| 11:57 | <gsnedders> | <http://diveintomark.org/archives/2008/03/09/no-fury-like-dracon-scorned#comment-11418>; — and someone asking whether he means it |
| 11:58 | <annevk> | yeah, funny |
| 11:59 | <annevk> | although the person asking if he means it points to a page talking about XML |
| 11:59 | <annevk> | I wonder if that person doesn't realize that that's XML-as-HTML or something |
| 12:10 | <gsnedders> | annevk: huh? what's XML-as-HTML? |
| 12:11 | <annevk> | anything XML (XHTML in this case) with a text/html media type... |
| 12:12 | <gsnedders> | annevk: I mean, in that context |
| 12:12 | <gsnedders> | annevk: the post he links to is about actual XML, Sam's blog that caused the issue is XML. What am I missing? |
| 12:12 | <annevk> | he points to something else in his comment |
| 12:12 | <annevk> | noah |
| 12:12 | <gsnedders> | annevk: that's about actual XML |
| 12:13 | <gsnedders> | annevk: it's all about things breaking due to trackbacks (which have no defined encoding) |
| 12:13 | <annevk> | no it's not |
| 12:13 | <annevk> | those pages have a text/html media type |
| 12:13 | <gsnedders> | Or at least the example given in it is XML? |
| 12:14 | <gsnedders> | those ones? at the time they didn't |
| 12:14 | <annevk> | they did |
| 12:14 | <gsnedders> | oh, I'm forgetting what's there. |
| 12:14 | <gsnedders> | The main body of the post is about actual XML |
| 12:17 | <gsnedders> | annevk: Noah's point I think, is that it can happen |
| 14:58 | <gsnedders> | ooo… Apache 2.2.8 (unlike 2.0.54) returns 501 for a FO"O method |
| 15:43 | <gsnedders> | token = %x21 / %x23 - %x27 / %x2A - %x2B / %x2D - %x2E / %x30 - %x39 / %x41 - %x5A / %x5E - %x7A / %x7C / %x7E |
| 15:43 | <gsnedders> | RFC2616 as ABNF. It doesn't map that easily. |
| 16:06 | <weinig> | Hixie: ping |
| 16:24 | <gsnedders> | does, "If a request being parsed, the server MUST respond with 400 (Bad Request); if a response being parsed, the client MUST notify the user of the error (e.g., by displaying a graphical error message, by sounding an audible warning, by logging the error in an error log).", seem reasonable? |
| 16:32 | <Philip`> | Logging an error in an error log does not notify the user |
| 16:46 | <gsnedders> | Philip`: That's really my issue. How do I phrase it any better, though? |
| 16:50 | <Philip`> | What happens if there isn't a user? (e.g. if your HTTP client is part of search engine) |
| 16:51 | <Philip`> | (*search engine web crawler or whatever they're called nowadays) |
| 16:52 | <gsnedders> | true. it probably doesn't make sense to either log it then. |
| 16:57 | <gsnedders> | SHOULD report the error? |
| 16:58 | <annevk> | search engines have users too |
| 16:58 | <annevk> | you should probably define different classes of products though |
| 16:58 | <annevk> | at least for clients |
| 16:59 | <annevk> | so you can be specific |
| 17:00 | <gsnedders> | what sort of classes? … and robots? |
| 17:00 | <gsnedders> | what can you call the non-robot class, though? |
| 17:00 | <gsnedders> | user-controlled user-agent |
| 17:01 | <annevk> | web browsers? |
| 17:01 | <gsnedders> | annevk: but things like cURL too |
| 17:01 | <annevk> | http://www.whatwg.org/specs/web-apps/current-work/multipage/section-conformance.html#conformance |
| 17:49 | <weinig> | annevk: hi |
| 17:50 | <weinig> | annevk: Do you think that the SECURITY_ERR you define in the XHR spec will be the same one Hixie eventually defines for HTML5? |
| 18:11 | <Hixie> | weinig: pong |
| 18:12 | <weinig> | hey Hixie, I was curious about the definition of a Security Exception |
| 18:12 | <Hixie> | yeah |
| 18:12 | <Hixie> | dunno yet |
| 18:12 | <Hixie> | probably the same as xhr's |
| 18:12 | <weinig> | Hixie: ok, that's fair |
| 18:13 | <weinig> | Hixie: so adding it to the list of Core DOM ones |
| 18:13 | <weinig> | as numero 18 |
| 18:14 | <weinig> | Hixie: Firefox currently uses code 1000 and Opera uses code undefined for Canvas tainting security exceptions |
| 18:16 | <annevk> | weinig, that's my plan anyway :) |
| 18:16 | <SadEagle> | weinig: heh, I used INVALID_ACCESS_ERR :-) |
| 18:18 | <Hixie> | weinig: 18 seems fair to me, it's listed on http://wiki.whatwg.org/wiki/Exception_Codes |
| 18:22 | <weinig> | Hixie: cool |
| 19:07 | <annevk> | hsivonen, re: "elements are case-sensitive", for real? that's not really consistent |
| 19:08 | <hsivonen> | annevk: if we were dealing only with MathML, I'd agree with you |
| 19:08 | <Hixie> | i wonder if IE's implementation of namespaces is actually going to be a problem for IE implementing what hsivonen is suggesting |
| 19:09 | <annevk> | IE would need an XML DOM for HTML... that might be tricky for them |
| 19:09 | <annevk> | "XML DOM" |
| 19:09 | <hsivonen> | annevk: but keeping DOM consistency with camel casing is not feasible with allowing the SVG WG do their thing without a lookup table in the parser |
| 19:10 | <Hixie> | lookup table it is, then ;-) |
| 19:10 | <hsivonen> | annevk: other than case-sensitivity, how far did my email go off the rails? |
| 19:10 | <Hixie> | this namespace thing is gonna suck so much. whatever i end up doing, 70% of people will be pissed at me |
| 19:12 | <annevk> | rest seems fine |
| 19:12 | <hsivonen> | annevk: thanks |
| 19:12 | <annevk> | Hixie, at least 90% of the Web community won't care :) |
| 19:13 | <Philip`> | Hixie: write multiple competing specifications, so that everyone is happy with at least one, and then let implementors choose which one will survive |
| 19:13 | <Philip`> | and then abuse will be aimed at implementors for making the wrong choice, rather than at you |
| 19:16 | <Dashiva> | Nah, he's the one who'll have to enter it into the standard :) |
| 19:18 | <Hixie> | Philip`: heh |
| 19:20 | <Hixie> | man i wish people would use the '-- ' signature convention |
| 19:20 | <Hixie> | i'm tired of having to cut out signatures :-P |
| 19:25 | <Hixie> | good lord |
| 19:25 | Hixie | cuts out a huge discussion on the relative merits of <dl> vs <table> |
| 19:39 | <annevk> | hsivonen, so my thinking is that a hardcoded lookup table is not that bad as new elements will always require some work |
| 19:40 | <hsivonen> | annevk: that's true |
| 19:41 | <Dashiva> | What if: SVG had used lowercase |
| 19:41 | <Dashiva> | Would things be a lot easier, or just a tiny bit? |
| 19:41 | <annevk> | tiny bit |
| 19:42 | <annevk> | lookup tables are not intrinsically hard |
| 19:42 | <annevk> | it would be a lot better though, from a consistency point of view with the rest of the Web platform |
| 19:43 | <Philip`> | New elements may require only work that can be emulated by a DOM-manipulating script library for UAs that don't properly support the new feature yet |
| 19:43 | <Philip`> | but that doesn't work if the new element won't be parsed properly |
| 19:45 | <annevk> | so introduce lowercase elements :) |
| 19:59 | <virtuelv> | tangentially relevant to html 5, http://schrodinger.sourceforge.net/press/DiracAnnounce060308.html |
| 20:01 | <Dashiva> | virtuelv: You mean the encoding error? ;) |
| 20:01 | <virtuelv> | :p |
| 20:05 | hsivonen | notes "Dirac professional profiles are currently in the process of being standardized by SMPTE as VC-2 and standardization of full Dirac is planned for this year. " |
| 20:06 | <annevk> | that's better than Ogg Theora, no? |
| 20:07 | <roc> | yeah |
| 20:07 | <roc> | patent status unknown, though |
| 20:07 | <roc> | and likely to be bad |
| 20:08 | <annevk> | I thought they claimed they had that covered somehow? |
| 20:08 | <annevk> | probably too good to be true |
| 20:09 | <roc> | not as far as I know |
| 20:09 | <roc> | theora's also not looking good |
| 20:09 | <annevk> | "While the BBC own some patents on Dirac, they have irrevocably granted a royalty-free licence for their Dirac-related patents to everyone. In addition, the BBC have checked (by extensive patent search) that Dirac does not infringe any third party patents, enabling the public to use Dirac for any imaginable purpose." |
| 20:10 | <annevk> | from Wikipedia |
| 20:10 | <roc> | oh really? |
| 20:10 | <roc> | interesting |
| 20:10 | <annevk> | I remember having read that on the Dirac FAQ too |
| 20:10 | <virtuelv> | I think the patent thing was one of Dirac's driving forces |
| 20:10 | <roc> | I'll ask Chris to look into it |
| 20:11 | <hsivonen> | annevk: see Wikipedia talk page |
| 20:11 | <roc> | but I'm sure it's already been looked into so the situation can't be as simple as that |
| 20:11 | <Hixie> | my understanding is that dirac isn't an especially compelling codec, but i'm certainly no expert |
| 20:11 | <annevk> | http://dirac.sourceforge.net/faq.html#a6B |
| 20:12 | <annevk> | sketches it slightly differently |
| 20:13 | <jgraham_> | Hixie: Did you say you were working tables at the moment? Will it be any help if I post the latest iteration of the smart span algorithm for table headers to the list sooner rather than later? |
| 20:13 | <Hixie> | yes |
| 20:13 | <Hixie> | and yes |
| 20:13 | <hsivonen> | unsourced statements on wikipedia aren't of best thing for proving absence of patent threats |
| 20:13 | <hsivonen> | s/of best/the best/ |
| 20:13 | <jgraham_> | hsivonen: It's nice to come home and find that someone has already written a long email for you :) |
| 20:14 | <jgraham_> | s/for you/so you don't have to/ |
| 20:14 | <hsivonen> | jgraham_: :-) |
| 20:14 | <jgraham_> | Hixie: OK, will do |
| 20:16 | <jwalden> | need the baseline necessarily be hyper-compelling? |
| 20:16 | <jwalden> | something moderately competitive, even if not top-of-the-line, seems good enough |
| 20:16 | <jwalden> | at least for starters |
| 20:18 | <gsnedders> | jwalden: it needs to be good enough to get implemented, and be known not to have patents |
| 20:19 | roc | whistles |
| 20:19 | roc | wishes this whole issue wasn't mired in secrecy |
| 20:20 | <jwalden> | finding codecs with implementations isn't really a problem |
| 20:21 | <jwalden> | and they claim they're on the patent thing |
| 20:21 | <jwalden> | or we can fud ourselves into doing nothing |
| 20:22 | <roc> | we're beyond FUD |
| 20:27 | <gsnedders> | "The tokeniser/parser MUST stop processing immediately. If a request is being parsed, the server MUST respond with 400 (Bad Request); if a response is being parsed, the client SHOULD report the error." — Does that sound all right? |
| 20:44 | <Philip`> | What does "report the error" mean? |
| 20:48 | <gsnedders> | yeah, I thought copying what XML says probably wasn't the best idea :) |
| 21:36 | <Philip`> | http://diveintomark.org/archives/2008/03/09/no-fury-like-dracon-scorned#comment-11430 - U+FFFF wins again! |
| 21:38 | <gsnedders> | Philip`: it's too easy. it's not even fun trying nowadays |
| 21:39 | <Philip`> | I still can't resist when someone makes an explicit challenge :-( |
| 21:39 | <gsnedders> | Philip`: So childish :P |
| 21:39 | <annevk> | Philip`, awesome :) |
| 21:39 | <hasather> | haha |
| 21:40 | <hsivonen> | Philip`++ |
| 21:41 | <annevk> | did you take a screenshot? |
| 21:42 | <Philip`> | annevk: No |
| 21:43 | <gsnedders> | http://stuff.gsnedders.com/plasmasturm_U+FFFF.pdf |
| 21:44 | <Philip`> | I'm not sure Safari's error reporting is an improvement over Firefox's or Opera's |
| 21:44 | <gavin_> | did that guy take down his site because of the problem you caused? |
| 21:44 | <Philip`> | I don't get that pink box when I look at it myself, but the error text is similarly unreadable |
| 21:45 | <gsnedders> | make that .png |
| 21:45 | <gsnedders> | paparazzi is less than ideal at exporting to pdf |
| 21:45 | <hsivonen> | http://hsivonen.iki.fi/screen/plasmasturm-ffff.png |
| 21:45 | <hsivonen> | a .png |
| 21:46 | <gsnedders> | http://stuff.gsnedders.com/plasmasturm_U+FFFF.png |
| 21:46 | <gsnedders> | a .png :) |
| 21:47 | <Philip`> | At least the page renders fine in Konqueror |
| 21:48 | <roc> | doesn't that make Konqueror evil? |
| 21:48 | <Philip`> | Yes |
| 21:48 | <Philip`> | (As far as I'm aware, it uses a text/html parser for application/xhtml+xml) |
| 21:48 | SadEagle | hides |
| 21:48 | <SadEagle> | I blame QXML! |
| 22:02 | jgraham__ | wonders where he lost all his table headers tests |
| 22:03 | <jgraham__> | Lesson learnt: revision control is less useful when you've only made imaginary checkins |
| 22:06 | <Dashiva> | Kinda like how keeping backups on the same drive as the original data isn't really backed up |
| 22:17 | <Philip`> | (I suppose I should refrain from pointing at http://blog.codedread.com/?s=%00 and http://golem.ph.utexas.edu/cgi-bin/MT-3.0/mt-search.cgi?Template=%00 ) |
| 22:21 | <jgraham__> | Philip`: Clearly someone needs to flip your evil bit :) |
| 22:22 | gsnedders | flips Philip`'s evil bit |
| 22:29 | <Philip`> | I'm not sure whether that means my evil bit is currently set or not |
| 22:30 | <jgraham__> | Does that make it an evil qbit? |
| 22:33 | Hixie | looks at Philip` to collapse his wavefunction |
| 22:34 | Philip` | refuses to collapse, and instead entangles Hixie |
| 22:34 | <Hixie> | yikes |
| 22:34 | <gsnedders> | this isn't looking good. |
| 22:35 | <SadEagle> | hixie's evil bit is always set, though |
| 22:35 | Hixie | wonders what would happen if he looked at _himself_ |
| 22:35 | <Dashiva> | Hixie: Ask the cat |
| 22:36 | <Hixie> | SadEagle: so that means Philip`'s is unset, right? |
| 22:36 | <Hixie> | assuming conservation of evilness |
| 22:36 | <SadEagle> | aren't you the physicist here, or am I confusing things? |
| 22:36 | <Hixie> | jgraham__ and Philip` are more physicsy, i |
| 22:36 | <Hixie> | have been out of hte loop for years |
| 22:36 | <Philip`> | Conservation of evilness doesn't tell you anything unless you know how much evil the closed system started with |
| 22:37 | <Hixie> | when then how does entangling me help? |
| 22:37 | gsnedders | needs to decide whether to continue to be physicsy after next year |
| 22:37 | <Philip`> | I'm not physicsy, I only did it for a year and then didn't mind abandoning it :-) |
| 22:37 | <Hixie> | wait i thought you were phiscsy right now? |
| 22:38 | <Philip`> | I'm computery |
| 22:38 | <Hixie> | oh ok |
| 22:38 | <jgraham__> | I only did one course on quantum computation, which was a while ago |
| 22:39 | Philip` | just about understood some of the quantum maths stuff, but totally failed to follow the explanations of quantum algorithms |
| 22:39 | <jgraham__> | But I think if Hixe and Philip have entangled qubits you could entangle them to have the same value or different values |
| 22:40 | <jwalden> | mm, qubits |
| 22:40 | jwalden | longs to play qubit in a Scrabble game |
| 22:40 | gsnedders | has never done quantum physics/computation |
| 22:40 | <Philip`> | You could have an entangled state like (|00>+|11>)/sqrt(2) so they'd both be the same value, or like (|01>+|10>)/sqrt(2) so they'd be different |
| 22:40 | <gsnedders> | and the only time I've tried to get my head around it, I came out understanding less than when I started |
| 22:42 | <Philip`> | gsnedders: It's kind of fun in a slightly insane way :-) |
| 22:43 | gsnedders | ought to email you and James about May, now I know exactly when I'll be around |
| 22:43 | Philip` | has no idea exactly when he'll be around |
| 22:44 | <annevk> | Hixie, on the Window IDL frames should be moved under "self" |
| 22:44 | <annevk> | Hixie, it's not about other browsing contexts per se as it simply points to itself |
| 22:44 | <Philip`> | (except that I'm rarely not around, and tend not to have any kind of inflexible scheduling, so things tend to be alright) |
| 22:44 | <gsnedders> | Philip`, jgraham__: the Saturday will be taken up more or less entirely with my Grandmother's memorial service, and being with family — the only possible time would be before lunch, but I really want to head around Cam. then. I'm going back on Sunday, which leaves on Friday evening |
| 22:45 | <gsnedders> | s/on/only/ |
| 22:45 | <jgraham__> | gsnedders: Friday evening will be fine for me (as in "this is long enough before May that, barring any unexpected significant event, I can make it fine") |
| 22:46 | <gsnedders> | jgraham__: well, if something huge comes up (or the train is _really_ late, then I couldn't go either :P |
| 22:47 | <gsnedders> | (but when I say _really_ late, I mean _I've never seen it that late_ late, when I've gone to Cam. by trains far too many times) |
| 22:47 | Philip` | is in the same situation as jgraham__ |
| 22:48 | <gsnedders> | which leaves location, but that's probably better dealt with in May :P |
| 22:48 | <jgraham__> | gsnedders: Sure. Location can be improvised |
| 22:49 | <gsnedders> | (it would probably suit me best to have dinner/supper/whatever-you-want-to-call-it with you two) |
| 22:50 | <jgraham__> | Also fine. |
| 22:50 | <jgraham__> | (for me) |
| 22:50 | Philip` | sometimes calls it lunch, depending on when he wakes up |
| 22:50 | <gsnedders> | But there are plenty of places to eat in Cambridge |
| 22:50 | <gsnedders> | Philip`: oh, my mother always calls lunch/supper the wrong way around |
| 22:55 | <annevk> | jgraham__, looking through your code, does it work with scope=ROW ? |
| 22:55 | <jgraham__> | annevk: No, I noticed that bug when I wrote the email |
| 22:55 | <jgraham__> | I suppose I could check in the fix :) |
| 22:55 | gsnedders | waves good night |
| 22:55 | <jgraham__> | nn |
| 22:57 | <annevk> | i see |
| 23:07 | <Hixie> | annevk: send mail, i'm in meetings a lot right now |
| 23:15 | <annevk> | done |
| 23:40 | <Hixie> | thanks |
