: although i can see the case being made for those, the simple values of lowercase false and true are sufficient |
| 21:17 | <doodlewarrior> | theres a difference between true and truthy |
| 21:17 | <doodlewarrior> | true/false should definitely be supported |
| 21:17 | <doodlewarrior> | truthy/falsy much less critical |
| 21:18 | <takkaria> | the rule at the moment is: if you have an attribute present, it is present, otherwise, it is not |
| 21:18 | <takkaria> | o |
| 21:18 | <takkaria> | your rule is just a different one that makes just as much sense |
| 21:19 | <doodlewarrior> | im a usability guy, not an engineer |
| 21:19 | <doodlewarrior> | so not-breaking the user's expectations is what i think about all day :) |
| 21:20 | <doodlewarrior> | to someone who's designed in another major language, the a Boolean has a meaning |
| 21:20 | <doodlewarrior> | which includes true/falseness |
| 21:20 | <doodlewarrior> | to call something a boolean and have it behave differently is a major breakage of the designer's expectations |
| 21:20 | <Philip`> | doodlewarrior: Even if pages don't currently rely on disabled=false meaning disabled=disabled, there are scripts like http://code.google.com/p/qwin/source/browse/trunk/content/preferences.js?r=10#94 that rely on hasAttribute('disabled') telling you whether the element is disabled or not |
| 21:21 | <takkaria> | well, the user's expectation is set if they read that "if an attribute is present, then it is present, and if it is not, it is not, regardless of its value" |
| 21:21 | <takkaria> | it's just as boolean as what you propose, just in a different way |
| 21:22 | <doodlewarrior> | Philip`: nice counterexample |
| 21:22 | <doodlewarrior> | i was thinking more about user agents than javascript |
| 21:22 | Philip` | failed to find examples of reliance on disabled=false in actual markup :-( |
| 21:23 | <doodlewarrior> | i figured standards-mode/html5 UAs could check for false before interpretting as true |
| 21:23 | <doodlewarrior> | javascripts like you mention do raise a breakage-risk |
| 21:24 | <Philip`> | If we have to break people's expectations, it's best to do it immediately when they're writing their content (and the validator complains, or form controls are disabled when they shouldn't be, etc), rather than breaking the then-correct expectations that people had when writing scripts years ago |
| 21:25 | <doodlewarrior> | im not sure thats correct, but i see your point |
| 21:26 | <doodlewarrior> | there are plenty of designers who never bother validating |
| 21:26 | <Philip`> | Also there's the usual problem that if we defined new processing for disabled="false", users of old (i.e. current) web browsers will get exactly the opposite effect than the author intended, and the author may have only tested in futuristic browsers and would never be aware of the problem |
| 21:29 | <doodlewarrior> | old browsers break a lot of things. really, this is not a change that would probably happen for another generation for exactly the reason you mention |
| 21:29 | <doodlewarrior> | but if bad behavior is never made obsolete, it will never go away |
| 21:29 | <doodlewarrior> | hence why i brought it up |
| 21:29 | <doodlewarrior> | at least now i know why it's there |
| 21:34 | <gsnedders> | doodlewarrior: What bad behaviour that has been made obsolete in HTML has ever gone away? |
| 21:35 | <Philip`> | Hopefully it will go away when someone designs a new web language and learns from all the mistakes of HTML, but for now we're kind of stuck with what we've got |
| 21:35 | <gsnedders> | doodlewarrior: the plaintext element was marked as obsolete from the first draft HTML standard, and still has to have special parsing rules in HTML 5 far-too-many-years later |
| 21:38 | <annevk> | Philip`, just replacing HTML is not quite going to cut it, I think |
| 21:39 | <annevk> | yay, RB returns |
| 21:50 | <Philip`> | I might suggest an interjection other than "yay" |
| 21:50 | <takkaria> | more allegations of the parsing algorithm being severely broken :) |
| 21:54 | <Dashiva> | Hixie surely is the master of unfinished sen |
| 23:32 | gsnedders | blames jgraham for him falling in love |
| 23:32 | <gsnedders> | :P |
| 23:33 | <gsnedders> | (It's fun giving not quite enough information for things to quite make sense) |
| 23:46 | <jgraham> | gsnedders: I can't help it I just have that efect on people :p |
| 23:46 | <gsnedders> | jgraham: Not with you, you silly! :P |
| 23:47 | <jgraham> | Ah yes, I was confused between love and repulsion. Sorry the concepts are pretty similar |
| 23:47 | gsnedders | is listening to Starálfur by Sigur Rós from Ágætis byrjun |
| 23:47 | <gsnedders> | W |
| 23:47 | <gsnedders> | *ith that song |
| 23:50 | <jgraham> | gsnedders: If you have only fallen in love with one song you really need to listen to more Sigur Ros :) |
| 23:50 | gsnedders | has been listening to most of their free stuff |
| 23:50 | jgraham | would get the accents right but for some reason terminal.app isn't playing ball |
| 23:50 | <gsnedders> | That's the only one I've totally fallen for :P |
| 23:52 | <jgraham> | I recommend the DVD "Heima" which collects recordings they made when doing a series of free, unannounced(?) gigs round iceland. It really evokes a sense that the music is the product of the landscape in whih they grew up |
