| 00:10 | <bewest> | what character do you stop at to determine the end of a broken entity? |
| 00:10 | <bewest> | or what kinds of characters? |
| 00:11 | <zcorpan> | what's a broken entity? |
| 00:13 | <Lachlan> | bewest, the algorithm consumes as many characters as possible until it either matches one of the entity references, or can't do so, in which case it just outputs the consumed characters |
| 00:13 | <Dashiva> | Mhm... |
| 02:10 | Philip` | wonders if anything implements the "flavor" colour, or has plans to do so |
| 02:14 | <mpt> | Philip`, MacIE5 is the only browser that ever did afaik |
| 02:15 | <mpt> | And it's a value that only ever made sense for Rev 1 iMacs |
| 02:16 | <mpt> | and that it ever found its way into the CSS spec is a mystery of our times |
| 02:16 | mpt | winks at tantek |
| 02:17 | <Philip`> | Ah, that makes sense |
| 02:17 | <Philip`> | (Well, not much sense, but at least slightly more than none :-) ) |
| 02:23 | mpt | wonders if it's been removed from CSS2.1 on the grounds of not having multiple interoperable impelementations, and if not, why not |
| 02:27 | <Philip`> | flavor doesn't seem to be in any of CSS 1, 2 or 2.1, though it is in 3 |
| 02:32 | <mpt> | CSS colors are a disaster zone |
| 02:32 | <mpt> | Most of the UI colors make sense only in Windows 95, 98, and 2000 |
| 02:35 | <bewest> | ? |
| 02:35 | <bewest> | red |
| 02:35 | <bewest> | blue |
| 02:35 | <bewest> | silver |
| 02:36 | <Philip`> | inactivecaptiontext |
| 02:37 | <Hixie> | mpt: yeah, 'appearance' is supposed to fix that. the system colours are deprecated. |
| 02:38 | <Philip`> | It seems odd how graytext is defined to handle the case of monochrome displays, though nothing else is and you'd presumably end up with quite a broken interface on such a display |
| 02:40 | <Hixie> | i thought graytext was for disabled controls, not monochrome displays |
| 02:41 | <Philip`> | It says "Grayed (disabled) text. This color is set to #000 if the current display driver does not support a solid gray color." |
| 02:41 | <Philip`> | (I suppose you could get multichrome displays that don't support grey, but that seems unlikely) |
| 03:48 | <mpt> | Philip`, the Apple II or the Mac Classic :-) |
| 03:48 | <mpt> | It would be rather difficult to get a graphical Web browser working on them, though |
| 04:23 | <mpt> | agggh |
| 04:24 | <mpt> | Hixie, I wish to make a complaint |
| 04:24 | <mpt> | I've had two users now saying "Opera passes Acid2, therefore the CSS problem I have with your site can't be an Opera bug" |
| 04:28 | <mpt> | or at least implying it |
| 04:50 | <jruderman> | blockquote { flavor: cherry; } |
| 06:20 | <Hixie> | mpt: sorry, i don't take responsibility for the world's morons |
| 06:25 | <mpt> | "Acid2 is a test page, written to help browser vendors ensure proper support for web standards in their products" could be "Acid2 is a test page, written to highlight *some* high-profile bugs and sought-after features in Web browsers" |
| 06:25 | <mpt> | or something |
| 06:25 | <mpt> | anyway, hometime for me |
| 06:25 | <Hixie> | i didn't write that |
| 06:25 | <Hixie> | so... |
| 06:26 | <mpt> | ah, you wrote the thing next to it |
| 06:26 | <mpt> | ok, I'll e-mail them |
| 07:04 | <met_> | http://solutoire.com/plotr/ (charts with canvas) see also discussion under http://www.linux.com/article.pl?sid=07/05/04/1457210 |
| 08:41 | <mikeday> | is it possible to implement support for HTML forms using something like XBL? |
| 08:44 | <othermaciej> | depends on whether you have the other needed primitive capabilities |
| 08:51 | <mikeday> | eg. JavaScript support? |
| 08:52 | <mikeday> | as I understand it, XBL can add bits to the DOM transparently, |
| 08:52 | <mikeday> | but that won't magically make the added bits interactive |
| 08:53 | <mikeday> | although it could presumably reproduce the appearance of HTML forms, by setting style |
| 08:55 | <mikeday> | and handle things like <select size="4">, which can't really be done with pure CSS |
| 08:56 | <othermaciej> | you would need to have an XBL binding not just on the controls but also on <form> and you would need enough networking APIs |
| 08:56 | <othermaciej> | <input type="file"> can't be done with XBL |
| 09:05 | <mikeday> | true |
| 09:05 | <mikeday> | haven't really thought that far ahead, I'm just curious if it's possible to make interactive buttons and things |
| 09:06 | <othermaciej> | yes, XBL can bind event listeners |
| 09:06 | <mikeday> | ah of course |
| 09:06 | <mikeday> | so it could turn <button>Hello</button> |
| 09:07 | <mikeday> | into a more complicated tree, eg. to add borders or other decorations |
| 09:07 | <mikeday> | and add event listeners to make it clickable, and change the style when clicked |
| 09:07 | <mikeday> | (and possibly fire off other events or whatever) |
| 09:10 | <mikeday> | presumably it would also need to update the DOM properties that HTML form elements are supposed to have |
| 09:10 | <mikeday> | so that other scripts on the page could query the state of the form controls and get the expected results |
| 09:10 | <mikeday> | eg. is this checkbox selected or not |
| 09:11 | <mikeday> | the final submit would be a bit tricky, as you said |
| 09:11 | <mikeday> | and file input wouldn't be possible without some kind of extension |
| 09:12 | <mikeday> | still, it sounds halfway plausible :) |
| 09:24 | <nickshanks> | hello mike |
| 09:24 | <nickshanks> | how is your evening? |
| 09:27 | <mikeday> | hi nickshanks |
| 09:27 | <mikeday> | good thanks, and you? :) |
| 09:28 | <nickshanks> | i am waiting for a 500 GB perpendicular serial ata hard drive to arrive |
| 09:28 | <nickshanks> | since i have 37.2 MB left on my boot drive |
| 09:28 | <mikeday> | hmm, what exactly is perpendicular about it? |
| 09:28 | <nickshanks> | the magnetic 'bits' that the read-write head toggles |
| 09:29 | <nickshanks> | http://www.google.com/search?q=perpendicular+recording |
| 09:29 | <mikeday> | I can't say I've ever much worried about the orientation of my magnetic bits :) |
| 09:29 | <nickshanks> | mikeday: what cool neat-o-matic features are you working on today? |
| 09:30 | <mikeday> | none, just the Prince 6.0 release |
| 09:30 | <mikeday> | it includes basic styling for HTML forms, which is a feature |
| 09:30 | <mikeday> | non-interactive though |
| 09:31 | <mikeday> | and not all aspects of form styling are reproduced yet, as that's hard to do with CSS only |
| 09:31 | <nickshanks> | does PDF support styled forms in that way |
| 09:32 | <mikeday> | well, we're not generating PDF forms at all yet, only things that look like form controls |
| 09:32 | <mikeday> | but I think it does, as you can just draw arbitrary stuff on the page and then make it interactive |
| 09:32 | <nickshanks> | sounds like a feature for 6.1 or 6.2 :) |
| 09:35 | <mikeday> | yeah, or 7.0 :P |
| 09:36 | <mikeday> | actually I think it would be great to be able to use Prince to make PDF forms |
| 09:36 | <mikeday> | it's something we've planning to do for ages, but there's been a lot of other stuff to do as well. |
| 09:36 | <nickshanks> | yes, it would take if from a book publishing tool to a corporate marketing tool |
| 09:37 | <nickshanks> | open up new markets, etc |
| 09:37 | <mikeday> | right |
| 09:37 | <mikeday> | I think that SVG is also a good platform for forms, |
| 09:37 | <nickshanks> | what do most of your customers use the app for? |
| 09:37 | <mikeday> | which are often designed in a page-centric manner, following a grid etc. |
| 09:38 | <mikeday> | some use it for producing printed receipts, some for dynamically generated reports, some for technical documentation, |
| 09:38 | <mikeday> | some for books or academic papers |
| 09:38 | <mikeday> | there's a fair bit of variety. |
| 09:38 | <nickshanks> | hmm, not seen it used for technical papers |
| 09:39 | <nickshanks> | all the ones i read are written in latex |
| 09:39 | <mikeday> | right, we hope to change that :) |
| 09:39 | <nickshanks> | hmm. i'm not sure changing that would help anything other than your bank account |
| 09:40 | <mikeday> | hah |
| 09:40 | <nickshanks> | LaTeX seems like a good solution to me |
| 09:40 | <mikeday> | believe it or not, there are other user agents capable of rendering HTML + CSS |
| 09:40 | <mikeday> | we call them "browsers" |
| 09:40 | <nickshanks> | ooooo! |
| 09:40 | <mikeday> | if web standards aren't a good solution for technical papers, fix web standards |
| 09:41 | <nickshanks> | that's why we're in this channel |
| 09:41 | <mikeday> | indeed. |
| 09:41 | <nickshanks> | is there an extant SVG to PDF conversion tool? |
| 09:41 | <nickshanks> | that retains vectors, rather than rasterising |
| 09:42 | <KevinMarks> | safari + print? |
| 09:42 | <nickshanks> | (no, Adobe Illustrator doesn't count) |
| 09:42 | <mikeday> | Batik |
| 09:42 | <mikeday> | I think you have to rasterise sometimes though, eg. for filters, which we don't support yet |
| 09:45 | <mikeday> | I think that (X)HTML + SVG + MathML is a decent enough combination for technical papers, anyway. |
| 09:46 | <nickshanks> | mikeday: can you consider adding Prince to http://en.wikipedia.org/wiki/Comparison_of_layout_engines_%28SVG%29 |
| 09:46 | <mikeday> | sorry, that would be conflict of interest under Wikipedia policy |
| 09:46 | <mikeday> | how about you do it? :) |
| 09:47 | <nickshanks> | i don't know any of the facts |
| 09:47 | <nickshanks> | i already added prince to http://en.wikipedia.org/wiki/Comparison_of_layout_engines_%28CSS%29 |
| 09:47 | <mikeday> | http://www.princexml.com/doc/svg/ |
| 09:47 | <mikeday> | gives SVG support for Prince 5.1 |
| 09:48 | <mikeday> | we've added a bit more in 6.0 |
| 09:49 | <nickshanks> | wikipedia policy doesn't prohibit adding content that might represent a conflict of interest, just that you provide a reference. leeway is much greater for filling in omitted data in a "hey, you forgot me!" manner |
| 09:50 | <othermaciej> | nickshanks: Safari can print to PDF, though I', not sure how well it works for PDF |
| 09:52 | <nickshanks> | othermaciej: i am thinking of SVG in <img src="foo.svg" style="width:100%"> elements, where the svg is converted to PDF and then passed to NSImage, and stays sharp as you resize the window |
| 09:52 | <mikeday> | yeah, it's mainly that updating Wikipedia would take time that I don't have right now |
| 09:52 | <mikeday> | (he says, while chatting on irc) |
| 09:53 | <othermaciej> | nickshanks: I don't understand what you are thinking of, then, since neither WebKit nor any other browser engine handle PDF that way |
| 09:53 | <othermaciej> | er, SVG |
| 09:53 | <nickshanks> | mikeday: it may also bring in customers. wikipedia is an effective advertising tool |
| 09:54 | <mikeday> | actually, there are several ? on the CSS comparison page that I can fix |
| 09:54 | <nickshanks> | mikeday: just 5 mins a day tending to the article would help make it more complete, and you're the best-placed person to do that |
| 09:55 | <othermaciej> | on http://en.wikipedia.org/wiki/Comparison_of_layout_engines and related, someone should really rename WebCore to WebKit |
| 09:55 | <othermaciej> | that's the name of the engine that people actually use |
| 09:55 | <mikeday> | I can't help wondering if all this stuff really belongs on Wikipedia though |
| 09:56 | <nickshanks> | mikeday: probably not, but it's wiki nature keeps it more up to date than other sites that try and do the same |
| 10:02 | <mikeday> | by the way nickshanks, we should be able to fix the BOM issue that was affecting Prince on HTML documents |
| 10:02 | <mikeday> | as we are planning to write a new HTML parser, following the HTML5 spec |
| 10:04 | <nickshanks> | did i complain about that? can't remember the issue |
| 10:04 | <mikeday> | libxml2 HTML parser chokes on a UTF-8 BOM unfortunately |
| 10:04 | <othermaciej> | information about layout engines seems as valid to me as info about various animal species or physical lawys |
| 10:04 | <othermaciej> | *laws |
| 10:05 | <mikeday> | othermaciej, yes |
| 10:05 | <mikeday> | but as always, the web gives highest priority to information about itself |
| 10:05 | <mikeday> | that, and exhaustive listings of all the Pokemon, Simpsons episodes... |
| 10:06 | <nickshanks> | mike: it represents what people want to read about |
| 10:06 | <mikeday> | people being "geeks on the web" :) |
| 10:06 | <othermaciej> | anyway, I'm just kinda suggesting someone could fix this since per wikipedia my expert knowledge of the webkit project disqualifies me from writing about it |
| 10:06 | <mikeday> | it represents what people want to *write* about |
| 10:06 | <nickshanks> | not necessarily what a learned individual should read |
| 10:07 | <nickshanks> | yeah, sorry, being a wiki i tend to think of reading and writing as the same thing |
| 10:07 | <mikeday> | no one with expert knowledge should be contributing to an encyclopaedia, right :) |
| 10:08 | <nickshanks> | sure they can, as long as they avoid their area of expertise ;) |
| 10:09 | <mikeday> | we only want opinions untainted by any detailed knowledge of the subject in question :) |
| 10:10 | <nickshanks> | we should get jukka, tina and lachy to write the wikipedia article on HTML5 then ;) |
| 10:11 | <othermaciej> | lachy would be tainted by quite a bit of knowledge about HTML5 |
| 10:11 | mikeday | reads the article on HTML5 |
| 10:11 | <mikeday> | short, and to the point |
| 10:11 | <mikeday> | I like the fact that it's in the "Future products" category |
| 10:11 | <mikeday> | looks like hsivonen put that in :) |
| 10:17 | <mikeday> | hmm, it would be nice sometimes if Linux had binary compatibility across versions |
| 10:18 | <mikeday> | (...and a pony). |
| 10:19 | nickshanks | read that as "it would be nice if Linux sometimes had binary compatibility across versions" |
| 10:19 | <nickshanks> | like, on thursdays and saturdays... |
| 10:24 | <mikeday> | heh |
| 10:24 | <mikeday> | that's pretty much what it feels like now |
| 10:24 | <mikeday> | things may or may not run, depending on glibc and the phase of the moon |
| 10:24 | <mikeday> | one may well ask why we bother with dynamic linking at all |
| 10:25 | <mikeday> | we hardly need to worry about saving disk space |
| 10:25 | <nickshanks> | see now, if you moved to the sun, the moon would always be in the same phase |
| 10:25 | <nickshanks> | and you'd get reliability |
| 10:25 | <mikeday> | and the simplification of dropping dynamic loading and relocation would more than make up for it |
| 10:26 | <nickshanks> | the best would be to ship the lib you built against, and then give the user the choice to use his own if it's a newer version |
| 10:27 | <nickshanks> | that way the user can always fall back on the version you built against if anything goes wrong with the new lib |
| 10:27 | <mikeday> | yeah, but if you moved to Sun, you'd have to run Solaris, not Linux |
| 10:27 | <nickshanks> | that's fine with me |
| 10:28 | <nickshanks> | my astronomy dept. has lots of Suns |
| 10:28 | <mikeday> | astronomers are known for their fascination with suns |
| 10:29 | <mikeday> | ...and a weak joke collapses under the strain. |
| 10:29 | <nickshanks> | i wasn't joking |
| 10:29 | <mikeday> | 'scuse me |
| 10:29 | <nickshanks> | i did astrophysics at uni and we all worked on Solaris machines |
| 10:30 | <nickshanks> | writing f77 and stuff |
| 10:31 | <BenWard> | Not sure how new or old this is — not been in the channel very long this morning — but I trust someone has seen that the WHATWG blog has been hacked and spammed? |
| 10:31 | <othermaciej> | lovely |
| 10:33 | <Dashiva> | Let the blame passing begin! |
| 10:35 | <nickshanks> | clearly it's TimBL's fault for inventing HTML and so causing the WHATWG blog to exist |
| 10:56 | <BenWard> | No no no, it's Microsoft's fault for not producing a robust enough proprietary internet publishing solution, requiring us to develop these open standards in the first place. |
| 10:57 | <Lachy> | yay! Our first spam articles :-) http://blog.whatwg.org/ |
| 10:58 | <BenWard> | … or, it's because you're running Wordpress 2.0.5 and not 2.0.10. |
| 10:59 | <Lachy> | yeah, I need to update WP one day |
| 10:59 | <nickshanks> | WP 2.2 is the latest |
| 10:59 | <othermaciej> | this is a strangely ironic URL: http://whatwg.majda.cz/wallpapers/5-gt-2-black-simple.svg |
| 11:00 | <BenWard> | 2.2 is the latest, but the 2.0 branch is still being patched for security (hence 2.0.10) and shouldn't break any plug-ins that might be running. |
| 11:01 | <nickshanks> | anyway, have to install hard disk and some OSes, bbl |
| 11:07 | <Lachy> | I just reduced the rights of new users to contributor instead of author. That will allow them to register and write posts, but can't publish themselves. Once they prove to not be a spammer, I can upgrade them |
| 11:39 | <ROBOd> | good day. spammers found the whatwg blog? |
| 11:39 | <ROBOd> | i've got two spam articles from the blog, via atom feed |
| 11:43 | <Lachy> | ROBOd, yeah, I've already deleted them |
| 11:43 | <ROBOd> | i saw that |
| 12:11 | <zcorpan> | i think i'm starting to understand role a bit better now |
| 12:12 | <zcorpan> | it's "needed" because <input type=checkbox> isn't stylable in current browsers. so when you need a fancy checkbox, you use <img onclick>. but then screen readers don't know it's a checkbox. |
| 12:13 | <zcorpan> | i think it's the wrong way to go about it. the right way is to make <input type=checkbox> stylable. |
| 12:13 | <zcorpan> | or am i missing something? |
| 12:17 | <peepo> | hi folks it's a rare event.. |
| 12:18 | <peepo> | oops |
| 12:19 | <zcorpan> | [type=checkbox] { content:url(unchecked.png); } [type=checkbox]:checked { content:url(checked.png); } |
| 12:44 | <mikeday> | zcorpan, role is for stylable form controls? I thought it was more than that. |
| 12:45 | <Dashiva> | role is magical |
| 12:45 | <Dashiva> | Somehow everyone will know and agree on which roles mean what, and never use them incorrectly or make up their own roles |
| 12:47 | <mikeday> | what code will access the role attribute? |
| 12:47 | <mikeday> | I mean, CSS could, with [role=foo], but I'm assuming that's undesirable. |
| 12:48 | <mikeday> | screen readers? |
| 12:49 | <mikeday> | search engines? |
| 12:51 | <mikeday> | automatic text summarisers? |
| 12:51 | <mikeday> | throw me a bone here, people :/ |
| 12:51 | <Dashiva> | The people talking most about it are the accessability people, so screen readers |
| 12:52 | <mikeday> | hmm. |
| 12:52 | <mikeday> | <i role="shipname">Titanic</i> for example? |
| 12:53 | <Dashiva> | Sure |
| 12:53 | <mikeday> | how is role supposed to connect to RDF, anyway? |
| 12:53 | <mikeday> | I mean how does it differ from class? |
| 12:53 | <Dashiva> | See my initial statement |
| 12:55 | <mikeday> | hmm, from what I see on the mailing list, people want to say role="foo:shipname" |
| 12:55 | <mikeday> | where foo is a currently bound namespace prefix |
| 12:55 | <mikeday> | -1 for qnames in content... |
| 12:56 | <zcorpan_> | i don't see the point in linking to RDF |
| 12:56 | <mikeday> | how exactly will user agents "learn" about new roles |
| 12:56 | <zcorpan_> | is it really intended that UAs will read the RDF and then magically know what to do with the role? |
| 12:56 | <mikeday> | exactly! |
| 12:56 | <mikeday> | that was the kind of insane magic thinking that came out when XML Namespaces where designed |
| 12:57 | <mikeday> | "a namespace is a URL, user agent dereferences URL, downloads schema for namespace (why?), magically knows what to do" |
| 12:57 | <zcorpan_> | was it? |
| 12:58 | <mikeday> | RDDL was a language designed for that very purpose |
| 12:58 | <mikeday> | to add namespace metadata to an XHTML page that sits on the namespace URL |
| 12:58 | zcorpan_ | sighs |
| 12:58 | <mikeday> | so that hypothetical user agents could go and download it and learn new tricks |
| 12:59 | <mikeday> | I don't know how many smart people burnt processor cycles on that idea |
| 12:59 | <mikeday> | the fact that most user agents don't even bother downloading DTDs should have been a hint |
| 12:59 | <mikeday> | (Prince does, actually :) |
| 12:59 | <zcorpan_> | does it learn new tricks from DTDs? :) |
| 12:59 | <mikeday> | sadly, no :) |
| 13:00 | <mikeday> | but wouldn't it be awesome if you could just define new elements in the DTD, |
| 13:00 | <mikeday> | and they got all kinds of custom rendering behaviour automatically, |
| 13:00 | <mikeday> | and we didn't have to do any work! That'd rock. |
| 13:00 | <mikeday> | ...and a pony. |
| 13:02 | <mikeday> | I mean, the only thing you can do in RDF that doesn't involve writing code, |
| 13:02 | <mikeday> | is tell the user agent that a foreign element is equivalent to some other known element |
| 13:02 | <mikeday> | like, if you see a <foo>, treat it just like you currently treat <blockquote> |
| 13:03 | <mikeday> | now that might be handy, from time to time, but it's hardly a world changing feature. |
| 13:03 | <zcorpan_> | i meant with my first remark that it appears that that was what spawned the role idea in the first place. then they thought "hey, this can be useful for other stuff too!" or something |
| 13:03 | <zcorpan_> | and made it even more complicated |
| 13:03 | <mikeday> | so it was originally proposed for styling form controls? |
| 13:03 | <mikeday> | seems like improved CSS selectors would solve that one |
| 13:04 | <zcorpan_> | dunno, just speculating. i don't understand the problem it's trying to solve really |
| 13:04 | <zcorpan_> | indeed |
| 13:04 | <zcorpan_> | or selectors are already powerful enough |
| 13:04 | <mikeday> | role was from XHTML 2 originally though, right? |
| 13:04 | <zcorpan_> | probably. does it make a difference? |
| 13:05 | <mikeday> | just trying to find out where it all began |
| 13:05 | <mikeday> | looks like it was an XHTML module: |
| 13:05 | <mikeday> | http://www.w3.org/TR/2006/WD-xhtml-role-20060725/ |
| 13:05 | <zcorpan_> | that came afterwards |
| 13:05 | <zcorpan_> | as an attempt to "ease migration to xhtml2" (or so the draft says) |
| 13:05 | <mikeday> | It is used by applications and assistive technologies to determine the purpose of UI widgets. |
| 13:05 | <mikeday> | ah, okay |
| 13:05 | <mikeday> | (UI widgets sentence should be in quotes) |
| 13:06 | <zcorpan_> | yeah |
| 13:06 | <zcorpan_> | thing is there already is HTML markup for UI widgets! |
| 13:06 | <zcorpan_> | <button>, <a href>, <input type=checkbox> |
| 13:06 | <mikeday> | the example has it on a list: <ul role="navigation wai:sitemap"> |
| 13:06 | <zcorpan_> | <nav> |
| 13:06 | <mikeday> | heh :) |
| 13:07 | <zcorpan_> | <nav><h2>Sitemap</h2><ul>... |
| 13:07 | <mikeday> | to be honest, some of these role values remind me of rel="" |
| 13:08 | <mikeday> | I can sort of see what they're trying to do with it, and it's not *too* insane |
| 13:08 | <mikeday> | but the idea of qnames in attributes, and linking it to RDF via namespace URLs is just stupid |
| 13:08 | <zcorpan_> | role seems to be like an accessibility patch for poor CSS support in browsers and badly written markup |
| 13:08 | <mikeday> | the whole point is to have a small selection of values that user agents understand |
| 13:08 | <zcorpan_> | which might be a good thing |
| 13:08 | <mikeday> | some of it could also be handy for mobile rendering, |
| 13:09 | <mikeday> | like when you want to chop the page up into bits to fit on a small screen |
| 13:09 | <mikeday> | you want to know where to chop, and role could help with that |
| 13:09 | <zcorpan_> | i mean, if there is old content with <img onclick>, it's easier to add role=checkbox than to change it to <input type=checkbox> and restyle it |
| 13:09 | <mikeday> | but again, define a small number of roles, make sure user agents understand them, ditch the namespace stuff |
| 13:09 | <zcorpan_> | yeah |
| 13:09 | <mikeday> | people might go nuts inventing new role values, but that's fine |
| 13:10 | <mikeday> | since it's not useful for styling, only accessibility nuts will go nuts, which is okay |
| 13:11 | <mikeday> | hmm, just like the rel attribute, actually |
| 13:11 | <mikeday> | rel="shortcut icon" |
| 13:11 | <mikeday> | and the hook for OpenID |
| 13:12 | <mikeday> | rel="openid.server", works fine with no namespace support |
| 13:12 | <mikeday> | it's just that every W3C specification has to pay homage to RDF at some point... |
| 13:13 | <zcorpan_> | indeed, i don't see why namespaces are better than naming conventions |
| 13:14 | <zcorpan_> | naming conventions seem to be more successful in terms of usage and implementation |
| 13:14 | <zcorpan_> | microformats are naming conventions |
| 13:15 | <zcorpan_> | some implementations of RSS rely on the convention of how the namespace prefixes are written! |
| 13:15 | <zcorpan_> | (i think) |
| 13:17 | <mikeday> | yes |
| 13:17 | <mikeday> | pseudo-XML formats, not really XML |
| 13:17 | <mikeday> | but namespaces should never have been defined using URLs to begin with |
| 13:17 | <mikeday> | http://www.xml.com/pub/a/2005/04/13/namespace-uris.html |
| 13:22 | <mikeday> | must go |
| 13:23 | mikeday | waves |
| 13:23 | <zcorpan> | cya |
| 13:23 | <zcorpan> | have a predefined set of prefixes, and if you don't want to use a prefix, then you have to declare the namespace |
| 13:23 | <zcorpan> | <html:html/> == <html xmlns="html"/> |
| 13:23 | <zcorpan> | that could have worked |
| 13:24 | <zcorpan> | unknown prefixes are just that, handled the same way as unknown namespaces are today |
| 15:02 | <met_> | http://www.readwriteweb.com/archives/poll_update_mozilla_google_favored.php |
| 15:18 | <zcorpan_> | "Mozilla (open source, microformats, browser-based)" translates pretty much to "HTML5" :) |
| 15:21 | <met_> | mozilla is still better buzzworld than html5 8-) |
| 15:21 | <met_> | *buzzword |
| 15:23 | <Philip`> | met_: I noticed you sent quite a few people to my site :-) |
| 15:26 | <met_> | Philip` i wrote about canvas in my blog 8-) and you doom-like game is perfect example |
| 15:27 | <met_> | today I have found http://solutoire.com/plotr/ - js library for generating charts in canvas |
| 17:18 | <zcorpan_> | somehow i think svg is more suitable for static charts |
| 17:19 | <zcorpan_> | although plotr is probably easy to use for dynamic charts too |
| 17:32 | <Philip`> | Plotr seems to be heavily based on PlotKit, which does have SVG as an option |
| 17:33 | <Philip`> | but it's perhaps a problem that IE doesn't do SVG, whereas you can sort of make it do <canvas> |
| 17:33 | <Philip`> | although it's also a difficulty that <canvas> can't do text, whereas SVG can |
| 17:44 | <zcorpan_> | you can make IE do SVG the same way you can make it do canvas |
| 17:44 | <zcorpan_> | (with VML) |
| 17:50 | <Philip`> | Has anybody actually done that? |
| 17:54 | <zcorpan_> | yeah |
| 17:54 | <zcorpan_> | http://starkravingfinkle.org/blog/2006/03/svg-in-ie/ |
| 18:02 | <Philip`> | Ah, looks neat |
| 18:02 | <Philip`> | (Seems it won't work for SVG generated at run-time, though) |
| 18:11 | <ianloic> | in theory you could though, right? |
| 18:11 | <ianloic> | does IE support the dom mutation events? |
| 20:44 | <met_> | http://www.davidflanagan.com/blog/2007_05.html#000128 |
| 20:59 | zcorpan_ | expects at least 15 new subscribers who will request dashed lines |
| 21:03 | <jruderman> | met_: i'm not surprised. i think i told hixie to expect things like that when he started rejecting things with "not enough requests" as the reason. |
| 21:04 | <jruderman> | zcorpan_: given that it's flanagan's blog... only 15? :P |
| 21:04 | <gsnedders> | dashed lines plz. kthxbai. |
| 21:05 | <Philip`> | I think that was "not enough demand" rather than "not enough requests" - people who are actually writing canvas content and finding that they'd benefit from native support for dashed lines are a more significant indicator of demand than people who just pop in to say it'd be nice :-) |
| 21:09 | <Hixie> | yeah |
| 21:09 | <Hixie> | case in point, the MNG fiasco |
| 21:09 | <Hixie> | where there were more people voting for MNG support in Firefox than there were MNGs created and put on the Web |
| 21:09 | <Hixie> | it's not a matter of votes |
| 21:10 | <Hixie> | it's a matter of organic demand |
| 21:10 | <Hixie> | if everyone comes in and says "hey this guy asked me if i could ask for dashed lines", those all count as one request |
| 21:11 | <Lachy> | are there people using canvas who are creating their own dashed lines? |
| 21:11 | <Hixie> | nice work handling the blog spam btw |
| 21:11 | <Hixie> | Lachy: i've seen one, so far |
| 21:11 | <Lachy> | was it difficult to do? |
| 21:11 | <Hixie> | Lachy: they handled it quite easily (with a pattern) |
| 21:11 | <Hixie> | but they were just doing straight lines |
| 21:11 | <Lachy> | ok |
| 21:13 | <Philip`> | There's http://canvaspaint.org/ with the selection tool |
| 21:13 | <Philip`> | which is presumably that one |
| 21:15 | <Hixie> | yeah |
| 21:15 | <Hixie> | and you did a demo page with dashed lines, which was pretty neat |
| 21:15 | <Hixie> | even a dashed circle! |
| 21:16 | <Lachy> | canvaspaint is awesome |
| 21:17 | <Philip`> | My one seems to break a bit on corners, though |
| 21:18 | <Philip`> | and implementing decent Bezier curves in JS would be less trivial than one might prefer, compared to just setting a dash option |
| 21:19 | <Hixie> | would anyone doing a graph really use bezier curves? |
| 21:20 | <Hixie> | dashed lines are so Web 1.0. I didn't say it in the thread, but come on, who uses dashed lines these days. They're so old-school. |
| 21:20 | <Hixie> | it's all about bold lines and contrasting colours now. |
| 21:22 | <zcorpan_> | semi-transparent bold lines with contrasting colors :) |
| 21:23 | <Hixie> | exactly! |
| 21:23 | <Philip`> | http://www.mactech.com/articles/mactech/Vol.21/21.08/Mineralogy101/fig10.jpg - but you can do such useful things with dashes |
| 21:23 | Philip` | can't think of any sensible cases for drawing dashed non-straight lines |
| 21:24 | <Hixie> | that was sort of why i wasn't convinced |
| 21:24 | <Hixie> | but anyway |
| 21:24 | <Hixie> | gotta go |
| 21:24 | <Hixie> | bbiab |
| 21:25 | <Dashiva> | A dashed bezier curve! |
| 21:25 | <met_> | what about charts? there are often non-solid lines |
| 21:32 | <Philip`> | met_: Those normally just use straight lines, and you can simulate dashes easily by drawing lots of shorter lines |
| 21:33 | <zcorpan_> | googling for "dynamic chart" doesn't give any curved dashed lines on the first 3 pages of results afaict |
| 21:34 | <zcorpan_> | some with straight dashed lines, though |
| 21:34 | <met_> | Philip` a lot of work 8-) |
| 21:35 | met_ | foun dotted when googling xy chart |
| 21:39 | <mpt> | there's those conspiracy charts |
| 21:39 | <mpt> | which I now won't be able to find |
| 21:39 | <mpt> | but they have curved dashed lines between all the people |
| 21:42 | <mpt> | http://www.pierogi2000.com/flatfile/lombardi.html |
| 21:42 | <mpt> | ok, they have curved dashed lines between some of the people |
| 21:43 | <zcorpan_> | would you generate such a chart with canvas? well, perhaps, linking XFN people together real-time or something :) |
| 22:44 | <tantek> | zcorpan, indeed - see what http://xhtmlfriends.net/ does |
