| 00:00 | <othermaciej> | probably worth pointing out |
| 00:23 | Philip` | wonders what the proportion of facts vs namespace religion will be |
| 00:25 | Philip` | won't point out obvious mistakes like the failure to use a URI as the namespace attribute value |
| 00:33 | Philip` | likes how he finally has enough RAM to run three web browsers at once, plus a virtual machine with Windows and another five web browsers, without anything slowing down noticeably |
| 00:40 | tantek | thinks the proposal is one of the more hilarious things he's seen on the public-html list in quite some time. |
| 00:40 | <tantek> | s/distributed extensibility/proprietary extensibility/ |
| 00:40 | <tantek> | for more humor value (and perhaps even "truer" statements) |
| 00:41 | tantek | is resisting the temptation to write a massive "bitch-slap" (technical term) email to public-html in response to the proposal. (Microsoft people ought to know better) |
| 00:46 | <Philip`> | tantek: Please do resist, it doesn't sound like it would contribute to productive discussion :-p |
| 00:53 | Philip` | reads a bit further |
| 00:54 | <Philip`> | The scripting and CSS examples seem crazy - why would you want to process/style elements based solely on their namespace? |
| 00:54 | tantek | resists by attempting to focus on higher priority things. |
| 00:55 | <tantek> | Philip` - in this case another way of saying 'crazy' is 'completely made up and unrelated to any real world practice', what we in #microformats call "purely theoretical" and handily dismiss as not worth discussing. |
| 00:56 | <Philip`> | "The proposal is made in a modular way so that any components causing unacceptable compatibility concerns could be abandoned." - indeed, we can just abandon the entire thing and tell people never to touch elements or attributes containing the character ':' because it's a compatibility nightmare |
| 01:06 | <othermaciej> | Philip`: it would be useful to do have test data on attributes too |
| 01:10 | <othermaciej> | how do you get at HTTP headers in PHP? |
| 01:10 | <othermaciej> | (sorry for asking all my dumb PHP questions here) |
| 01:17 | <Philip`> | othermaciej: Like that? [on list] |
| 01:18 | <othermaciej> | Philip`: namespaces don't have that tagUrn property in IE? |
| 01:18 | <othermaciej> | er, attributes rather |
| 01:18 | <othermaciej> | Philip`: does the Microsoft proposal specify how xmlns:foo attributes should look in the DOM? |
| 01:19 | <othermaciej> | Philip`: that might be another good thing to test |
| 01:19 | <Philip`> | othermaciej: $_SERVER['HTTP_UPPERCASED_HEADER_NAME'] perhaps? |
| 01:19 | <othermaciej> | Philip`: thank you sir |
| 01:19 | <Philip`> | othermaciej: IE doesn't do any namespace processing of attributes at all, so tagUrn is always just undefined |
| 01:19 | <othermaciej> | I neglected to uppercase |
| 01:20 | <Philip`> | (so I excluded it from the email) |
| 01:20 | <othermaciej> | Philip`: might be good to state that, since it's sort of the substitute namespaceURI for elements |
| 01:20 | <Philip`> | I don't think they say anything about xmlns:foo, but all text/html treat it identically to any other attribute with or without a colon |
| 01:20 | <Philip`> | s/treat/browsers treat/ |
| 01:22 | Philip` | won't bother sending another email to state anything now, since it seems better to go to bed |
| 01:23 | <Philip`> | (Also, most of this stuff has been discussed years ago anyway) |
| 01:23 | <Philip`> | (xmlns-dom.html was 1.5 years old) |
| 01:25 | <Philip`> | (Also the details don't seem too important when their claim of IE compatibility appears to be fundamentally bogus) |
| 01:25 | Philip` | goes to bed, and will try to predict how many "+1" responses the proposal gets by tomorrow |
| 07:20 | <hsivonen> | tantek: please do write your concerns about what is wrong with the "decentralized extensibility" proposal to the list |
| 09:04 | <tantek> | hsivonen - in short, the justifications provided for "decentralized extensibility" in Microsoft's proposal are either theoretical, solved by other mechanisms in HTML5, or are a poor rationalization for what amounts to proprietary extensions, which are known to be bad, and thus should be discouraged as far as standard formats are concerned. |
| 09:04 | <Hixie> | i encourage you to post that to the list, i think most people on the list are desensitised to me and henri saying this by now |
| 09:05 | <Philip`> | Hixie: You need a sock puppet |
| 09:06 | <tantek> | perhaps in the (PDT) morning, time for me to sleep on it. |
| 09:43 | <hsivonen> | are iframes supposed to delay the onload event of the parent? |
| 09:43 | <zcorpan_> | yes |
| 09:43 | <hsivonen> | ok |
| 09:44 | <zcorpan_> | "When there is an active parser in the iframe, and when anything in the iframe is delaying the load event of the iframe's browsing context's active document, the iframe must delay the load event of its document." |
| 09:44 | <hsivonen> | so, on on both www.aol.com and www.cnn.com I see more things trying to unblock onload than blocking onload |
| 09:44 | <hsivonen> | and on www.aol.com I see attempts to unblock onload periodically |
| 09:45 | <zcorpan_> | how do you unblock onload? |
| 09:45 | <hsivonen> | zcorpan_: I mean internal onload unblocking code |
| 09:45 | <zcorpan_> | ah |
| 09:46 | hsivonen | goes set a breakpoint to catch the periodic thing |
| 09:47 | <zcorpan_> | hsivonen: i did some research on <script><!-- yesterday |
| 09:47 | <hsivonen> | ha! I have innerHTML setter accidentally unblocking onload |
| 09:47 | hsivonen | wonders how innerHTML itself and images inserted by innerHTML should interact with onload |
| 09:48 | <hsivonen> | zcorpan_: did you manage to disqualify any proposals? |
| 09:48 | <zcorpan_> | hsivonen: yes |
| 09:48 | <hsivonen> | cool |
| 09:48 | <Philip`> | All of them? |
| 09:49 | <hsivonen> | so proposal 3 is the only one not marked as FAIL |
| 09:49 | <zcorpan_> | Philip`: i think proposal #3 breaks the least out of the lot |
| 09:49 | <zcorpan_> | but it still breaks some pages |
| 09:50 | <hsivonen> | intuitively, proposal 3 scares me more than what's specced now |
| 09:50 | <jgraham> | hsivonen: Why? |
| 09:50 | jgraham | doesn't really like any of the proposals |
| 09:50 | <hsivonen> | jgraham: proposal 3 breaks random sites out there |
| 09:50 | <zcorpan_> | what's specced now breaks more sites out there :) |
| 09:51 | <hsivonen> | jgraham: what's currently been reported about the spec is that it breaks one developer-oriented tool and one attempt of writing test cases |
| 09:51 | <hsivonen> | zcorpan_: ok. no one has bothered to note the other sites broken by the spec on bugzilla.mozilla.org |
| 09:51 | <zcorpan_> | hsivonen: what's specced now also breaks all of these http://philip.html5.org/data/script-close-in-escape-without-script-open-2.txt |
| 09:52 | <hsivonen> | ooh. scary. |
| 09:52 | <Philip`> | I propose getting rid of <script> |
| 09:52 | <zcorpan_> | proposal #3 only breaks 3 sites in there or so |
| 09:52 | <hsivonen> | //--><!--</script> |
| 09:52 | <hsivonen> | that's pattern is so, so, sad |
| 09:53 | Philip` | wonders if reparsing is really so bad, given that everybody does it |
| 09:54 | <hsivonen> | what are people thinking when they finish their script not with //--> but with //--><!-- |
| 09:54 | jgraham | wonders the same as Philip` |
| 09:54 | <zcorpan_> | they are clearly confused by all the comment tags "//", "<!--" and "-->" so don't know when to use which |
| 09:55 | <zcorpan_> | and it doesn't matter because it all works in browsers today |
| 09:56 | <hsivonen> | zcorpan_, Philip`: thanks for analyzing the proposals |
| 09:57 | <zcorpan_> | maybe we also need to look at pages that do have <script inside escaped text span, too |
| 09:58 | <zcorpan_> | to find pages with the second kind of breakage of #3 |
| 09:58 | <Philip`> | hsivonen: I blame Hixie, for writing a document that suggests <script><!--//--><![CDATA[//><!-- ... |
| 09:59 | <Philip`> | thus causing people to completely give up all hope of understanding commenting, and to just randomly insert comment characters until it works |
| 09:59 | <zcorpan_> | oh right, i forgot the other two comment tags "<![CDATA[" and "]]>" |
| 10:00 | gsnedders | calls theoretical on the whole security hole Hixie worries about here |
| 10:00 | <hsivonen> | Philip`: we have a problem with debating complexity on the Web |
| 10:01 | <hsivonen> | Philip`: if you say that something is impossible, you are wrong, but if you show that the possibility is ridiculously complex, there's always someone who copies and pastes it |
| 10:02 | <hsivonen> | it's like every complex piece of badness is an attractive nuisance |
| 10:02 | jgraham | suggests that the whoile situation would have been much simpler if <!-- --> just indicated a comment in an inline script block from the start |
| 10:03 | <Philip`> | hsivonen: Maybe we should convert any demonstrations of ridiculous complexity into images rather than text, so people can't copy-and-paste and fewer will manually type it all out |
| 10:04 | <hsivonen> | Philip`: that would be inaccessible |
| 10:04 | <zcorpan_> | Philip`: just do ondragstart="return false" |
| 10:04 | <jgraham> | (since HTML-aware text editors tend to be insuffciently HTML-aware to make their comment function notice that it is working in a script block rather than in some random HTML |
| 10:04 | <jgraham> | hsivonen: Not with alt text |
| 10:04 | <hsivonen> | Philip`: also, consider all the typos people would make when typing it |
| 10:05 | hsivonen | assumed alt text got copied |
| 10:05 | <jgraham> | But people won't expect it to work when they can't select it |
| 10:06 | <jgraham> | (well I know you can select images but people generally wouldn't select an image of text expecting the alt text to be the text) |
| 10:08 | <Philip`> | hsivonen: The inaccessibility has to be balanced against the humanity of not subjecting screen reader users to a spoken version of <script><!--//--><![CDATA[//><!-- |
| 10:12 | <hsivonen> | now that I've seen what kind of browser bugs are possible with innerHTML during load, innerHTML during load scares me a bit even for cases that aren't browser bugs |
| 10:17 | Philip` | wonders why Laurens Holst picked, of all the legitimate examples of insane complexities in HTML5, one that is actually very simple in HTML and is only confusing because of promotion of XHTML-as-text/html |
| 10:21 | hsivonen | wonders if Julian programs with namespaces, writes namespace-wise correct code and likes the experience |
| 10:25 | <hsivonen> | Yay. I fixed the innerHTML setting during load issue |
| 10:25 | <hsivonen> | It worries me that even the fix was non-obvious |
| 10:29 | <Philip`> | hsivonen: You don't necessarily have to like the experience of working with namespaces, you just have to not dislike it enough to counteract the benefits |
| 10:29 | <hsivonen> | Philip`: good point. |
| 10:29 | <hsivonen> | I rank the benefits the be negative. |
| 10:30 | <Philip`> | That sounds like a type error |
| 10:30 | <Philip`> | Benefits are by definition positive :-p |
| 10:30 | <Philip`> | s/positive/non-negative/ |
| 10:52 | <mookid> | anyone got any recommendations on web based open source scrum tools (preferably LAMP stack)? |
| 10:53 | Philip` | doesn't know many open source rugby simulators |
| 11:29 | hsivonen | wonders why texts that discuss multithreading don't explain how the multiple runtime stacks are typically laid out in the address space of the process |
| 11:30 | <hsivonen> | while texts that discuss processes commonly mention the address layout of the stack and the heap |
| 11:32 | <Philip`> | Is there much to discuss? |
| 11:32 | <hsivonen> | Philip`: saying how they are laid out |
| 11:32 | <Philip`> | I thought it was just like thread n has a stack at base-n*max_stack_size and that's about it and otherwise it's the same as a single-threaded process |
| 11:33 | <hsivonen> | Philip`: well, that's the bit that usually goes unmentioned |
| 11:33 | <hsivonen> | btw, what mechanism is there to prevent a stack longer than max_stack_size overwriting the next stack? |
| 11:34 | <Philip`> | You could put a non-writable page between stacks |
| 11:34 | <hsivonen> | OK |
| 11:34 | <Philip`> | Don't know whether that happens in practice, though |
| 11:35 | <hsivonen> | well, that's the thing. it's easy to solve these problems in theory, but no one ever says how they are solved in real systems |
| 11:35 | <Philip`> | http://gcc.gnu.org/onlinedocs/gcc-4.1.2/gnat_ugn_unw/Stack-Overflow-Checking.html - "For most operating systems, gcc does not perform stack overflow checking by default." |
| 11:37 | <hsivonen> | generating additional instructions seems worse than wasting a page of virtual address space |
| 11:37 | <Philip`> | (I expect you'd get a segmentation fault though, since the stack would extend into unallocated address space) |
| 11:37 | <Philip`> | (in the single-threaded case) |
| 11:43 | <Philip`> | Anyway, just don't overflow your stack and it'll all be fine :-) |
| 11:44 | <Philip`> | (and certainly don't use recursive functions) |
| 13:34 | <annevk2> | if anyone knows a more correct Unicode symbol for average than x̅ please let know |
| 13:34 | <annevk2> | x̅ doesn't look quite right and is somewhat of a hack I suppose |
| 13:36 | <jgraham> | Is that supposed to be x with a line on top? |
| 13:36 | <jgraham> | If so that is one of two common ways I can think of of prepresenting average |
| 13:36 | <jgraham> | the other being <x> |
| 13:36 | <Philip`> | I suggest using more than one symbol, such as "mean(x)" |
| 13:37 | <Philip`> | which has the bonus that it doesn't like identical to "x" in my IRC client |
| 13:37 | <jgraham> | Philip`: It depends what you are trying to do |
| 13:37 | <jgraham> | If you are writing an equation <x> or so is nicer |
| 13:38 | <Philip`> | If I was writing an equation I'd write something like \xbar |
| 13:39 | <Philip`> | (rather than trying to do it with Unicode) |
| 13:40 | <annevk2> | see http://annevankesteren.nl/2009/10/tirsdagspils for the use case |
| 13:40 | <annevk2> | jgraham, i.e. <x> in HTML? |
| 13:41 | jgraham | assumes that it is not possible to do something like \overline{x^2} using pure unicode |
| 13:41 | <Philip`> | annevk2: Just write "average" or "avg" or "mean" |
| 13:41 | <jgraham> | annevk2: Yeah. There are maybe "more proper" symbols to use than < and > |
| 13:42 | <Philip`> | The mathematical symbols don't seem to make sense in this context |
| 13:42 | <Philip`> | e.g. because you're never defining what 'x' is |
| 13:42 | jgraham | wonders what the context is |
| 13:42 | <Philip`> | jgraham: See annevk2's link |
| 13:43 | <jgraham> | Oh, I missed that |
| 13:43 | <jgraham> | Obviously |
| 13:44 | <jgraham> | Yeah, what Philip` said |
| 13:44 | <jgraham> | You should say "mean" or something if you want to be precise |
| 13:45 | <Philip`> | You could show the mean, median and standard deviation, to make sure people have enough information to make a good decision |
| 13:45 | <annevk2> | lame |
| 13:46 | <annevk2> | but thanks |
| 14:09 | annevk2 | files more bugs |
| 14:10 | <zcorpan_> | annevk2: xml:lang? |
| 14:11 | <annevk2> | zcorpan_, guess my post-processor is not doing its job everywhere |
| 14:11 | <annevk2> | I should fix some of these things I suppose |
| 14:35 | <annevk2> | does anyone recall whether Hixie wanted bugs on normal text referencing impl-protected text? |
| 14:36 | <annevk2> | e.g. the definitions of the constants for the appcache API do that a lot |
| 14:37 | <zcorpan_> | annevk2: i think he doesn't want bugs on that |
| 14:40 | <annevk2> | k |
| 14:41 | annevk2 | feels a little bad for filing all these bugs now |
| 15:06 | <annevk2> | http://www.w3.org/2001/tag/2009/09/HTMLIssuesRevised.pdf |
| 15:06 | <annevk2> | (from an Excel file?!) |
| 15:08 | <Philip`> | It's really terrible that people might use a spreadsheet program to write a spreadsheet |
| 15:08 | <annevk2> | agreed |
| 15:09 | <hsivonen> | annevk2: what's "?!" about Excel? |
| 15:10 | <annevk2> | surprise |
| 15:11 | <hsivonen> | so each TAG member assigned number of beans per issue |
| 15:11 | <hsivonen> | and the number can be > 1 |
| 15:11 | <hsivonen> | and there's no clear bean maximum |
| 15:11 | <annevk2> | I wasn't quite sure how the number game worked |
| 15:12 | <Philip`> | Maybe the implicit maximum is 10, but if you want you can turn it up to 11 |
| 15:13 | <hsivonen> | hmm. TimBL gave 4 beans to RDF triples but otherwise only one other TAG member gave a bean |
| 15:14 | <hsivonen> | at least the XPath thing got 0 beans |
| 15:14 | <Philip`> | Are the ones in purple magic beans? |
| 15:15 | <hsivonen> | hmm. lack of doc.write in XML got a bean |
| 15:17 | <hsivonen> | what does issue 7 (relationship of DOM & serialization) entail? |
| 15:18 | <Philip`> | That document indicates no TAG member wants to look for positive feedback |
| 15:18 | <Philip`> | hsivonen: Maybe that was the disabled=banana thing? |
| 15:19 | <hsivonen> | Philip`: disabled=banana is issue 13 |
| 15:20 | <Philip`> | hsivonen: That's what the layout indicates, but Henry's question seemed to be about issue 7 |
| 15:21 | <Philip`> | (particularly about what DOM you get from the serialisation disabled="banana") |
| 15:24 | <Philip`> | The interaction of prefix binding and table foster-parenting seems like a really boring issue |
| 15:24 | <Philip`> | since there's only one way it can possibly work that is not totally insane |
| 15:25 | <hsivonen> | Philip`: I hope we have the same idea of what isn't totally insane |
| 15:26 | <Philip`> | Anything that depends on the source characters, not on the parsed DOM, is totally insane |
| 15:26 | <hsivonen> | good |
| 15:57 | Philip` | wonders if there's a good JS editor plugin for Eclipse |
| 16:31 | <Dashiva> | Great game: http://www.stfj.net/art/2009/loselose/ |
| 16:34 | <TabAtkins> | Heh, interesting. |
| 16:35 | <Philip`> | Sadly its deep meaning is rendered futile by virtual machine snapshots that remove all possible consequences |
| 16:35 | <TabAtkins> | Actually, that just brings up further existential questions revolving around simulation and the consequences of Singularity. |
| 16:36 | <TabAtkins> | Is it truly killing if there is an exact copy that can be restored? How can we modify our ethics to handle this situation consistently? |
| 16:38 | <Philip`> | http://home.wildit.net.au/hellohelloben/mystery.html is a game with far more morally serious and unavoidable life-and-death consequences |
| 16:39 | <Dashiva> | The opening screen doesn't make me want to press start |
| 16:39 | <TabAtkins> | Are we doing something unethical when we create simulated life while talking with someone (modeling them mentally to predict reactions)? What about when a hyperintelligent AI is able to simulate us to arbitrary precision while talking to us, then deleting the simulations when done? |
| 16:40 | <Philip`> | Dashiva: It's one of the games of decade, according to http://www.eurogamer.net/articles/eurogamer-editors-games-of-the-decade-article?page=3 |
| 16:41 | <TabAtkins> | Oh jeezus haploid christ. |
| 16:42 | <Philip`> | TabAtkins: If I delete all the files on your hard disk and you have a backup, does that mean I didn't really delete all the files on your hard disk? |
| 16:42 | <TabAtkins> | Philip`, ethically, did you? What if those files are sapient entities? Did you commit murder? |
| 16:43 | Philip` | is reminded of Glasshouse, where everyone has backups of themselves, and murder is common but considered a far less serious crime than identity theft |
| 16:43 | TabAtkins | liked that book. |
| 16:44 | <TabAtkins> | Though that only barely touches on the issues of identity and existence in the face of simulation. |
| 16:45 | <Philip`> | The existence of backups doesn't change the definition of "murder", it just changes how much we care about it and how much effort we put into trying to stop it |
| 16:46 | <Philip`> | which are much more concrete ideas than the concept of 'ethics' |
| 16:46 | <Dashiva> | But you lose everything since the latest backup |
| 16:46 | <TabAtkins> | Hmm? Ethics are vitally important when you're talking about the possibility of an action being duplicated across trillions or more conscious entities in a moment. |
| 16:47 | <TabAtkins> | Such as in the simulation spaces of a hyperintelligent AI. |
| 16:47 | <Philip`> | Dashiva: Well, you need to work out a decent backup schedule |
| 16:47 | <TabAtkins> | Is it ethical to torture someone to death if you can restore them back to before the torture started with no memories of it? |
| 16:47 | <Philip`> | Knowing most people, they wouldn't actually back themselves up until they'd already been killed, and then it'd be too late |
| 16:48 | <Dashiva> | The real question is how they do the reanimation |
| 16:48 | <TabAtkins> | (I'd like to say no, but I'm honestly not sure if that's correct, given that it's only an issue of computation resources that distinguishes the poor copies of others I create in my mind and copies that are functionally indistinguishable from real people.) |
| 16:49 | <TabAtkins> | (And I don't think it's unethical to torture a mental image. Probably.) |
| 16:49 | <Dashiva> | In associated news: Squids are one of the four fundamental forces of the universe, alongside the Weak Force, the Strong Force and Gravity. |
| 16:49 | <Philip`> | Dungeon Keeper had a torture chamber |
| 16:49 | <Philip`> | That was a fun game, so I'll decide it's not unethical |
| 16:49 | <TabAtkins> | This is your fault, Dashiva. |
| 16:50 | <TabAtkins> | Philip`, I'll see if you change your mind when the AI who is executing this universe in its mind decides to torture everyone to death. |
| 16:51 | <Dashiva> | Actually, lose/lose reminds me of Only You Can Save Mankind |
| 16:51 | <Philip`> | Also your cursor was a Hand of Evil, and you could slap all your imps |
| 16:52 | TabAtkins | has fond memories of being an evil god in Black and White. |
| 16:52 | <TabAtkins> | My people led a life of unrestricted hedonism, as long as they were willing to give me about half their babies. |
| 16:53 | <Philip`> | Dashiva: I always wanted a copy of the real-time journey-to-alpha-centauri game from that |
| 16:56 | <TabAtkins> | Philip`, go get one. It's just a few jumps from the Wikipedia article. |
| 16:57 | <Philip`> | TabAtkins: I stopped wanting it about ten years ago :-p |
| 16:57 | <TabAtkins> | Hehe. |
| 18:26 | <Fernandos> | hi |
| 18:26 | <TabAtkins> | yo |
| 18:27 | <Fernandos> | can someone help me debugging a site made with basic html5 |
| 18:27 | <TabAtkins> | Possibly. Shoot us the link. |
| 18:27 | <Fernandos> | http://fif-moebel.de |
| 18:28 | <Fernandos> | I get 3 errors which make no sense to me. |
| 18:28 | <Fernandos> | I doubt the validator is wrong http://html5.validator.nu/?doc=http%3A%2F%2Ffif-moebel.de%2F&showimagereport=yes |
| 18:28 | <Fernandos> | I thought <meta http-equiv="language" content="de" /> is correct and neccessary for seo |
| 18:29 | <Fernandos> | I used media query like media="speech" with a speech stylesheet |
| 18:29 | <Fernandos> | but it tells me that is an invalid media query. |
| 18:29 | <Fernandos> | I know it was aural before and now it is speech, whyever.. |
| 18:30 | <TabAtkins> | CSS2 changed it from aural to speech, but it's possible that HTML didn't update itself. |
| 18:30 | <Fernandos> | last thing, <meta http-equiv="pics-Label"... is correct but why does it tell me it's an error? |
| 18:31 | <Fernandos> | I know it's all a working draft some even just experimental, but hey html5 is not alpha software |
| 18:31 | <Fernandos> | I respect the work you do and it's government grade quality or higher in usual. |
| 18:33 | <Fernandos> | any ideas |
| 18:34 | <TabAtkins> | Hmm. |
| 18:34 | <Philip`> | Fernandos: You might want http-equiv="content-language" instead |
| 18:34 | <gsnedders> | Only a few meta elements are allowed, almost all are ignored anyway. |
| 18:34 | <Philip`> | though that should give a warning telling you to use <html lang="de"> instead |
| 18:35 | <Philip`> | (See http://www.whatwg.org/specs/web-apps/current-work/multipage/semantics.html#attr-meta-http-equiv-content-language ) |
| 18:35 | <Fernandos> | Philip`: I just removed http-equiv="content-language" because it told me it is obsolete |
| 18:36 | <Fernandos> | I already use <html xmlns="http://www.w3.org/1999/xhtml"; lang="de"> |
| 18:37 | <Philip`> | HTML5 only allows a handful of http-equiv values, plus anything that's on http://wiki.whatwg.org/wiki/PragmaExtensions |
| 18:37 | <Philip`> | I don't know whether pics-label satisfies the criteria to be added to that page |
| 18:37 | <TabAtkins> | The media=speech issue seems to be a validator bug. HTML5 references MQ, which says that "speech" should be a recognized value. |
| 18:38 | <Philip`> | http://about.validator.nu/#reporting-bugs for validator bugs |
| 18:38 | <Fernandos> | TabAtkins: indeed. but the pics-label? |
| 18:38 | <Fernandos> | which is the official html5 validator? |
| 18:38 | <TabAtkins> | Fernandos: See Philip`'s response. |
| 18:39 | <Philip`> | Fernandos: There is no official HTML5 validator |
| 18:39 | <Philip`> | The idea is to encourage competition between validators, rather than blessing one as 'official' |
| 18:39 | <Philip`> | (though there's only one proper implementation so far) |
| 18:41 | <Fernandos> | that is? |
| 18:43 | <Philip`> | That implementation is used by validator.nu and validator.w3.org |
| 18:44 | <webben> | pics-label might be worth allowing |
| 18:44 | <webben> | only a few bits of software support it |
| 18:44 | <webben> | but one of them is MS Content Advisor. |
| 18:44 | <webben> | i read up about this a while back |
| 18:45 | <webben> | and apparently if you turn on Content Advisor and then don't configure it, it blocks webpages without the pics-label |
| 18:45 | <webben> | otoh you /can/ also just use the HTTP header so I guess it depends on how convenient one wants it to be. |
| 18:48 | <TabAtkins> | Filed a bug on the validator flagging media=speech as invalid. |
| 18:49 | <Fernandos> | thx TabAtkins |
| 19:52 | <Philip`> | Hmm, a microsoft.com comment on the canvas API |
| 19:53 | <Philip`> | Clearly a great opportunity to read too much into the existence of an email |
| 19:53 | <Philip`> | (http://lists.w3.org/Archives/Public/public-canvas-api/2009OctDec/0000.html for those not subscribed there) |
| 19:53 | <TabAtkins> | Never pass up an opportunity to misinterpret something. |
| 20:00 | <zcorpan_> | (?i)<script[^>]*>([^<]|<(?!/script[\s/>]))*<!--([^-]|-(?!->))*<script[\s/>]([^-]|-(?!->))*</script[\s/>] |
| 20:01 | <zcorpan_> | script close in escape *with* script open |
| 20:01 | <zcorpan_> | Philip`: any chance you could take that regexp for a ride? :) |
| 20:01 | zcorpan_ | hopes he got the regexp right |
| 21:22 | <zcorpan_> | " I think the HTML5 way of incorporating |
| 21:22 | <zcorpan_> | SVG and MathML is less magic in the sense that there are no spells for |
| 21:22 | <zcorpan_> | the author to cast." -- http://markmail.org/message/uywp4yfc4lvz6erd |
| 21:22 | <zcorpan_> | heh |
| 21:38 | <zcorpan_> | hsivonen: shouldn't the html5 dom viewer hide the Uhhhhhh syntax from the user? |
| 21:39 | <zcorpan_> | hmm maybe not, since it would be hard to figure out why scripts don't work as intended |
| 21:41 | <zcorpan_> | "Namespaces in HTML5" wouldn't make any sense, because HTML5 is layered on top of DOM Level 2 which is layered on top of Namespaces in XML which is layered on top of XML |
| 21:41 | <zcorpan_> | so Namespaces in HTML5 would have to violate layering and completely change HTML5's parsing rules |
| 21:41 | <zcorpan_> | if it's supposed to change namespaceURI of elements parsed from text/html |
| 21:43 | <TabAtkins> | It's an exciting new form of Namespaces. |
| 21:44 | <zcorpan_> | <com.netscape.blink> form of namespaces could be layered on top of HTML5 |
| 21:44 | <zcorpan_> | but that wouldn't change parsing or namespaceURI |
| 21:45 | <TabAtkins> | I have no idea why poor-man's namespaces is so unacceptable. It's easy and wonderful and great. |
| 21:46 | <TabAtkins> | I mean, I'm unbelievably frightened of a <calendar> element appearing in a Microsoft proposal. There will be something crazy like Outlook integration involved there. But <com.microsoft.calendar>? That seems downright sensible. |
| 21:49 | <zcorpan_> | doesn't ie8 also have a restriction that you can't use xmlns on elements whose tag name happens to match a known html tag name (like "table")? |
| 21:49 | <Philip`> | "Namespaces in XML" only works because everybody implements a Namespaces in XML processor, and nobody implements a plain XML processor |
| 21:50 | <Philip`> | so there's interoperability |
| 21:50 | <Philip`> | It'd be a nightmare if half the people implemented Namespaces in XML, and half didn't, and you had to write hacky scripts to parse raw xmlns:* attributes to get consistent application behaviour |
| 21:50 | <zcorpan_> | indeed |
| 21:51 | <zcorpan_> | the xml core wg should really merge those two specs, and remove the vanilla XML profile |
| 21:51 | <zcorpan_> | and the external subset |
| 21:51 | <zcorpan_> | and define more default entities |
| 21:52 | <Philip`> | I assume it originally made sense because they wanted to push XML to REC before finishing the Namespaces design |
| 21:52 | <Philip`> | but the layering seems potentially harmful from a technical perspective |
| 21:52 | <Philip`> | and therefore not a good thing to emulate |
| 21:52 | <zcorpan_> | and maybe move back to 4th ed productions, because 5th ed breaks with the rest of the toolchain so no-one wants to implement it |
| 21:54 | <jgraham> | But apart from that verything is good |
| 21:55 | <virtuelv> | the one bit about zcorpan_'s proposal that worries me is "define more default entities" |
| 21:55 | <TabAtkins> | I like entities... |
| 21:55 | <Philip`> | I like parsers that don't have hundred-kilobyte entity lookup tables |
| 21:55 | jgraham | cannot believe that we are still having the "should we include XML-style namespaces in HTML even though a large fraction of the people that deal with them in XML hate them" discussion |
| 21:55 | <zcorpan_> | virtuelv: either that or define magic FPIs like browsers do |
| 21:56 | <zcorpan_> | virtuelv: personally i'm fine with either but more default entities would be more author-friendly |
| 21:56 | <jgraham> | Philip`: It seems like a small, seldom paid cost |
| 21:56 | <zcorpan_> | i still don't understand why we have the mathml entities at all |
| 21:57 | <zcorpan_> | Hixie thought it was a simple thing to add them to text/html, so he just went ahead and did so? |
| 21:57 | <jgraham> | virtuelv: Shouldn't you be asleep or something? |
| 21:58 | <zcorpan_> | why did the mathml wg mint the entities in the first place? |
| 21:58 | <jgraham> | (at least I assume you are being woken up all night) |
| 21:58 | <jgraham> | zcorpan_: LaTeX |
| 21:58 | <Philip`> | zcorpan_: I expect they're necessary for the use-case of copy-and-pasting MathML markup (that currently exists, or is generated by current tools) into HTML |
| 21:59 | <zcorpan_> | jgraham: what about it? |
| 21:59 | <zcorpan_> | Philip`: i thought the main use case was to enable math input in a blog and have the blog software emit mathml markup |
| 22:00 | <jgraham> | zcorpan_: People who expect to author MathML by hand don't want to spend their whole life looking up the unmemorable codepoint for "counter clockwise contour integral" |
| 22:00 | <jgraham> | when they have already learnt the latex code for it |
| 22:00 | <jgraham> | so instead you make the LaTeX code work in HTML |
| 22:00 | <jgraham> | well in MathML |
| 22:01 | <zcorpan_> | but you don't write mathml by hand |
| 22:01 | <zcorpan_> | you write latex |
| 22:01 | <jgraham> | zcorpan_: I'm not sure that the MathML working group would agree with that position |
| 22:02 | <jgraham> | (actually since I have written MathML by hnd, I'm not sure I would agree with that position either. But I can't say the experience is one I will be looking to repeat) |
| 22:02 | <zcorpan_> | i've written mathml by hand, too |
| 22:03 | <zcorpan_> | i'm not going to do it again, either |
| 22:03 | <Philip`> | zcorpan_: Not many people have blog software that supports MathML, and more have e.g. Microsoft Word which can export equations as MathML |
| 22:03 | <zcorpan_> | Philip`: does Word emit mathml entities? |
| 22:04 | <Philip`> | zcorpan_: If I remember correctly, no |
| 22:04 | <zcorpan_> | are there any other commonly-used tools that emit mathml with entities? |
| 22:05 | <Philip`> | <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"; xmlns:m="http://schemas.openxmlformats.org/officeDocument/2006/math"><mml:msup><mml:mrow><mml:mi>e</mml:mi></mml:mrow><mml:mrow><mml:mi mathvariant="italic">iπ</mml:mi></mml:mrow></mml:msup></mml:math> |
| 22:05 | <Philip`> | (You'd just need a find-and-replace on "mml:" before pasting it into HTML) |
| 22:05 | <jgraham> | If you believe MathML is never written by hand the entities make no sense. But the theory is clearly that it is written by hand otherwise ideas like ⁢ would be nonsensical |
| 22:06 | <zcorpan_> | jgraham: indeed |
| 22:06 | <zcorpan_> | i think the entities make no sense :) |
| 22:06 | <jgraham> | (As, possibly, would the whole of conten MathML but maybe Mathematica or something can produce that) |
| 22:06 | <jgraham> | *content |
| 22:08 | jgraham | is quite sad at how many hard-to-hand-author languages have ended up on the web |
| 22:08 | <TabAtkins> | I'm happy CSS has stayed as easy as it is. |
| 22:08 | <jgraham> | Yeah CSS is more or less OK. Although selectors kinda sucks |
| 22:09 | <jgraham> | (for complex selectors) |
| 22:09 | <Philip`> | I find writing PNGs by hand to be a bit challenging |
| 22:09 | <Philip`> | Should have stuck with PPM |
| 22:09 | <TabAtkins> | Yeah, I find complex selectors pretty easy, but I know I'm atypical. |
| 22:10 | <jgraham> | Philip`: I think images are a different case. Argualy that excuses SVG too |
| 22:10 | <zcorpan_> | Philip`: btw, any chance you could do a dump for (?i)<script[^>]*>([^<]|<(?!/script[\s/>]))*<!--([^-]|-(?!->))*<script[\s/>]([^-]|-(?!->))*</script[\s/>] |
| 22:11 | <jgraham> | TabAtkins: The main problem is guessing which arbitarity-chosen symbol will be used for the operation you want |
| 22:11 | <Philip`> | zcorpan_: Oops, I forgot you asked already |
| 22:11 | <Philip`> | zcorpan_: I could probably try that later this evening (in maybe an hour), if I don't forget aain |
| 22:11 | <Philip`> | s//g/ |
| 22:11 | <zcorpan_> | ok |
| 22:11 | <TabAtkins> | jgraham: Ah, yeah, that's a very valid criticism. |
| 22:12 | <zcorpan_> | has anyone checked what the compat situation is with the new entities, btw? |
| 22:13 | TabAtkins | wishes IE would treat alt-text properly rather than calling it out as extra-special-missing-image-right-here. |
| 22:14 | <zcorpan_> | do pages use e.g. ℘ and expect it to be "℘" and not ℘ |
| 22:23 | <virtuelv> | jgraham: me, not so much. baby's being fed right now |
| 22:23 | <virtuelv> | zcorpan_: my problem with defining additional entities is that we don't know what kind of off-web breakage it will cause |
| 22:24 | <virtuelv> | and by "don't know", I mean we lack data, and means to collect them |
| 22:25 | <virtuelv> | I would like to think that authors haven't been braindead enough to use the HTML 4.01 defined entities for something else in their own formats |
| 22:25 | <virtuelv> | but I certainly wouldn't discount it |
| 22:28 | <virtuelv> | either way, now is bedtime |
| 22:33 | <zcorpan_> | virtuelv: yes, good point |
| 22:40 | <TabAtkins> | Ah. Interesting. Julian's gone for the "insane" option. |
| 23:21 | <zcorpan_> | Philip`: an hour has past :) |
| 23:26 | gsnedders | waits for "two hours have past" |
| 23:29 | <zcorpan_> | i guess the exit criteria for web idl should be two independent specs that use web idl and themselves have passed their exit criteria |
| 23:36 | <Hixie> | zcorpan_: for each feature |
| 23:37 | <zcorpan_> | Hixie: uh yeah |
| 23:38 | <zcorpan_> | i guess that might be too high a bar, given that it's unlikely that there'll be two specs for each feature |
| 23:38 | <zcorpan_> | maybe one spec for each feature |