| 00:00 | <shelleyp> | People will be starting to create HTML5 that no browser currently maintains, and they'll want to know if it is future valie |
| 00:00 | <shelleyp> | Sorry, valid |
| 00:01 | <shelleyp> | If the spec is precise enough, no browser has to implement it in order to test for conformity |
| 00:03 | <shelleyp> | Now the issue is: is XML within SVG conforming or not, in HTML? |
| 00:03 | <shelleyp> | Because if it isn't, then it's not very useful |
| 00:04 | <shelleyp> | Does anyone else have any questions they would like to ask me specifically on my recent email to HTML WG? |
| 00:06 | <shelleyp> | Does anyone have anything they want to say directly to me? To have a real discussion? |
| 00:07 | <shelleyp> | I thought not |
| 00:10 | <TabAtkins> | ... |
| 00:11 | <Lachy> | sorry shelley, was afk. The answer to your question is that no, other XML markup is not conforming in SVG, other than xlink attributes and xml:id, xml:lang and xml:space |
| 00:11 | <Lachy> | guess she'll have to read that in the logs. |
| 00:14 | <TabAtkins> | I wonder if she understands that IRC is only slightly less asynchronous than email? |
| 00:26 | <shelleyp> | Lachy: thanks for the response. That's extremely disappointing. A show stopper. |
| 00:29 | <shelleyp> | Lachy: That's not clear in the document either. I don't want to file a bug, though, because I don't agree with this behavior |
| 00:29 | <roc> | SVG in HTML is still incredibly useful without that |
| 00:29 | <shelleyp> | roc: As much as I'm a supporter of SVG, I don't necessarily agree |
| 00:30 | <shelleyp> | roc: most of the SVG you can get that's public domain has CC in RDF/XML |
| 00:34 | <webben> | shelleyp: So the concern is you wouldn't be able to paste such clipart into an HTML5 doc? |
| 00:35 | <webben> | e.g. http://openclipart.org/people/gubrww/gubrww_gold_nugget_brooch.svg for example? |
| 00:35 | <shelleyp> | webben: Yes, that's a real issue. Especially if the CC license doesn't allow modification of the SVG |
| 00:36 | <shelleyp> | webben: it's more than that, though. I tried out SVG in HTML5 with the FF nightly, and the external XML works beautifully |
| 00:37 | <webben> | shelleyp: Potentially stupid question. Looking at that SVG source, doesn't rdf:about="" point to the current document URL? |
| 00:37 | <webben> | wouldn't that mean, if you did paste it into a host XML document, the rdf:about would start pointing to the current doc rather than the embedded SVG fragment? |
| 00:38 | <roc> | shelleyp: I agree that's a real issue, but it's going to be really useful in any case |
| 00:38 | <shelleyp> | webben: No, because the SVG document is loaded into its own SVGdocument object |
| 00:38 | <webben> | okay. |
| 00:38 | <webben> | but what's the url of that object? |
| 00:39 | <webben> | or rather doesn't "" still resolve to the same URL? |
| 00:39 | <shelleyp> | webben: What would it be, if you embed the same SVG in an XHTML document, which is valid in HTML5 land? |
| 00:40 | <webben> | That's approx. what I'm asking, yep. |
| 00:42 | <rubys2> | shelleyp: why don't you want to file a bug? |
| 00:43 | <shelleyp> | rubys2: What would be the bug? That we can't use external vocabularies? Or that it's not well documented that we can't? |
| 00:44 | <webben> | 2 bugs? |
| 00:45 | <rubys2> | shelleyp: whichever one you would advocate. |
| 00:46 | <shelleyp> | rubys2: OK |
| 00:47 | <rubys2> | as far as I can tell, you aren't asking browsers to behave differently, you simply don't want conformance checkers to complain when you copy and paste a typical SVG document produced by inkscape or whatever into a text/html page. |
| 00:47 | <shelleyp> | rubys2: Yes, that's it. I'm not asking for changes to the DOM because of namespaces. I'm just asking that browsers work like the FF nightly worked |
| 00:49 | <rubys2> | If spec'ed, I'm willing to make the changes to validator.nu. |
| 00:49 | <AryehGregor> | Ah, I see the response to my objection starts somewhere around line 2500. |
| 00:49 | <shelleyp> | rubys2: But the only way for the page to be conforming, if I understand it, is to codify this behavior in the HTML5 spec, yes? |
| 00:50 | <rubys2> | The spec does need to be clarified one way or another. And I find the copy/paste from inkscape argument compelling. |
| 00:51 | <shelleyp> | rubys2: I think it's very compelling. I actually assumed this was the behavior, and wrote an article for ALA, which I now need to pull because of this assumption |
| 00:52 | <rubys2> | I wouldn't pull it... simply annotate it as an open bug report and/or issue. |
| 00:52 | <rubys2> | (and I would suggest going with a bug report to start) |
| 00:53 | <shelleyp> | rubys2: OK, will do a bug. I want to go read that line 2500 AryehGregor just pointed out |
| 00:53 | <AryehGregor> | shelleyp, http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2008-April/014372.html |
| 00:53 | <shelleyp> | AryehGregor: thanks -- possible search term I can look for? |
| 00:54 | <AryehGregor> | It's scattered. But a lot of the interesting stuff on HTML vs. XML syntax for embedded SVG seems to start around "I don't think it would take magic". |
| 00:54 | <AryehGregor> | I didn't pay much attention to the namespaces issue, though. |
| 00:55 | <AryehGregor> | I was just reading the whole post. |
| 00:56 | <AryehGregor> | . . . am just reading the whole post, that is. |
| 00:56 | AryehGregor | predicts it will take another 10 minutes or so to finish |
| 00:58 | AryehGregor | is amused by the "Maths is a key part of our heritage" argument instead of the usual "It's a proven standard that implementors have shown interest in incorporating into HTML" |
| 00:59 | <rubys2> | quickly scanning that email, I don't see anything that backs up Henri's assertion. |
| 00:59 | <AryehGregor> | Which one? |
| 01:00 | AryehGregor | finishes reading |
| 01:01 | <AryehGregor> | Hmm, I was pessimistic, more like 4.5 minutes. |
| 01:01 | <rubys2> | I may have missed it, but I don't see anything in http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2008-April/014372.html that supports http://twitter.com/hsivonen/status/3797923550 |
| 01:02 | <AryehGregor> | Well, that e-mail isn't normative, the spec is normative. |
| 01:02 | <shelleyp> | rubys2: I've been looking too, primarily starting later in the document, and I can't see a definitive answer on that either |
| 01:03 | <rubys2> | aryehGregor: I don't see anything in the spec that backs up Hsivonen's assertion either. |
| 01:03 | <shelleyp> | rubys2: hence my confusion on this subject, and surprise with getting the dc:foo answer |
| 01:04 | <AryehGregor> | I'd assume it's somewhere in the bowels of section 9. |
| 01:04 | <AryehGregor> | http://www.whatwg.org/specs/web-apps/current-work/#parsing-main-inforeign |
| 01:04 | <rubys2> | it doesn't surprise me that hsivonen believes it, but I don't believe it has been decided yet either way. |
| 01:05 | rubys2 | is not sure he can load the one page version of the spec on this netbook... |
| 01:05 | AryehGregor | isn't going to try reading chapter 9 |
| 01:06 | <AryehGregor> | Er, sorry about that. |
| 01:06 | <AryehGregor> | I don't usually use it myself. |
| 01:06 | <AryehGregor> | Only when I'm doing a search of some kind. |
| 01:06 | <AryehGregor> | http://www.whatwg.org/specs/web-apps/current-work/multipage/syntax.html#parsing-main-inforeign |
| 01:06 | <AryehGregor> | Possibly somewhere there. |
| 01:08 | <shelleyp> | * Shelley frantically waves rubys2 away from loading the one page WhatWG spec on his netbook -- visions of burned wires |
| 01:09 | <AryehGregor> | You can say /me to do the action thing. |
| 01:09 | <AryehGregor> | /me demonstrates. |
| 01:09 | AryehGregor | demonstrates |
| 01:09 | <AryehGregor> | . |
| 01:09 | <AryehGregor> | (the extra newline wasn't intentional) |
| 01:10 | <TabAtkins> | rubys2: If you didn't know, you can always just take the anchor from a link into the 1-page document, append it to the root url of the multipage document, and a script'll pick it up and take you to the correct place in the multipage. |
| 01:11 | shelleyp | thanks AryehGregor |
| 01:11 | <AryehGregor> | TabAtkins, sneaky. |
| 01:11 | <TabAtkins> | I just found about it a few weeks ago. |
| 01:11 | <webben> | 9.1.2 says only HTML, MathML and SVG elements are "allowed". |
| 01:12 | <webben> | (http://dev.w3.org/html5/spec/syntax.html) |
| 01:12 | <rubys2> | neither 9.2.4.5 nor 9.2.4.7 disallow colons |
| 01:13 | <Lachy> | what exactly are you trying to find out from the spec? |
| 01:13 | <rubys2> | ok, so a bug report on 9.1.2 citing the use case of copy/paste from inkscape would be in order |
| 01:14 | <shelleyp> | rubys2: OK, will do |
| 01:14 | <Lachy> | scanning the logs quickly didn't make it clear what you're looking for |
| 01:14 | <webben> | Lachy: Where dc:foo is disallowed. |
| 01:14 | <webben> | Lachy: as per http://twitter.com/hsivonen/status/3797923550 |
| 01:15 | <rubys2> | Lachy: more context: http://twitter.com/shelleypowers/status/3801626489 |
| 01:15 | <Lachy> | presumably it's not allowed because it's an attribute in no namespace with the localname "dc:foo", as opposed to "foo" attribute in the "dc" namespace, and I guess SVG doesn't allow non-namespaced, undefined attributes |
| 01:16 | <Lachy> | but that's just a guess. I'm not really familiar with SVG's conformance requirements on the issue |
| 01:16 | <webben> | oh dc:foo is an attribute not an element? |
| 01:16 | <webben> | if so 9.1.2 doesn't apply |
| 01:17 | <rubys2> | webben: shelley's example has plenty of elements |
| 01:17 | <Lachy> | oh, if you're asking about elements, then it should be clear that only HTML, SVG and MathML elements are permitted |
| 01:17 | <webben> | I'm not sure what shelleyp's example was. My openclipart example has elements too. |
| 01:17 | <webben> | Just talking about dc:foo |
| 01:18 | <webben> | The spec seems less obvious about elements. |
| 01:18 | <rubys2> | it would be helpful if such usages weren't flagged... again, nobody is asking that the browser do anything different with such elements and attributes. |
| 01:18 | <webben> | *about attributes |
| 01:18 | <Lachy> | I just assumed you were talking about an attribute, since I didn't think the spec was in any way ambiguous about undefined elements not being permitted. |
| 01:19 | <webben> | okay |
| 01:21 | <rubys2> | Lachy: if svg were to allow "dc:foo" (note: with the colon) it wouldn't be undefined. So the bug report could go against svg. |
| 01:22 | rubys2 | thinks this is all good fodder for his svgopen keynote :-) |
| 01:23 | <Lachy> | rubys2, then you would presumably run into issues about defining elements with localnames containing colons, which clash with XML namespaces |
| 01:23 | shelleyp | thinks that how SVG will work with HTML in HTML5 will generate lots of interest |
| 01:25 | <rubys2> | Lachy: http://www.w3.org/TR/html-design-principles/#priority-of-constituencies |
| 01:25 | <rubys2> | users will *definitely* copy/paste documents from inkscape into HTML and expect it to "just work" |
| 01:25 | <Lachy> | it will just work. Just won't be conforming. |
| 01:25 | <Lachy> | I don't see why flagging useless markup is a bad thing |
| 01:26 | <rubys2> | Can I give every last one of them your phone number? :-) |
| 01:26 | <AryehGregor> | Lachy, useless markup is conforming in a lot of other cases to ease interoperability with existing content. |
| 01:26 | <rubys2> | it can all be a talisman for all I care. |
| 01:26 | <AryehGregor> | Like xmlns, xml:lang, etc. (at least where those were required in XHTML). |
| 01:27 | <AryehGregor> | (required or permitted or whatever) |
| 01:27 | <Lachy> | AryehGregor, yes, but only in a few very specific cases. Allowing random junk doesn't seem like such a good idea |
| 01:27 | <AryehGregor> | No opinion here. |
| 01:27 | <rubys2> | things intentionally put in SVG to capture license information (for example) hardly qualifies as "random junk" |
| 01:28 | <rubys2> | BBIAB. Will check logs. |
| 01:28 | <takkaria> | I think it's pretty important to warn about dc:* if they're not put in the right namespace, since that's not what you would expect |
| 01:28 | <AryehGregor> | If there are particular examples that are very common, it seems reasonable to special-case those. |
| 01:28 | <AryehGregor> | At least by HTML 5's philosophy of compatibility with existing content even at the expense of ugly or useless markup. |
| 01:30 | <Lachy> | in the few limited cases where useless talismans are permitted, they would have virtually no effect even if they did work. e.g. xmlns="" need to have the value of the namespace in which the element is already placed |
| 01:30 | <TabAtkins> | OMG I'M DONE WITH IT. |
| 01:30 | <TabAtkins> | http://www.xanthir.com/test.php?width=200&height=100&radial-gradient=top 20px left 20%, ellipse cover,yellow 20px,green,red,blue |
| 01:30 | <takkaria> | division by zero at like 221? |
| 01:30 | <AryehGregor> | Is that supposed to be a URL? |
| 01:30 | <Lachy> | that wouldn't be the case with allowing <dc:*> and <cc:*> elements, since it would create the false impression that they actually do anything useful in HTML |
| 01:30 | <TabAtkins> | That's like 12 hours of work getting that to work. |
| 01:30 | <TabAtkins> | Yeah, copy the whole thing. |
| 01:31 | <TabAtkins> | http://www.xanthir.com/test.php?width=200&height=100&radial-gradient=top%2020px%20left%2020%,%20ellipse%20cover,yellow%2020px,green,red,blue |
| 01:32 | <AryehGregor> | Is that moving, or is it just my imagination? I *think* it's just my imagination, but . . . |
| 01:32 | <takkaria> | Lachy: the uselessness is fine, I think the thing that's important is that they'd be in the DOM in a very different form to what one would expect |
| 01:32 | <TabAtkins> | Just your imagination. ^_^ The eye-raping colors are good for testing. |
| 01:32 | <Lachy> | takkaria, yeah, that too |
| 01:42 | <TabAtkins> | Well, correction, there are a couple of important bugs in my parser that need fixing. But, given appropriate inputs, it displays as intended! |
| 02:07 | <AryehGregor> | Okay, so why is "Attribute size not allowed on element input at this point"? <input id=mw-input-contextchars name=wpcontextchars size=20 value=50 min=0 type=number> |
| 02:07 | <AryehGregor> | Is size invalid on type=number, or is the validator wrong? |
| 02:08 | <Hixie> | http://www.whatwg.org/specs/web-apps/current-work/#input-type-attr-summary |
| 02:09 | <AryehGregor> | So why is that? Presumably in some cases you'd expect the user to be entering a number like 6, and other times you'd expect them to want to enter floating-point numbers to 47 decimal places. |
| 02:10 | <AryehGregor> | Of course, size is kind of dodgy anyway, I always felt it seemed more appropriate to do with CSS. |
| 04:01 | <Hixie> | wtf is the API that does the equivalent of getElmentsBySelector() called? |
| 04:01 | <Hixie> | selectNodes()? |
| 04:01 | <Hixie> | oh, querySelector(). |
| 06:58 | <zcorpan> | Philip`: http://webreflection.blogspot.com/2009/09/font-face-we-are-already-doing-wrong.html |
| 07:51 | <hsivonen> | I see my spec reading ability has been questioned in the logs. |
| 07:55 | <othermaciej> | it has? |
| 07:58 | Philip` | questions othermaciej's IRC log reading ability |
| 07:58 | <othermaciej> | I'm reading the logs now, but I'm not sure I'll be able to figure out what hsivonen was referring to |
| 08:00 | <othermaciej> | the Dublin Core controversy? |
| 08:01 | <hsivonen> | othermaciej: yes |
| 08:01 | <othermaciej> | does HTML5 specify anywhere that any SVG elements or attributes are valid, or does it defer that completely to SVG? |
| 08:01 | <hsivonen> | othermaciej: it defers to SVG |
| 08:02 | <othermaciej> | hsivonen: through the "other applicable specifications" clause, or in some more explicit way? |
| 08:02 | <hsivonen> | othermaciej: but SVG 1.1 requires the stuff that goes in <metadata> to be expressed "in a manner conforming with the "Namespaces in XML" Recommendation", and having a non-NCName local name doesn't qualify |
| 08:02 | <hsivonen> | othermaciej: it's basically like other applicable specs, except only "in a manner conforming with the "Namespaces in XML" Recommendation" |
| 08:03 | <hsivonen> | maybe there's list email by now that I can reply to |
| 08:03 | <othermaciej> | there were some public-html posts that I assume were on this topic |
| 08:04 | <othermaciej> | it seems to me that 'expressed in a manner conforming with the "Namespaces in XML" Recommendation' is not a very clear rule as applied to SVG-in-text/html |
| 08:06 | <othermaciej> | it's not even very precise in the context of XML - I would guess it means something like namspace well-formed, but that's not meaningful for a fragment |
| 08:07 | <othermaciej> | however, it does seem to me that attributes and elements which parse into different {namespaceURI, localName} pairs in HTML and XML are problematic |
| 08:08 | <othermaciej> | besides making them clearly nonconforming inside <svg:metadata>, another possibility is to just apply xmlns processing inside <metadata> or even just inside <svg>, but then presumably people would use it as a hook to inject arbitrary namespaced content |
| 08:13 | <hsivonen> | I've posted about this to public-html before, but Shelley wasn't a WG participant back then. |
| 08:13 | <hsivonen> | I'll look up the archive URL. |
| 08:13 | <hsivonen> | short version: there are multiple possible solutions but they are all bad |
| 08:14 | <othermaciej> | that seems par for the course |
| 08:14 | <othermaciej> | shelley is technically not a WG participant now, but she did post to the list on this topic |
| 08:27 | <erlehmann> | hsivonen, i dont really get why the current solution to have SVG content inside an iframe or image is not preferred |
| 08:29 | <Hixie> | (me either) |
| 08:34 | <othermaciej> | image isn't very good if you want to change the svg dynamically or have interactive features in it |
| 08:35 | <othermaciej> | iframe does allow those, although it requires your svg to live in a defined opaque rectangular box |
| 08:36 | <othermaciej> | that being said, there are many use cases for vector graphics that do work fine with those and don't need inline svg |
| 08:36 | <hsivonen> | Hixie: clearly, the "how to read this specification" part doesn't teach people that reading a Hixie spec isn't like reading a Dave Winer spec |
| 08:37 | <Hixie> | do you want bold 96px text that says "don't read between the lines"? |
| 08:39 | <othermaciej> | hsivonen: your statement that SVG and MathML requirements defined in terms of XML formalisms are to be applied to the DOM after parsing from text/html does not seem overwhelmingly clear from any spec |
| 08:39 | <Philip`> | 96px text is easy to miss when it just occurs once in a giant document |
| 08:40 | <Philip`> | Maybe you should put it in 2px text located between every line |
| 08:40 | <othermaciej> | hsivonen: I do believe that is the intent of HTML5, but one could just as easily assume SVG and MathML requirements are applied at a source text level instead of a DOM level, or something |
| 08:41 | <Hixie> | Philip`: hah |
| 08:41 | <othermaciej> | short of some clear spec language to say otherwise |
| 08:49 | <Hixie> | woo |
| 08:50 | <othermaciej> | woo? |
| 08:50 | <Hixie> | outstanding e-mails are now below outstanding bugs and outstanding XXX markers. And I blogged. |
| 08:52 | <othermaciej> | you catblogged! |
| 08:53 | <hsivonen> | othermaciej: I'll post proof to the list shortly. |
| 08:54 | <Hixie> | yeah well most of the things i'd like to say about subjects y'all would be interested in would likely just piss people off |
| 08:54 | <hsivonen> | othermaciej: evidently, it's not overwhelmingly clear to everyone |
| 08:55 | <hsivonen> | othermaciej: it is overwhelmingly clear from the implementation POV |
| 08:55 | <hsivonen> | othermaciej: because the code I have for RDF/XML simply won't run in the text/html case |
| 08:55 | <othermaciej> | hsivonen: it's clear what to do from the implementation POV, yes |
| 08:55 | <othermaciej> | (or do you mean validator implementation?) |
| 08:58 | <hsivonen> | othermaciej: yes |
| 08:58 | <othermaciej> | the fact that it naturally falls out of how you implemented the validator doesn't necessarily mean it's clearly required by spec |
| 08:58 | <hsivonen> | I have trouble finding the right search terms to locate my earlier treatment of this topic |
| 08:58 | <othermaciej> | I'm not saying this to argue in favor of dc:foo, btw, I'm just wary of ending up with an SVG+CSS-style gap between the specs |
| 09:01 | <hsivonen> | oh well. my earlier message didn't use obvious words, I suppose |
| 09:05 | <othermaciej> | shouldn't some spec (probably HTML5) say that SVG and MathML conformance requirements are to be applied to the parsed DOM? |
| 09:09 | <othermaciej> | SVG 1.1 does actually have a definition of "Conforming SVG Document Fragments" but it's all XML-specific |
| 09:10 | <othermaciej> | one of the requirements is XML well-formedness - I am not sure it is meaningful to ask whether a DOM is well-formed XML, since well-formedness is defined in terms of text |
| 09:12 | <Hixie> | i would hope SVG defines the conformance requirements of SVG nodes in _any_ DOM, whether HTML is present or not |
| 09:12 | <Hixie> | if not, that's more of a problem for SVG than for HTML |
| 09:12 | <othermaciej> | SVG is not defined in terms of a DOM |
| 09:14 | <othermaciej> | it seems to be defined in terms of XML concepts which are ultimately defined at a textual level |
| 09:15 | <othermaciej> | I don't think SVG 1.1 was written with any non-XML syntax in mind |
| 09:15 | <Philip`> | What about createElement etc? |
| 09:15 | <othermaciej> | or MathML 3.0 (the only MathML version I looked at) |
| 09:15 | <Philip`> | (which is effectively a non-XML syntax) |
| 09:16 | <othermaciej> | I don't think SVG 1.1 defines conformance for API-created DOM trees |
| 09:17 | <othermaciej> | (or MathML) |
| 09:24 | <othermaciej> | does HTML5 itself have conformance requirements for script-constructed DOMs? I seem to recall a requirement that the DOM must not be put in a nonconforming state by script, but I can't find it offhand |
| 09:28 | <hsivonen> | Maybe I should maintain a list of public-html highlights about SVG so that I could drop URLs instead of a vague "search the archives" recommendation |
| 09:28 | <othermaciej> | hsivonen: I'm reading your longer email |
| 09:34 | <othermaciej> | hsivonen: I still think it is not properly specified anywhere that SVG 1.1 requirements should be applied to the DOM |
| 09:35 | <othermaciej> | hsivonen: I do agree that it's the sane way to do things |
| 09:55 | <hsivonen> | othermaciej: being clear about that point seems like actionable feedback to Hixie |
| 09:55 | <othermaciej> | hsivonen: I think so too, though I'm also having trouble thinking about how to define it clearly |
| 09:56 | <othermaciej> | if you look at SVG 1.1 and follow all the normative references, it is defined in terms of text |
| 09:56 | <othermaciej> | I'm not sure about specs that use RelaxNG instead of DTDs as the formalism |
| 09:56 | <hsivonen> | othermaciej: seems like a bug in SVG specs |
| 09:57 | <othermaciej> | hsivonen: updating the relevant other specs might be the way to go |
| 10:00 | <othermaciej> | however, Namespaces in XML appears to be defined in terms of text, and I don't think anyone is up for changing it |
| 10:01 | <othermaciej> | actually, it looks like the Namespace constraints could be interpreted in ways that are not purely textual, but "namespace-well-formed" is definitely defined in terms of text |
| 10:02 | <hsivonen> | XML is one of the best-written W3C specs, but the failure to roll the infoset concept into the XML spec itself and defining everything in terms of bits on the wire is an unfortunate failure |
| 10:02 | <jgraham> | hsivonen: I'm not sure how making stuff that was intended to be namespaced XML in html non-conforrming is theoretical purity |
| 10:03 | <hsivonen> | jgraham: fair enough |
| 10:03 | <jgraham> | Since the content will not work as expected it seems like it helps users |
| 10:03 | <hsivonen> | jgraham: I considered conformance in general to be more in the direction of purity than "works in Firefox" |
| 10:03 | <othermaciej> | in the case of <metadata>, I think the common case is that no one expects the content to have any effect |
| 10:04 | <hsivonen> | I have to admit I was a bit surprised to find that Hixie had made it impossible to have any children of <metadata> that would conform per SVG |
| 10:04 | <jgraham> | othermaciej: Presumably people expect that someone will parse it and do whatever it is that people do with metadata |
| 10:04 | <hsivonen> | I had had a faulty impression that <metadata> worked like SVG <title> as far as parsing goes and {http://www.w3.org/1999/xhtml}meta were allowed |
| 10:05 | <othermaciej> | it's true that if anyone parses it according to HTML5 rules, it won't do what is expected |
| 10:06 | <othermaciej> | (unless they process it in some technically incorrect way, like using regexps or DOM Level 1 APIs) |
| 10:06 | <hsivonen> | unless, of course, "do whatever it is that people do with metadata" and "ignore" are equivalent... |
| 10:16 | <othermaciej> | it seems like things would be better if XML, Namespaces in XML and Infoset were all one spec and higher level specs could be defined in terms of infoset instead of in terms of XML grammar productions |
| 10:17 | <Hixie> | i've been saying that for years |
| 10:17 | <Hixie> | good luck convincing anyone |
| 10:17 | <hsivonen> | othermaciej: I agree |
| 10:18 | <hsivonen> | Hixie: out of curiosity, why didn't you make <metadata> parse like <title>? |
| 10:18 | <Hixie> | what's <metadata>? |
| 10:18 | <hsivonen> | Hixie: an SVG tag |
| 10:18 | <Hixie> | never heard of it |
| 10:18 | <hsivonen> | Hixie: parse like <title> in SVG foreign lands |
| 10:18 | <Hixie> | (or at least i don't recall hearing of it) |
| 10:18 | <jgraham> | hsivonen: It seems that your asumption in #3 is different to your assumption in #1 |
| 10:19 | <hsivonen> | jgraham: could be |
| 10:19 | <othermaciej> | (I recall there's already a mapping from Infoset to DOM, so XHTML5 is already indirectly defined in terms of infoset) |
| 10:19 | <othermaciej> | I hadn't heard of <metadata> until today |
| 10:20 | <hsivonen> | jgraham: yes, I'm assuming that if something is conforming, it's legitimate to assume it "works" |
| 10:20 | <Hixie> | hsivonen: making <metadata> switch back to html seems to make sense; file a bug |
| 10:20 | Hixie | really doesn't understand why people are so obsessed with metadata |
| 10:21 | <Hixie> | and by "people" i really mean software architects |
| 10:21 | <Hixie> | because few other people care |
| 10:23 | <hsivonen> | whoops! filing that bug would be bad, since "meta" is on the break out list |
| 10:23 | <hsivonen> | I don't care enough to argue for changing *that* too |
| 10:24 | <hsivonen> | umm. no |
| 10:24 | <hsivonen> | now I'm confused |
| 10:24 | <hsivonen> | the break out list wouldn't apply |
| 10:27 | <hsivonen> | ok. it was a bad idea still but for a different reason |
| 10:34 | <Lachy> | hsivonen, are you aware of this bug in your HTML5 parser in Firefox. http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Ctextarea%3E%26%23xA%3B%26%23xA%3Btest%3C%2Ftextarea%3E |
| 10:35 | <Lachy> | it seems all subsequent new lines done with char refs get ignored, when only the first is supposed to be ignored. Compare the result with the old parser, or any other browser |
| 11:06 | <Hixie> | hsivonen: yt? |
| 11:06 | <Hixie> | hsivonen: what do i need to change here http://www.whatwg.org/specs/web-apps/current-work/#interactions-with-xpath-and-xslt for http://www.w3.org/Bugs/Public/show_bug.cgi?id=7059#c18 ? |
| 11:08 | <annevk3> | this HTML5 cheat sheet that's making its rounds on twitter is based on a very old version of HTML5 :/ |
| 11:09 | <annevk3> | it still has stuff like <datatemplate> |
| 11:09 | <Hixie> | i don't even remember what that is |
| 11:09 | <othermaciej> | that was for the repetition model |
| 11:09 | <Hixie> | the first one or the second one? |
| 11:09 | <othermaciej> | that you and hyatt came up with long lng ago to make XForms people happy |
| 11:09 | <annevk3> | second |
| 11:10 | <Hixie> | ah |
| 11:10 | <annevk3> | the one that nobody got |
| 11:10 | <annevk3> | apart from Philip` iirc |
| 11:11 | <annevk3> | ah fun, the source is w3schools.com |
| 11:11 | <annevk3> | didn't see that before |
| 11:11 | <othermaciej> | Hixie: you'd have to also remove a case where XPath 1.0 name expressions match |
| 11:12 | <Hixie> | "also"? |
| 11:12 | <othermaciej> | Hixie: specifically, if you have a situation just like the 6 bullet conditions for the extra match, but the element is in no namespace, the name expression must not match the node, even though XPath 1.0 says it does |
| 11:13 | <Philip`> | I think I only got it for an evening and then I forgot how it worked |
| 11:13 | <othermaciej> | Hixie: you currently have added a case where XPath 1.0 name expressions must match nodes in an HTML DOM that they wouldn't otherwise |
| 11:13 | <othermaciej> | Hixie: but you also left all the old matching intact |
| 11:14 | <Hixie> | "you" is the wrong pronoun there |
| 11:14 | <othermaciej> | the xpath people want to replace what would have matched instead of adding |
| 11:14 | <othermaciej> | well |
| 11:14 | <Hixie> | i just pasted this text straight from hsivonen, iirc |
| 11:14 | <othermaciej> | maybe hsivonen should propose some new text then |
| 11:14 | <Hixie> | i was hoping he might |
| 11:15 | <othermaciej> | I think I understand the issue well enough to sort of suggest something, but not enough to be sure it would fit into the XPath spec the right way |
| 11:15 | <hsivonen> | Lachy: thanks for finding the textarea bug |
| 11:16 | <hsivonen> | Hixie: you need to defined that additionally, an XPath Name expression doesn't match where XPath 1.0 says it would match if |
| 11:16 | <hsivonen> | * The mane expression has no namespace |
| 11:17 | <hsivonen> | and |
| 11:17 | <hsivonen> | * The expression is being tested against an element node |
| 11:17 | <hsivonen> | and |
| 11:17 | <hsivonen> | * the element's document is an HTML document |
| 11:18 | <hsivonen> | s/mane/name/ |
| 11:18 | <othermaciej> | hsivonen: it seems odd that the XPath folks want what is arguably an even greater violation of XPath 1.0 |
| 11:18 | <hsivonen> | Hixie's email about MathML and SVG to public-html confuses me |
| 11:18 | <hsivonen> | othermaciej: :-) |
| 11:19 | <Hixie> | hsivonen: So start a new paragraph that says "However, a name expression must not evaluate to matching a node when the following conditions are met:" then your bullet points? |
| 11:19 | <Hixie> | hsivonen: what if both sets of conditions match? |
| 11:20 | <othermaciej> | hsivonen: shouldn't you also have a bullet of * the element is in the null namespace |
| 11:20 | <othermaciej> | (or "no namespace" or whatever is the right way to express that) |
| 11:24 | <hsivonen> | Hixie: oops. you also need othermaciej's point |
| 11:24 | <Hixie> | ok |
| 11:24 | <hsivonen> | Hixie: add: and * the element is not in a namespace |
| 11:24 | <othermaciej> | that makes the two sets of conditions mutually exclusive |
| 11:24 | <hsivonen> | right |
| 11:26 | <Hixie> | k |
| 11:26 | <Hixie> | checked in |
| 11:31 | <Lachy> | Hixie, is it intentional that that the authoring requirements expressed in #character-references for disallowed code points don't match the implementation requirements #tokenizing-character-references? |
| 11:31 | <Hixie> | probably not |
| 11:31 | <Hixie> | how do they differ? |
| 11:32 | <Lachy> | it seems the implementation requirements require parse errors for a lot more characters than are listed in the authoring requirements |
| 11:32 | <Lachy> | e.g. U+0001 to U+0008 generate a parse error, but aren't listed in the authoring requirements |
| 11:33 | <Hixie> | unintentional; file a bug |
| 11:34 | <Lachy> | ok |
| 11:34 | <Hixie> | 9.1.3 Text is similarly wrong |
| 11:35 | <Lachy> | how is that one wrong? |
| 11:36 | <Hixie> | doesn't disallow control characters |
| 11:36 | <Lachy> | is it supposed to forbid the same character ranges? |
| 11:36 | <Lachy> | ok. I'll mention that in the bug |
| 11:37 | <Hixie> | thanks |
| 11:39 | Hixie | tries to work out what elements block </li> from seeing the LI node it should close |
| 11:43 | <Lachy> | bug filed. |
| 11:50 | <Lachy> | why don't people responding to the fakepath thread realise that whatwg is CC'd twice, and that the unnecessary duplicate messages are annoying? |
| 11:50 | <annevk3> | why are people responding to that thread in the first place? |
| 11:50 | <annevk3> | :p |
| 11:51 | <Lachy> | good point :-) |
| 11:52 | <othermaciej> | I think I had a valid reason to respond, but my mail client seemed to be smart enough to remove the duplicate Cc |
| 11:59 | <foolip> | Hixie: is there any particular reason for "Tidy up how 'abort' and 'emptied' events fire when you call load() while it's already loading."? (http://html5.org/tools/web-apps-tracker?from=3755&to=3756) |
| 12:00 | <Hixie> | foolip: the spec was contradictory before, so something had to be fixed; the way i fixed it was to make sure 'abort' doesn't fire until after the element is in a 'not loading' state |
| 12:00 | <Hixie> | foolip: (before, one kind of abort would have it be still loading, and the other kind would have it be post-loading) |
| 12:02 | <Lachy> | hsivonen, you may want to do something to improve the obviousness of the UI for validator.nu. Apparently someone I know has used it frequently, but never knew he could switch to File Upload or Text Field |
| 12:04 | <hsivonen> | Lachy: I am aware that the UI feature is non-obvious |
| 12:04 | <hsivonen> | Lachy: It's hard to improve it, because I don't understand why it's not obvious |
| 12:05 | <Lachy> | I guess people just see the <select> and read it as a label, and never stop to think what else it might have to offer |
| 12:05 | <foolip> | Hixie: hmm, ok. only nuisance is that this is more state to keep track of internally. will try it and see |
| 12:05 | <annevk3> | what the W3C has is a lot more explicit |
| 12:05 | <Lachy> | and maybe because people don't expect that changing the value of a select field will replace other parts of the UI |
| 12:06 | <Lachy> | using a tab like approach, like the w3c validator, might be better |
| 12:06 | <Hixie> | foolip: it didn't seem like much state to keep track of, unless i'm missing something |
| 12:06 | <hsivonen> | why does OS X reset my default browser to Safari every time I log out and back in? |
| 12:06 | <Hixie> | foolip: just one boolean while you reset some variables, no? |
| 12:07 | <Lachy> | hsivonen, my guess is there's a corrupted plist somewhere in ~/Library |
| 12:07 | <annevk3> | hsivonen, Apple doesn't have a monopoly so they can do evil things? :) |
| 12:08 | hsivonen | wonders which plist to zap |
| 12:08 | <Lachy> | I'd start with the com.apple.Safari plist |
| 12:08 | <foolip> | Hixie: yes, but this algorithm will "yield" while running the synchronous events, so any "local" state that's kept across a synchronous events must actually be stored globally per element. |
| 12:08 | <Lachy> | or find out which plist gets modified when you change the default browser option in Safari's prefs dialog |
| 12:09 | <hsivonen> | Lachy: this stuff lives in the legacy Internet Config back end, AFAIK |
| 12:09 | <foolip> | Hixie: there are of course lots of such state elsewhere, I'm just saying that it's a nuisance and it's nice to know if it's for a good reason |
| 12:09 | <hsivonen> | Lachy: even though Safari is the front end now |
| 12:09 | <foolip> | meh, I fail at English |
| 12:10 | <hsivonen> | com.apple.internetpref.plist looks promising |
| 12:10 | <Hixie> | foolip: ah, yeah, if your sync events aren't just nested event loops, that could be more painful |
| 12:10 | <hsivonen> | hmm. wrong plist |
| 12:11 | <hsivonen> | ha! there's com.apple.internetconfig.plist that doesn't change and com.apple.internetconfigpriv.plist that changes |
| 12:11 | <foolip> | Hixie: right, it's of course all very implementation-specific, I have no idea how other implementors deal with this stuff |
| 12:13 | <hsivonen> | still wrong plist |
| 12:13 | <hsivonen> | sigh |
| 12:22 | <othermaciej> | hsivonen: making the three options visible at once (with tabs for example) would make it more obvious, putting the select in the controls column with its own label would also probably make it more obvious |
| 12:22 | <Hixie> | bed time |
| 12:22 | <Hixie> | nn |
| 12:27 | <hsivonen> | aaargh. the definition of element in scope changed instead of </li> using a new concept |
| 12:27 | <hsivonen> | Hixie: thanks for figuring it out in any case |
| 12:28 | <Hixie> | i actually did use a new concept |
| 12:28 | <Hixie> | i just defined it in terms of the previous concept |
| 12:28 | <Hixie> | there were two algorithms that were the same except for a list |
| 12:28 | <Hixie> | there are now three the same except for a list |
| 12:28 | <Hixie> | and i factored out the common part |
| 12:28 | <Hixie> | i thought reuse was better than new things :-) |
| 12:33 | <hsivonen> | Hixie: your latest email makes it seem that my earlier reply to Shelley was indeed accurate |
| 12:35 | <Hixie> | how so? |
| 12:36 | <hsivonen> | Hixie: that the language of where <svg> and <math> are allowed applies only if you invoke an SVG spec and a MathML spec to make those elements allowed anywhere at all and HTML5 itself doesn't make them allowed |
| 12:37 | <Hixie> | they're not "non-conforming as far as HTML5 itself goes" |
| 12:37 | <hsivonen> | Hixie: what are they if they aren't allowed by HTML5 itself? |
| 12:37 | <Hixie> | they're conforming as far as HTML5 itself goes, it just doesn't go far enough to make them conforming in an absolute sense |
| 12:37 | <hsivonen> | not-allowed but not non-conforming? |
| 12:38 | <Hixie> | i don't think it makes sense to talk about their conformance without invoking the SVG spec |
| 12:38 | <Hixie> | and MathML spec |
| 12:38 | <hsivonen> | I don't understand what conforming but not in the absolute sense means |
| 12:39 | <Hixie> | HTML5 makes them conditionally conforming, the condition being that some other spec defines that they MAY be used |
| 12:39 | <hsivonen> | Hixie: right, so if the condition isn't met, they are non-conforming |
| 12:39 | <Hixie> | yes, but that's not "non-conforming as far as HTML5 itself goes", it's non-conforming as far as a validator goes. |
| 12:39 | <Hixie> | i interpret "non-conforming as far as HTML5 itself goes" to mean that HTML5 says "MUST NOT" |
| 12:40 | <hsivonen> | ah |
| 12:41 | <Hixie> | to be honest i find this discussion to be somewhat akin to trying to draw on the head of a pin with a piece of chalk |
| 12:42 | <Hixie> | which is to say, it's splitting hairs, like you said :-) |
| 12:43 | <hsivonen> | Hixie: sure. I just want to know how the extension theory works when the TAG asks about it. |
| 12:43 | <hsivonen> | also, I wanted to check that my spec reading skills aren't totally out of order |
| 12:43 | <Hixie> | your spec reading skills have a precision greater than my spec writing skills |
| 12:44 | <Hixie> | s/precision/resolution/ |
| 12:53 | <Lachy> | I was just asked again about why <nav> without an heading generates an untitled section in the outline, by a different person from who asked in here yesterday. |
| 12:53 | <Lachy> | but he also like the idea of the outline tool outputting an implied default heading like "Navigation" |
| 12:54 | <Hixie> | i like that all these people are using the outliner |
| 12:54 | <Hixie> | is it helping them catch the issue we were talking about? |
| 12:54 | <annevk3> | naturalWidth? |
| 12:55 | <Hixie> | if people use it a lot, i'd love to be able to get away without adding an element for it |
| 12:55 | <Lachy> | I don't know. This was from Bruce Lawson, who's already aware of the section abuse issue. |
| 12:57 | <Lachy> | Hixie, the spec's inline bug reporting tool should output the link to the bug as a clickable link in the page instead of using an alert |
| 12:58 | <Hixie> | copy and paste from the alert |
| 12:58 | <Hixie> | (i couldn't work out where to put the link) |
| 12:58 | Lachy | wonders if it would be appropiate to file a bug about the bug reporting tool using the bug reporting tool? |
| 12:58 | <jgraham> | The bug reporting tool shopuld allow you to have a shorter summary and a longer body |
| 12:59 | <Lachy> | the logical place would be to replace the text field with a message that said the bug was filed with a link to the bug |
| 12:59 | <jgraham> | Because all the long summaries are pretty annoying |
| 12:59 | <annevk3> | IRC should have an /ignore for people complaining about the bug reporting tool |
| 12:59 | <Lachy> | and then when the user clicks on another section of the page, the text field is put back so they can file a new bug |
| 13:00 | <Hixie> | Lachy: expect it to be WONTFIXed unless it comes with a patch :-) |
| 13:01 | <Lachy> | ok. I'll add that to my todo list, along with the redesign of the other spec stuff I'm working on |
| 13:01 | <annevk3> | actually, it's INVALID for the HTML WG bug tracker o_O |
| 13:01 | <Hixie> | heh |
| 13:01 | <annevk3> | anyway, naturalWidth is some IE/Opera extension to <img>? |
| 13:02 | <zcorpan> | i thought all browsers had naturalWidth |
| 13:03 | <Lachy> | I don't recall ever hearing about naturalWidth/Height before |
| 13:04 | <annevk3> | undefined in Opera |
| 13:05 | <annevk3> | maybe it works in Firefox/Safari and Hixie annotated the check-in wrong |
| 13:05 | <Hixie> | what did i annotate it as? |
| 13:06 | <annevk3> | as affecting Firefox/Safari |
| 13:07 | <annevk3> | i wonder if we can still rename videoHeight/Width |
| 13:08 | <zcorpan> | annevk3: renaming videoWidth was brought up ages ago |
| 13:08 | <zcorpan> | iirc |
| 13:09 | <annevk3> | it was renamed once |
| 13:09 | <zcorpan> | was it? |
| 13:09 | <annevk3> | from intrinsicWidth or so |
| 13:09 | zcorpan | can't recall that |
| 13:10 | <foolip> | is naturalWidth supposed to be the size in pixels of the image? |
| 13:11 | <foolip> | oh I think videoWidth has been renamed or redefined once, can't remember the details though |
| 13:11 | <Hixie> | annevk3: it affects firefox safari because they're the ones who implement it |
| 13:11 | <Lachy> | foolip, "The IDL attributes naturalWidth and naturalHeight must return the intrinsic width and height of the image, in CSS pixels, if the image is available, or else 0." |
| 13:11 | <Hixie> | ok i really must go to bed now |
| 13:11 | <Hixie> | nn |
| 13:11 | <foolip> | Lachy: right, like videoWidth |
| 13:12 | <annevk3> | I was clearly wrong: http://html5.org/tools/web-apps-tracker?from=684&to=685 |
| 13:12 | <annevk3> | in any case, it would make sense for them to use the same name |
| 13:14 | <zcorpan> | yes, but not essential, and videoWidth has been implemented in 4 browsers (i think) and shipped in two |
| 13:16 | <annevk3> | i still think it would be worth it |
| 13:16 | <annevk3> | but not enough to fight for |
| 13:24 | <hsivonen> | Why does vendor veto of <keygen> by MS receive a different response than vendor veto of Theora by Apple? |
| 13:25 | <zcorpan> | because people like Theora, but don't care about <keygen> |
| 13:27 | <hsivonen> | unfortunately, I have to care about <keygen>, because the HTML5 parser breaks Gecko's <keygen>, and I think Gecko can't be shipped without <keygen> |
| 13:27 | <jgraham> | I wonder what would happen if Microsoft announced that they would not support <video> |
| 13:27 | <zcorpan> | i think the simplest thing to do for gecko is to parse <keygen> into a macro of tags like the old parser does |
| 13:28 | <hsivonen> | parser macros suck |
| 13:28 | <hsivonen> | especially in the C++ land |
| 13:29 | <hsivonen> | where doing things differently may always have an effect on memory management strategies |
| 13:29 | <zcorpan> | hsivonen: do you think isindex should not be a macro? |
| 13:29 | <hsivonen> | zcorpan: well, let's put it this way: the macro has caused crashes due to memory management special cases, and it exposes the parser to string localization issues |
| 13:30 | <zcorpan> | i take that as a "yes" |
| 13:30 | <zcorpan> | i'm pretty sure isindex does not need to be a macro as far as web compat goes |
| 13:33 | <hsivonen> | is Flash supposed to have autoupdate these days? why do I have an old version of Flash? |
| 13:33 | <hsivonen> | why isn't Mac OS X software update updating Flash if it's Apple that installed that piece of Internet-exposed software on my computer in the first place |
| 13:35 | <annevk3> | hsivonen, it's downgrading it if you updated to Snow Leopard |
| 13:35 | <annevk3> | hsivonen, see gruber |
| 13:35 | hsivonen | wonders if preserving attributes on the formatting elements cloned by the AAA is essential for compat |
| 13:36 | <hsivonen> | annevk3: I haven't installed Snow Leopard yet |
| 13:36 | <hsivonen> | annevk3: waiting for end of quarter or 10.6.1 |
| 13:36 | <zcorpan> | hsivonen: yes, for <a href> at least |
| 13:36 | <hsivonen> | zcorpan: ok |
| 13:38 | <hsivonen> | I feel Apple/Adobe are in violation of my trust when it's www.mozilla.org rather than Apple or Adobe that alerts me about Flash updates |
| 13:38 | <zcorpan> | mozilla.org uses flash? |
| 13:39 | <hsivonen> | zcorpan: no, the page that loads after a Firefox update checks if you have the latest Flash |
| 13:39 | <zcorpan> | ah |
| 13:40 | <zcorpan> | does it alert if you don't have flash at all? |
| 13:40 | <hsivonen> | zcorpan: I haven't tested |
| 13:40 | <jgraham> | hsivonen: The lack of competition in the flash-player space obviously makes adobe not care much about competing on security/privacy |
| 13:40 | <jgraham> | Or indeed on end-user friendliness in general |
| 13:41 | <hsivonen> | maybe some day YouTube and Vimeo work in Firefox/Safari/Opera without plug-ins... |
| 13:41 | <hsivonen> | fortunately, I use Flashblock to reduce the attack surface |
| 13:41 | jgraham | would like non-flash non-sucky web based upload to flickr |
| 13:42 | hsivonen | just uses Python |
| 13:42 | <jgraham> | Which means being able to select multiple files at once and display a progress bar for the upload |
| 13:44 | <hsivonen> | making my Python uploader work from Finder/Nautilus without having to run exiftool separately has been on my todo list for months |
| 13:44 | <jgraham> | I typically combine uploading photos with browsing other people's photos so an uploader isn't that useful |
| 13:45 | <jgraham> | (also I never upload more than 10 photos at a time) |
| 13:45 | <hsivonen> | I have a 9-month backlog to upload... |
| 13:51 | <Philip`> | Do RDF/RDFa people not know/care that RDF requires XMLLiterals to be Exclusive Canonical XML? |
| 13:52 | <Philip`> | I've not seen anyone talking about that requirement |
| 13:52 | <Philip`> | and lots of their examples seem to violate it |
| 13:52 | <hsivonen> | Philip`: maybe you should send email |
| 13:53 | <hsivonen> | Philip`: though I believe it's a dead letter of the spec |
| 13:53 | <hsivonen> | Philip`: even though it's bad in theory not to have that requirement |
| 13:54 | <hsivonen> | Philip`: does it need to be exclusively canonical in RDFa or RDF/XML, though, or only in the data model? |
| 13:54 | <Philip`> | The problem with sending mail is they might decide that RDFa processors have to implement Exclusive Canonicalisation of XML, which would be annoying |
| 13:54 | <hsivonen> | in syntax vs. data model that is |
| 13:55 | <erlehmann> | oh, metadata. did you know, i had a wordpress plugin ready when ians mail came in "hey guys, lets talk about microdata again" |
| 13:55 | <Philip`> | hsivonen: I think the syntax can do whatever it wants - http://www.w3.org/TR/rdf-concepts/#dfn-rdf-XMLLiteral just restricts the lexical space of the abstract RDF datatype |
| 13:55 | <Philip`> | though I assume the syntaxes simply expect the strings in the syntax to be mapped directly onto that lexical space |
| 13:56 | <hsivonen> | Philip`: I think it would be architecturally unsound to require RDFa or RDF/XML syntax to be directly canonicalized |
| 13:56 | <hsivonen> | Philip`: as opposed to requiring the output of the processor to be |
| 13:57 | <Philip`> | hsivonen: True - I was thinking more of the N3/Turtle syntax |
| 13:58 | <Philip`> | (I'd expect RDF(a|/XML) to do serialisation into the appropriate form) |
| 13:59 | <hsivonen> | Philip`: oh do they have non-exclusively non-canonicalized N3/Turtle examples? |
| 13:59 | <Philip`> | (but I wouldn't expect N3 like "<> <blah> "<xml/>"^^http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral"; to be anything other than a direct mapping onto the lexical space) |
| 13:59 | <hsivonen> | Philip`: I agree that it's wrong to have non-canonical stuff in N3 |
| 14:00 | <Philip`> | hsivonen: http://html5.digitalbazaar.com/specs/rdfa.html does, in section 4.2 |
| 14:00 | <Philip`> | http://www.w3.org/TR/rdfa-syntax/ section 6.3.1.3 doesn't even bother to put the right namespace declarations in the output |
| 14:02 | <hsivonen> | Philip`: awesome |
| 14:02 | <Philip`> | Test cases like http://www.w3.org/2006/07/SWD/RDFa/testsuite/xhtml1-testcases/0011.sparql explicitly allow non-canonical XML in triples |
| 14:02 | <Philip`> | (though at least it requires the namespace declarations) |
| 14:06 | <Philip`> | Hixie: Are you watching the public-canvas-api list? (Erik Dahlstrom sent some feedback/questions there, but I don't know if anyone reads the list) |
| 14:06 | Philip` | supposes he should send mail about the XMLLiteral thing, even though it seems exceedingly uninteresting |
| 14:07 | gsnedders | sends plenty of exceedingly uninteresting feedback |
| 14:09 | <Philip`> | (Hmm, it seems C14n is somewhat incompatible with RDFa's use of CURIEs, because it only declares namespaces which are "visibly utilized", and CURIEs don't visibly utilize namespaces) |
| 14:09 | <gsnedders> | gah. NXEC wifi is being really slow |
| 14:09 | gsnedders | tries Opera Turbo, and finds it does make it quicker |
| 14:11 | <hsivonen> | Philip`: does exclusive c14n break qnames-in-content? |
| 14:13 | <Philip`> | hsivonen: http://www.w3.org/TR/2002/REC-xml-exc-c14n-20020718/#sec-Limitations |
| 14:13 | <Philip`> | i.e. yes, unless you do things to work around the breakage |
| 14:15 | <hsivonen> | Philip`: awesome |
| 14:15 | <hsivonen> | Philip`: you should definitely send email |
| 14:16 | <annevk3> | there's also http://www.w3.org/TR/C14N-issues/ |
| 14:17 | <gsnedders> | hsivonen: re: your tweet, I guess they assume Flash's auto-updating works itself |
| 14:17 | <annevk3> | anyway, isn't c14n just for the XML digital signatures and crap? |
| 14:17 | <hsivonen> | gsnedders: Flash has autoupdate? |
| 14:17 | <hsivonen> | gsnedders: I've never seen Flash autoupdate itself |
| 14:17 | <hsivonen> | annevk3: yes, but it's also required for theoretical purity of N3 |
| 14:17 | <gsnedders> | hsivonen: It's meant to. Apparently it's rather buggy. |
| 14:18 | <hsivonen> | annevk3: if you want to use string comparison on serialization to compare infoset equivalence, you need to canonicalize |
| 14:19 | <hsivonen> | annevk3: so if you want equivalent infosets to be equivalent when used as RDF literal values, you need the literals to be canonicalized in the model |
| 14:20 | <annevk3> | I see |
| 14:20 | <annevk3> | sounds like a lot of fun |
| 14:32 | gsnedders | is trying to write personal statement. oh the fun. |
| 14:32 | <Philip`> | Can't you copy-and-paste your old one? |
| 14:32 | <Philip`> | Or are you not the same person you once were? |
| 14:33 | <gsnedders> | I've changed. I'm a new man. |
| 14:33 | <Micheil> | hmm.. |
| 14:33 | <Micheil> | gsnedders: personal statement.. as in something you would stick in the about page of a website? |
| 14:34 | <gsnedders> | Micheil: http://www.ucas.com/students/startapplication/apply10/personalstatement/ |
| 14:35 | <Micheil> | uh, righteo... so a uni personal statement or something.. |
| 14:35 | <gsnedders> | Philip`: The opening sentence of my old one is a lie now, "I am a sixth year pupil at Madras College, currently studying Advanced Highers." |
| 14:35 | <Philip`> | gsnedders: I thought personal statements were meant to be lies |
| 14:36 | Micheil | thinks of sixth years and immediately thinks of the australian sixth year.. aka, when he was 12 O.o |
| 14:36 | <Philip`> | though preferably sufficiently subtle ones so you won't get caught out in short interviews |
| 14:37 | <Philip`> | "Ever since I was three years old I have had a yearning and a passion to learn about $SUBJECT" etc |
| 14:37 | <jgraham> | "I think I would be good at $SUBJECT" |
| 14:38 | gsnedders | was asked how he got involved in standards on phone interview for Opera |
| 14:38 | <gsnedders> | jgraham: Also, I'll spare you having to proof read it this time :P |
| 14:39 | <Philip`> | "I got involved in standards because Opera was rubbish and broken and didn't work on any sites and I wanted to try and make it marginally less useless"? |
| 14:40 | <Micheil> | gee.. that's nice to opera ;P |
| 14:40 | <gsnedders> | No: "Because of annevk2." |
| 14:40 | <Philip`> | Micheil: Well, the same applies to all other browsers too :-p |
| 14:40 | <Micheil> | yeah |
| 14:40 | <jgraham> | I assume gsnedders said "I got into standards because facebook quizzes hadn't been invented yet so I had copious amounts of free time" |
| 14:41 | <gsnedders> | jgraham: Hey, stop stalking me on Facebook using other people's accounts! (Or at least one other person!) |
| 14:43 | <Micheil> | wow.. it seems amazingly simple to convert a tidy xhtml1.1 document over to html5 style tags |
| 14:44 | <Micheil> | jgraham & gsnedders: are you guys both on the html5 / whatwg committee? |
| 14:45 | jgraham | denies being on any committees |
| 14:45 | <jgraham> | I am on some mailing lists |
| 14:45 | gsnedders | denies being on any committee… apart from the Habari Project Manangement Committee |
| 14:45 | <erlehmann> | Micheil, do it with regexes :D |
| 14:45 | <jgraham> | And I am technically a "member" of the HTML-WG |
| 14:45 | <gsnedders> | (But equally I am on some mailing lists) |
| 14:45 | <jgraham> | But that just means that I'm on a mailing list |
| 14:45 | <zcorpan> | there's a committee? |
| 14:46 | <erlehmann> | gsnedders, wait, you TOO are one of the habari folks ? they are everywhere now. |
| 14:46 | <gsnedders> | zcorpan: sshh! don't mention the committee! |
| 14:46 | <Micheil> | oh, committee, member, same thing really |
| 14:46 | <jgraham> | Well there is a WHATWG steering group thing |
| 14:46 | <gsnedders> | erlehmann: I have been since the days of 0.2 :) |
| 14:46 | <gsnedders> | erlehmann: I'm getting less and less happy with the direction of the project though |
| 14:46 | <Micheil> | (in other words, smart folks with ideas on making the web better) |
| 14:47 | beowulf | imagines a steering wheel so big it needs a group of people to turn it |
| 14:47 | jgraham | should set up the committee for the eradication of popular PHP-based blogging software |
| 14:47 | <erlehmann> | gsnedders, i should be sillent, as i abandoned s9y for WP |
| 14:47 | <Philip`> | s/popular //; s/-based blogging software// |
| 14:47 | <gsnedders> | jgraham: s/blogging// |
| 14:47 | <jgraham> | or CEPPBS for short. Our first meeting will be to discuss getting a better acnonym |
| 14:48 | <erlehmann> | jgraham, no. lets call it "The PHP Super Friends" |
| 14:48 | <gsnedders> | (or what Philip` said) |
| 14:48 | <Micheil> | gsnedders: hmm.. I don't suppose you've heard of a guy.. Matthias Bauer, have you? |
| 14:48 | <erlehmann> | moeffju ? |
| 14:48 | <Micheil> | yeah |
| 14:48 | <erlehmann> | with the green hair ? |
| 14:48 | <Micheil> | yeah |
| 14:48 | <erlehmann> | as i said, they are everywhere |
| 14:48 | <gsnedders> | erlehmann: Half my problems with Habari are things that can't be implemented easily in PHP. |
| 14:48 | <Philip`> | beowulf: Maybe it's like a ship, where only one person turns the wheel and everyone else tries to stop the mast falling over |
| 14:48 | <Micheil> | erlehmann: well, that proves my little connection in 7 degrees of separation.. |
| 14:48 | <gsnedders> | erlehmann: http://gsnedders.com/php-grievances |
| 14:49 | <gsnedders> | Micheil: Yeah |
| 14:49 | <gsnedders> | (Sorry, I have huge lag) |
| 14:49 | <erlehmann> | Micheil, no, it only prooves that communities are incestuous. |
| 14:49 | <Micheil> | gsnedders: I used to work with him on another site.. a while back, he knows his stuff |
| 14:49 | <erlehmann> | okcupid suggests nerds to me, with 96% match. but i know them already :D |
| 14:50 | <Micheil> | erlehmann: haha, that's kinda badass. |
| 14:50 | <erlehmann> | Micheil, are you one of the web-2-point-oh-germans too ? |
| 14:50 | <Micheil> | erlehmann: ah, no. I'm an aussie |
| 14:50 | <gsnedders> | Micheil: So do quite a few of us. Quite a lot of us who know stuff are wanting to write our own blogging software in something that isn't PHP :P |
| 14:51 | <Micheil> | erlehmann: and I've been around since before web 2.0, just not someone who is known by many people |
| 14:51 | <erlehmann> | gsnedders, reimplement bashblogger in zsh ! |
| 14:51 | <Philip`> | gsnedders: You should write it in C++ |
| 14:51 | <gsnedders> | erlehmann: No. bash > zsh (mostly) |
| 14:51 | <Micheil> | gsnedders: heh, I've been attempting to write something really simple in rails or django recently, just to power my own site |
| 14:51 | <erlehmann> | i want a python blogging system. |
| 14:51 | <erlehmann> | gsnedders, wat |
| 14:52 | Micheil | dislikes wordpress with a bit of a passion... |
| 14:52 | <erlehmann> | Micheil, use pylons |
| 14:52 | <Micheil> | erlehmann: I'll look it up |
| 14:52 | gsnedders | thinks WP was fairly good until the days after 2.0, and once wp.com launched |
| 14:52 | <erlehmann> | 2k lines of COBOL code with plugins in EXPECT. |
| 14:52 | <erlehmann> | that would be blogging software |
| 14:53 | <erlehmann> | gsnedders, still WP is attractive through cramming all new standards into itself. |
| 14:53 | jgraham | notes pytlons is not blogging software it is just a collection of components that could be used to make blogging software amongst other things |
| 14:53 | <jgraham> | *pylons |
| 14:53 | <Micheil> | gsnedders: well, in truth, I started part of my career through using wp.com and a small blog there, and then I kept having bugs with wordpress, which seemed to dislike me writing posts — so, I'll roll my own, then I can only blame myself if it doesn't work. |
| 14:54 | <gsnedders> | erlehmann: WP? standards? |
| 14:54 | <erlehmann> | ;_; |
| 14:54 | <gsnedders> | erlehmann: http://lists.automattic.com/pipermail/wp-hackers/2007-May/012987.html |
| 14:54 | <erlehmann> | forget what i said. i had to write a plugin just to get correct XHTML |
| 14:55 | <erlehmann> | and no it was not a theme issue |
| 14:56 | <gsnedders> | erlehmann: There are quite a few old bugs in XML stuff, including feeds which are XML. |
| 14:59 | annevk3 | is on some WHATWG comittee |
| 14:59 | <gsnedders> | annevk3: You and your cabal. |
| 15:00 | <annevk3> | there has been no communication since I joined and I'm told that before that all the communication was about me joining... |
| 15:01 | <Micheil> | ...wow.. the dojo mailing lists are busy tonight... |
| 15:10 | <Lachy> | I assume the existence of the whatwg steering committee is just to give the appearance of some kind of organisational structure, while not actually performing any useful functions |
| 15:12 | gsnedders | just wonders why at Hixie's latest blog post |
| 15:12 | <annevk3> | we have a useful function |
| 15:13 | <annevk3> | the need just isn't there to perform it |
| 15:13 | <annevk3> | (fortunately) |
| 15:19 | <Micheil> | annevk3: you don't like performing your function? |
| 15:20 | <Lachy> | Micheil, the function is to replace the editor if and when he can no longer funcation as an editor |
| 15:20 | <Micheil> | that doesn't make a whole lot of sense to me, but anyway |
| 15:20 | <Lachy> | basically, when Hixie can't do his job properly, the committee can take action to replace him |
| 15:20 | Micheil | always thought you lot helped write the specs |
| 15:21 | <Micheil> | ah, fair enough |
| 15:21 | <Lachy> | some of use edit our own specs |
| 15:21 | <gsnedders> | And formally whether to take on more work/editors. |
| 15:21 | <jgraham> | Some of us just make semi-abusive comments on irc |
| 15:21 | <Lachy> | but for HTML5, we just give feedback to Hixie |
| 15:21 | <Micheil> | jgraham: feel free to call me niave on this topic ;P |
| 15:22 | <gsnedders> | Micheil: It's spelt naïve dammit! |
| 15:22 | <gsnedders> | You naïve boy! |
| 15:22 | <Micheil> | ** jgraham: feel free to call me naïve on this topic ;P |
| 15:22 | <foolip> | who wants to bet who is more of a boy of Micheil and gsnedders? |
| 15:22 | <erlehmann> | Lachy, then Hixie shows it to his manatees, who push the balls into the tubes. The WHATWG spec generator reads the colors of the balls and proceeds with the commits. |
| 15:22 | <gsnedders> | This is a different topic, this is English :P |
| 15:23 | <Lachy> | erlehmann, LOL :-D |
| 15:23 | <Micheil> | foolip: actually, gsnedders would probably easily be my senior to ten years |
| 15:23 | <foolip> | Micheil: seriously, are you in kindergarten? |
| 15:23 | <Lachy> | Micheil, how old are you? |
| 15:23 | <gsnedders> | Micheil: Are you 7? |
| 15:23 | <Micheil> | no |
| 15:23 | <foolip> | lol |
| 15:24 | Micheil | was making a guess ;P |
| 15:24 | <erlehmann> | Micheil, talk to the manatees ;) |
| 15:24 | <jgraham> | erlehmann: Matanees? I always assumed he used dolphins |
| 15:24 | <Micheil> | gsnedders: although, by that comment, can I assume you're 17? |
| 15:24 | <jgraham> | s/Matanees/manatees/ |
| 15:25 | <foolip> | gsnedders has very little time left to use his age to dramatic effect in these situations |
| 15:25 | <Lachy> | Micheil, technically yes, we treat him like he's 12 :-P |
| 15:25 | <Micheil> | Lachy: ah, okay then, I stand corrected, we'd be the same age then ;P |
| 15:25 | <gsnedders> | Micheil: But how old are you? |
| 15:25 | <Micheil> | 17 (roughly) |
| 15:25 | <erlehmann> | jgraham, impossible ! everyone knows dolphins like namespaces ! |
| 15:25 | <Philip`> | What's the point in a committee to replace the editor, when someone could simply start up a new group with a new editor whenever they felt like it? |
| 15:26 | <erlehmann> | Philip`, legitimacy |
| 15:26 | <jgraham> | erlehmann: That's porpoises |
| 15:27 | <Micheil> | oh well, I'm off, it's getting rather late in aussieland. |
| 15:27 | <gsnedders> | Micheil: I am 17 |
| 15:27 | <Lachy> | Micheil, where in Aus are you from? |
| 15:27 | <Micheil> | umm.. NSW, regional |
| 15:27 | <gsnedders> | Micheil: and 4 months, and a random number of days |
| 15:27 | <Lachy> | Micheil, I'm from Parkes, originally. Anywhere nearby? |
| 15:28 | <Micheil> | actually, will many folk from here be at Web Directions South this year in october? |
| 15:28 | <jgraham> | gsnedders: Yopu might be younger tomorrow than today? |
| 15:28 | <Micheil> | umm.. a fair distance away, I'm closer to Griffith |
| 15:28 | <gsnedders> | jgraham: Maybe. |
| 15:28 | <Micheil> | Lachy: ^^ |
| 15:29 | <Micheil> | Lachy: although, if it's of any help I go to uni in wagga wagga.. ;P |
| 15:29 | <gsnedders> | uni? pff. |
| 15:29 | <Micheil> | yeah, uni. university. |
| 15:29 | gsnedders | just works for Opera :P |
| 15:30 | <Micheil> | supposedly we need a bit of paper to get by these days, and if they're offering me a position in the gifted and talented program, I'm not going to knock it back ;P |
| 15:30 | <Lachy> | Micheil, I'd like to go to WDS, but it depends if Opera sends me, which is pending some internal stuff |
| 15:30 | <Micheil> | Lachy: ah, I'll see you there if that does turn out — I'll also be holding an event for web week |
| 15:33 | <Micheil> | Lachy: are you actually still local to australia, or are you offshore now? |
| 15:34 | <erlehmann> | gsnedders, how come that at such a joung age? |
| 15:34 | <erlehmann> | s/joung/young/ |
| 15:35 | <Micheil> | erlehmann: age doesn't matter, talent and skill do. |
| 15:35 | <gsnedders> | erlehmann: Because wilhelm bullied me into applying. |
| 15:36 | <erlehmann> | Micheil, of course. but i, for example, had school till the age of 19, unable to attend anything more than semi-regularly at all. |
| 15:36 | <gsnedders> | erlehmann: I got moved up a year, so left at 17 |
| 15:36 | <Micheil> | erlehmann: I still have to attend school until 18 |
| 15:37 | <gsnedders> | Though I could've left with qualifications enough for Scottish unis at 16 |
| 15:37 | <jgraham> | ~sings~ you did much too much much too young. Now you're married with a kid when you should be having fun |
| 15:37 | <gsnedders> | Oh god. Not jgraham singing. |
| 15:37 | <Micheil> | erlehmann: this university thing prestarts me on my degree.. so.. it's start like starting it before the rest of the pack, while still doing the same stuff as the rest of the pack.. if that makes sense |
| 15:38 | gsnedders | covers his ears with headphones so he can't hear it |
| 15:38 | <jgraham> | Although that last bit strictly doesn't apply |
| 15:38 | <gsnedders> | How do you know? :P |
| 15:38 | <Micheil> | gsnedders is a teen dad |
| 15:39 | gsnedders | takes headphones off, finding jgraham has stopped singing |
| 15:40 | <gsnedders> | jgraham: Really, you should leave the singing to your girlfriend, then I wouldn't need to take such drastic action. |
| 15:40 | <erlehmann> | gsnedders, i got moved up a year to, but was late in school anyways. my lil brother will be graduating at 16 too. |
| 15:41 | Micheil | is in the standard year for his age |
| 15:42 | <gsnedders> | Micheil: n00b |
| 15:42 | Micheil | would like to add that he's supposedly the only person in this gifted and talented course at uni ;P |
| 15:42 | <Micheil> | gsnedders: ;P |
| 15:42 | <erlehmann> | TCP retransmit gets me everytime. i just used my mobile connection for phoning and i'm still logged in, yay \o/ |
| 15:42 | <erlehmann> | Micheil, i know that stuff, but it takes time. someone very bright i knew stoppen |
| 15:42 | <erlehmann> | d doing it to have his free time |
| 15:43 | <erlehmann> | anyways, after seeing how comparably easy it is to pass higher-up courses as a first-year student, i began to hate formal requirements that get in the way of learning. |
| 15:43 | <Micheil> | heh, I have no idea how I'm going yet in this course.. I get my first assessment back in a few weeks time |
| 15:44 | <Micheil> | anyway.. I was going to head off.. so I'll talk to you folks another day |
| 15:45 | <erlehmann> | Micheil, must be a lonely course then ;D |
| 15:45 | gsnedders | wonders why he's taken to listening to Death Cab for Cutie while travelling |
| 15:47 | <erlehmann> | gsnedders, listen to the streets |
| 15:47 | <Micheil> | erlehmann: actually, it's a standard university course, and I do the exact same assessments and work as those who are actually sort of "officially" in university, or of age to be |
| 15:47 | <gsnedders> | erlehmann: I don't own that :P |
| 15:48 | <erlehmann> | Micheil, that was kind of a joke. meh |
| 15:48 | <gsnedders> | erlehmann: Also, it's totally not my sort :P |
| 15:48 | <Micheil> | erlehmann: uh, I didn't get it ;P |
| 15:58 | <gsnedders> | "If you don't love me now/You'll never love me again/I can still hear you saying you'll never break the chain" |
| 16:00 | <erlehmann> | love is overrated |
| 16:00 | Micheil | prefers listening to Something for Kate over gsnedders or jgraham singing |
| 16:00 | <Micheil> | :D |
| 16:01 | <erlehmann> | i prefer Prince π :D |
| 16:02 | <gsnedders> | Micheil: It's better than me and jgraham in stereo. But you wouldn't be the first to say myself and jgraham are like jgraham in stereo. |
| 16:02 | <Micheil> | uhh.. I would rather not go there :D |
| 16:04 | <gsnedders> | (on my first day at Opera it was said that myself and jgraham look like brothers, scarily) |
| 16:04 | <Micheil> | erm.. woah |
| 16:04 | <Micheil> | haha |
| 16:05 | gsnedders | goes back to Evanescence |
| 16:07 | Micheil | goes to bed.. urgh.. 1am. |
| 16:07 | <gsnedders> | 1am isn't that late. |
| 16:07 | <gsnedders> | But my sleeping hours are really screwed up. |
| 16:08 | <Micheil> | gsnedders: 1am is late when you need to get up at 7am to go to school |
| 16:08 | <Micheil> | and it's also late when you need to travel 600km's each week for university |
| 16:08 | <gsnedders> | Micheil: Before I had CFS, I normally went to bed at 2am in such circumstances |
| 16:08 | <Micheil> | gsnedders: CFS? |
| 16:08 | <gsnedders> | Micheil: Chronic Fatigue Syndrome |
| 16:09 | <erlehmann> | being sleepy ? |
| 16:09 | <Micheil> | hmm.. so.. what's that do? |
| 16:09 | <gsnedders> | It makes you chronically fatigued. |
| 16:09 | <Micheil> | / any idea on the courses? |
| 16:11 | <Micheil> | hmm.. I'm off. |
| 16:43 | <erlehmann> | annevk2, a big thanks for tellin me about the File API. you ended the discussion in the best way imaginable :) |
| 16:58 | <gsnedders> | Someone find marcos and drag him onto IRC |
| 17:09 | gsnedders | mindlessly vandalises his blog some more |
| 17:21 | <Philip`> | annevk3: By the way, did you see http://philip.html5.org/data/charsets-2.html ? |
| 17:22 | Philip` | is just asking since he doesn't remember seeing any response |
| 17:23 | <annevk3> | oh sweet |
| 17:23 | <annevk3> | no, missed that |
| 17:23 | <annevk3> | is this from the botbot or whatever source? |
| 17:26 | <TabAtkins> | I like the dichotomy in that table - nearly all of them are either used nearly completely correctly, or nearly completely incorrectly. |
| 17:27 | <jgraham> | annevk3: The source is described at the top of the document |
| 17:27 | <jgraham> | (well in a file linked from the top of the document) |
| 17:27 | <jgraham> | (and yes, it is dotbot) |
| 17:28 | <takkaria> | TabAtkins: dichotomies, they're a love/hate thing |
| 17:28 | <takkaria> | I've been waiting to say that to someone for months, sorry |
| 17:29 | <TabAtkins> | You're a bad person, takkaria. |
| 17:29 | <jgraham> | takkaria: BTW should I be worried that I haven't seen you for weeks even though you are supposedly located a few tens of metres from me? |
| 17:29 | <jgraham> | (admittedly quite a few of those metres are vertical ones) |
| 17:29 | <beowulf> | underground bunker? |
| 17:30 | <takkaria> | jgraham: not really. I tend to eat lunch at my desk and haven't been around friday evenings because I've been travelling around after work |
| 17:30 | <jgraham> | beowulf: Nah, takkaria lives somewhere in the sky |
| 17:30 | <jgraham> | takkaria: Been anywhere nice? |
| 17:31 | <takkaria> | jgraham: Norrkoping, Stockholm, Goteborg. I think Stockholm was the nicest. :) |
| 17:31 | <jgraham> | Hmm, I have only been to one of those three places |
| 17:31 | <jgraham> | (but it was the nice one!) |
| 17:32 | <takkaria> | I figured that I'm leaving soon so I should make the most of being in the country |
| 17:32 | <jgraham> | When do you go? |
| 17:32 | <takkaria> | 12 days |
| 17:32 | <TabAtkins> | Hrm, my trig appears to not be working correctly. |
| 17:32 | <takkaria> | jgraham: are you at the kickoff tomorrow, btw? |
| 17:33 | <jgraham> | takkaria: Yes, although I haven't worked out why it has such a silly name yet |
| 17:34 | <takkaria> | jgraham: I'm still a bit confused about what it consists of, but I'm quite looking forward to it |
| 17:34 | <jgraham> | takkaria: It seems to consist of lots of oppertunities to kill yourself and/or other people fololowed by food |
| 17:35 | <takkaria> | hm, death photography |
| 17:36 | jgraham | has a theory he might enjoy taking pictures of other people trying to kill themselves more than trying to kill himself |
| 17:36 | <gsnedders> | jgraham: How was the core one, BTW? |
| 17:36 | <jgraham> | gsnedders: Surprisingly OK |
| 17:37 | <jgraham> | gsnedders: There are photos under http://hoppipolla.co.uk/410/teambuilding/ |
| 17:37 | <gsnedders> | Was my guess about who'd win (Lasse) right? |
| 17:37 | <jgraham> | gsnedders: He is in Nevada or something |
| 17:37 | <gsnedders> | Oh, yeah. |
| 17:37 | <jgraham> | Navada |
| 17:38 | <gsnedders> | That made me change my guess. I can't remember what I changed it to. |
| 17:38 | <gsnedders> | He's at Burning Man. |
| 17:38 | gsnedders | realizes looking at this on an NXEC wifi connection which is struggling to make 1KB/s is a bad idea |
| 17:39 | <TabAtkins> | Philip`, does a U in the second column mean "unknown"? |
| 17:39 | <gsnedders> | TabAtkins: Yes |
| 17:40 | <jgraham> | gsnedders: Yes, the photos are a little large |
| 17:52 | <Philip`> | TabAtkins: I don't think it's true they're all used nearly completely (in)correctly - I think my code that calculates correctness is bogus, because the output doesn't make sense |
| 17:53 | <TabAtkins> | Ah, k |
| 17:53 | <Philip`> | (There's a disclaimer at the top of the file about that, so it's not my fault!) |
| 17:53 | <TabAtkins> | Pfft. |
| 17:53 | <Philip`> | (Bugs aren't bugs if they're documented) |
| 17:53 | <TabAtkins> | They're features? |
| 17:54 | <gsnedders> | No, they're documented _and_ dressed up. |
| 17:57 | <Philip`> | TabAtkins: The "U"s have tooltips |
| 17:57 | <TabAtkins> | Ah, you need a "[title]{cursor:help;}" rule or whatever. |
| 17:58 | <TabAtkins> | I had no clue there was a tooltip there. |
| 17:59 | <TabAtkins> | For godsakes, I just discovered the bug in my gradient generator that was causing it to ignore the <angle> parameter was that I was using "pi" rather than "pi()". |
| 17:59 | <TabAtkins> | Who doesn't add pi into their language as a constant, I ask you? |
| 18:00 | TabAtkins | switched to using deg2rad() anyway. |
| 18:02 | <gsnedders> | TabAtkins: What language? PHP? |
| 18:02 | <TabAtkins> | gsnedders: yeah. |
| 18:02 | <gsnedders> | TabAtkins: PI |
| 18:02 | <TabAtkins> | too late. >_< |
| 18:02 | <TabAtkins> | Should have been using the conversion function anyway. |
| 18:02 | <gsnedders> | No, that's wrong. But there is a constant. |
| 18:02 | <TabAtkins> | M_PI |
| 18:03 | <TabAtkins> | For who-knows-what reason. |
| 18:03 | <gsnedders> | Lack of namespacing :P |
| 18:03 | <TabAtkins> | Yeah, they only introduced static variables and functions in php5, I think. |
| 18:04 | <gsnedders> | Static functions? They still don't exist! |
| 18:04 | <gsnedders> | Static methods exist since PHP5 |
| 18:04 | <TabAtkins> | Hrm? Sure they do. |
| 18:04 | <gsnedders> | But not static functions |
| 18:05 | <TabAtkins> | Oh, buh, whatever. The distinction is meaningless. |
| 18:05 | <TabAtkins> | They're just functions added to a class object rather than an object instantiated from a class. |
| 18:11 | <Philip`> | TabAtkins: The existence of the tooltip is indicated by the tooltip that pops up when you put the mouse over it |
| 18:11 | <Philip`> | so there's no need to also indicate it by a cursor change |
| 18:12 | <TabAtkins> | Philip` only if you *hold* the mouse cursor over it for a second or two. In the meantime your cursor is displaying an ordinary text cursor, and it's a pretty tiny mouse target to begin with. |
| 18:12 | <TabAtkins> | So you *do* need a cursor change to give people immediate feedback that there's something more if they'll just wait a second. |
| 18:12 | <Philip`> | TabAtkins: M_PI sounds like they just copied from the C standard library, which uses the M_ prefix for maths constants because otherwise it conflict with the giant mess that is the global namespace in a typical C program |
| 18:12 | <TabAtkins> | Philip`, makes sense. PHP is built on C anyway. |
| 18:12 | <Philip`> | TabAtkins: You need a browser/OS with faster tooltips :-p |
| 18:13 | <Philip`> | or that shows it in the status bar immediately |
| 18:13 | <Philip`> | TabAtkins: That doesn't make sense; pretty much everything is built on C, and it doesn't all copy C's rubbish naming conventions |
| 18:13 | <TabAtkins> | It's possible that FF is showing it in the status bar. I hide that anyway. |
| 18:14 | <TabAtkins> | Philip`: Yeah, but PHP was designed by idiots with no sense of consistency. |
| 18:14 | <gsnedders> | TabAtkins: That's untrue. |
| 18:14 | <Philip`> | TabAtkins: That seems a better explanation |
| 18:14 | <gsnedders> | (PHP wasn't designed.) |
| 18:14 | <TabAtkins> | gsnedders, point. |
| 18:15 | <gsnedders> | It's a series of features hacked on top of a language that wasn't designed to have them. |
| 18:15 | <gsnedders> | And those features coming from people writing patches with whatever API they want. |
| 18:18 | <TabAtkins> | All correct. |
| 18:18 | <TabAtkins> | And yet it's still one of the easiest languages to use. Le sigh. |
| 18:18 | <gsnedders> | It's one of the easiest languages to get something to work in. It's one of the hardest languages to write nice, clean, well-designed, maintainable code in. |
| 18:18 | <Philip`> | All languages evolve with features that they weren't originally designed for |
| 18:19 | <Philip`> | but usually they seem to have a more consistent design vision |
| 18:19 | <gsnedders> | Most other languages ultimately have some single gatekeeper |
| 18:19 | <TabAtkins> | gsnedders: I disagree. I write very good code in PHP. It's not hard. It's just *very easy* to write bad code in PHP. |
| 18:20 | <gsnedders> | TabAtkins: It depends what you want to do. Some things are very hard to write cleanly in PHP. |
| 18:20 | <TabAtkins> | Hmm, can you elaborate? I find that it supports all the basic features you want out of a language to implement design patterns well. |
| 18:21 | gsnedders | is getting into Edinburgh now, and this means going in a long tunnel, and wifi frequently sucks north of Edinburgh on these trains |
| 18:21 | <TabAtkins> | Lacking perhaps only namespaces, which are in php6 when it lands. |
| 18:21 | <gsnedders> | and 5.3, which is already out |
| 18:21 | <Philip`> | gsnedders: Why didn't you just bring a long network cable with you? |
| 18:21 | <TabAtkins> | Ah, didn't realize. |
| 18:21 | <gsnedders> | with "\" as a namespace separator |
| 18:22 | <gsnedders> | (w00t! Foo\Bar::magic(); — that looks lovely) |
| 18:22 | <Philip`> | Can you refer to namespaced things in a string? |
| 18:22 | <TabAtkins> | Heh, I actually sort of like the pseudo-directory structure it brings. ^_^ |
| 18:22 | <TabAtkins> | Philip`, just escape the \. |
| 18:23 | <gsnedders> | Philip`: I guess ${foo\bar} works. Dunno. |
| 18:23 | <gsnedders> | TabAtkins: directories are /, dammit! |
| 18:24 | <TabAtkins> | Thus it's a pseudo-directory, just like Windows has. |
| 18:24 | <Philip`> | TabAtkins: I mean something equivalent to Perl's "foo $bar::baz qux", because I'm wondering it's like "foo $bar\baz qux" since that would get horribly confusing when you print a newline after a variable |
| 18:24 | <Philip`> | s/it's/if it's/ |
| 18:25 | <TabAtkins> | Philip`, I dunno details, but I suspect it's as gsnedders says. |
| 18:25 | gsnedders | hasn't really looking into the new stuff in 5.3 much |
| 18:25 | <TabAtkins> | Or something close. |
| 18:25 | <gsnedders> | *looked |
| 18:25 | <takkaria> | in any case, it looks like an excellent addition to an otherwise umblemished language |
| 18:26 | <TabAtkins> | I just wish I had a box to run Hunchentoot on. And the patience to go learn a server other than Apache. |
| 18:26 | <gsnedders> | I tend to for all the distributable code I write promise to support the version of PHP that ships with the oldest support Debian release. (Currently 5.2.0.) |
| 18:27 | <TabAtkins> | Though I suppose I could just set up mod_lisp and take the CGI hit. |
| 18:28 | <Philip`> | TabAtkins: I thought the point of mod_* was to avoid the CGI hit |
| 18:28 | <Philip`> | since you can get the CGI hit by simply using CGI, and the special modules are intended to make things better |
| 18:28 | <TabAtkins> | Duh, yeah, you're right. I'd run hunchentoot behind mod_lisp. |
| 18:29 | TabAtkins | really doesn't understand servers or networking. |
| 18:29 | <Philip`> | Networking is easy, it's just a load of wires |
| 18:31 | <TabAtkins> | also: the internet is a giant web of pixie dust and magic smoke. |
| 18:32 | <gsnedders> | And tubes. Don't forget the tubes. |
| 18:32 | <TabAtkins> | Bah, I'll just use lisp as a CGI. That's easy to get running. |
| 19:45 | <Lachy> | it would be nice if we just developed a nice web based tool which could take an SVG image, strip out all the namespaced stuff that would be non-conforming in HTML and send back the image to the user ready to copy and paste into their HTML |
| 19:45 | <Lachy> | though, maybe it would have to replace the creative commons stuff with a simpler, conforming alternative that means the same thing |
| 20:09 | <TabAtkins> | Should be easy if you have an XML dom, right? |
| 20:14 | <annevk3> | should be fairly easy to make as a Python app |
| 20:14 | <annevk3> | (I say that because I know a bit of Python) |
| 20:27 | <Lachy> | yeah, and as long as the input is well formed XML, then it shouldn't be too hard. But if we want to handle some non-well-formed input, or input with namespace errors, and be able to output valid SVG for text/html, it would be harder |
| 20:29 | <Lachy> | btw, that Wikipedia logo in text/html page that Shelley made really makes Firefox slow to use |
| 20:30 | <annevk3> | you could use my XML5 code |
| 20:30 | <annevk3> | but euh, I thought the point was copy and pasting well-formed XML |
| 20:32 | <annevk3> | you could probably also just use html5lib and filter out the implied <html>, <head>, <body> crap |
| 20:32 | <Lachy> | yeah, but we know authors don't always make XML based SVG well formed |
| 20:32 | <annevk3> | we also know that content doesn't work |
| 20:33 | <annevk3> | othermaciej, hey |
| 20:34 | <Lachy> | using html5lib would put all the namespaced stuff like <dc:foo> and prefixed attributes into the wrong namespace, so filtering wouldn't be as simple as stripping out anything that isn't in the SVG, MathML or HTML namespace, nor an atribute in no namespace |
| 20:34 | <othermaciej> | hey annevk3 |
| 20:34 | <annevk3> | othermaciej, I was wondering if http://developer.apple.com/safari/library/documentation/iPhone/Conceptual/SafariJSDatabaseGuide/OfflineApplicationCache/OfflineApplicationCache.html is correct for WebKit as it seems like it contradicts HTML5 for fallback entries |
| 20:34 | <annevk3> | othermaciej, I asked in #webkit earlier but nobody was around that could help |
| 20:35 | <othermaciej> | annevk3: that documentation could be out of date, but it's also possible the implementation is out of date with the spec |
| 20:35 | <othermaciej> | annevk3: what's the specific point of contradiction? |
| 20:36 | <othermaciej> | annevk3: andersca is the person most likely to know offhand, for me to know the exact current behavior I would need to test |
| 20:37 | <annevk3> | othermaciej, it suggests that for a fallback thingy of /foo/ /bar/ a failed load of /foo/1.htm loads /bar/1.htm where as HTML5 suggests it should load /bar/ |
| 20:39 | <annevk3> | I haven't tested myself. Did find that the Firefox documentation does support my interpretation of HTML5... |
| 20:39 | <othermaciej> | annevk3: I'm not sure if it actually does that, but if so it's probably unintentional, or an error of interpretation |
| 20:40 | <annevk3> | kk, I pinged andersca just in case |
| 20:41 | virtuelv | changes topic to 'WHATWG (HTML5) -- http://www.whatwg.org/ -- Logs: http://krijnhoetmer.nl/irc-logs/ -- Vennligst legg igjen din logiske sans ved døren, takk!' |
| 20:44 | <jcranmer> | Leave your sense of logic at the door in... Norwegian? |
| 20:53 | <virtuelv> | jcranmer: sort of |
| 20:53 | <virtuelv> | the original seems to have been auto-translated initially |
| 20:53 | <Lachy> | virtuelv, yeah, google translate helped me a little bit with a few words in the original |
| 20:53 | <virtuelv> | it's now valid Norwegian, but a bit strained, and we would never say "thank you" |
| 20:54 | <Lachy> | virtuelv, that's just because norwegians are impolite :-) |
| 20:54 | <virtuelv> | Lachy: no, we're not |
| 20:54 | <virtuelv> | it's just that ", takk" in Norwegian implies that you expect compliance |
| 20:54 | <virtuelv> | so instead of being a polite request, it's now a fairly rude way of ordering people around |
| 20:55 | virtuelv | changes topic to 'WHATWG (HTML5) -- http://www.whatwg.org/ -- Logs: http://krijnhoetmer.nl/irc-logs/ -- Vær vennlig og legg igjen din logiske sans i døren' |
| 20:55 | <Lachy> | then let's change it back to the english version, which doesn't suffer from that problem |
| 20:55 | <virtuelv> | Now it, pretty directly is "Be polite and leave your sense of logic at the door" |
| 20:56 | <virtuelv> | s/directly/literally/ |
| 20:56 | <virtuelv> | but the implied translation changes "vær vennlig" to "please" |
| 20:56 | <Lachy> | Google translate says otherwise: "Please leave your logical sense at the door". But I suppose I can trust your translation a little more |
| 20:56 | <virtuelv> | please, though, directly would translate to "be kind" |
| 20:56 | <virtuelv> | or "be so kind" |
| 20:57 | <virtuelv> | (vær så snill) |
| 20:57 | <takkaria> | ah, like varsågod in Swedish |
| 20:57 | <Lachy> | I thought vennligst meant please |
| 20:57 | <virtuelv> | Lachy: depends on whether you want to order people around or be polite |
| 20:58 | <Lachy> | oh, so those signs I've seen around that say things like "Vennligst vent her" are ordering me around, instead of making a polite request? |
| 21:01 | <virtuelv> | Lachy: pretty much, yes |
| 21:03 | <virtuelv> | Lachy: the differences are subtle, though |
| 21:03 | <virtuelv> | it's not like in dutch airports |
| 21:03 | <virtuelv> | where their last call messages are saying "[Insert name], proceed to gate immediately, you are delaying the flight" |
| 21:07 | <Lachy> | I wish IRC were more resillient to this kind of server adjustments, without causing so many splits |
| 21:13 | <AryehGregor> | Lachy, every other chat protocol seems not to have this problem . . . |
| 21:14 | <takkaria> | IRC is *old* though |
| 21:14 | <Lachy> | AryehGregor, it's not really surprising. IRC is one of the oldest chat protocols still in common use. Newer protocols probably learned from its shortcomings |
| 22:13 | <Lachy> | othermaciej, I suspect the reason the W3C validator reports errors for those namespaced attributes is because its using DTD based validation, which has limitations in its ability to express conformance when namespaces and prefixes are involved. |
| 22:17 | <othermaciej> | Lachy: as far as I can tell, nothing in SVG lets you add arbitrary attributes in other namespaces to arbitrary elements |
| 22:17 | <othermaciej> | Lachy: that's only allowed in <metadata> |
| 22:18 | <Lachy> | I'm not too familiar with the SVG spec to know for sure, but I will take a look |
| 22:19 | <Lachy> | http://www.w3.org/TR/SVG/extend.html |
| 22:19 | <Lachy> | "SVG allows inclusion of elements from foreign namespaces anywhere with the SVG content." |
| 22:19 | <othermaciej> | ok, I was wrong |
| 22:19 | <othermaciej> | please correct me on the list |
| 22:19 | <Lachy> | "Additionally, SVG allows inclusion of attributes from foreign namespaces on any SVG element." |
| 22:20 | <Lachy> | ok |
