| 00:37 | <Philip`> | http://philip.html5.org/demos/apng/blending/ - highly useful |
| 00:37 | <Philip`> | http://philip.html5.org/demos/apng/sierpinski.png doesn't work so well in Opera since the blending gets broken :-( |
| 00:38 | <Philip`> | (White blended onto white really shouldn't give grey) |
| 01:39 | <Philip`> | The APNG spec seems to be a little loose with its use of normative language |
| 02:12 | <Hixie_> | ok the whatwg.org server is now running under DreamhostPS (a virtual server system) |
| 02:12 | <Hixie_> | the 500 errors were because i had the RAM setting too low |
| 02:12 | <Hixie_> | i doubled it |
| 02:12 | <Hixie_> | we'll see if that helps |
| 02:30 | <Hixie_> | fyi for anyone here who has an account on my dreamhost plan (lachlan, zcorpan, charlvn...) -- the ssh key changed this weekend. that's normal. sorry for any inconvenience. |
| 04:16 | <Lachy> | othermaciej, I have a good example for degrade gracefully |
| 04:17 | <Lachy> | oh, never mind, it's already in the list |
| 04:17 | <othermaciej> | heh :-) |
| 04:17 | <othermaciej> | what was it going to be? |
| 04:17 | <Lachy> | datalist |
| 04:17 | <Lachy> | I missed the first time I looked |
| 04:18 | <othermaciej> | I was thinking of adding <dialog> |
| 04:18 | <othermaciej> | since it works mostly without even the need for CSS styling |
| 04:18 | <Lachy> | you could do <input pattern="">, that was designed to be easy to simulate using script |
| 04:19 | <othermaciej> | but it does have the IE new element problem |
| 04:20 | <Lachy> | for dialog, one could style all dt and dd elements and then undo any unwanted styles for dl dt and dl dd |
| 04:21 | <othermaciej> | datalist is also not so great in IE |
| 04:21 | <Lachy> | why? |
| 04:21 | <othermaciej> | the contents of the options render |
| 04:21 | <Lachy> | not if you use <option value=""> instead <option>value</option> |
| 04:21 | <othermaciej> | and you get OPTION and /OPTION as void elements in the DOM |
| 04:21 | <othermaciej> | ah |
| 04:22 | <Lachy> | or you do what I did here using select http://html5.lachy.id.au/ |
| 04:23 | <othermaciej> | yeah, that's not bad, works even in IE with relatively minor tweaking of the markup |
| 04:23 | <othermaciej> | I might take a stab at rewriting the next two principles tonight or tomorrow |
| 04:24 | <othermaciej> | those will be fun |
| 04:24 | <Lachy> | cowpaths and wheel? |
| 04:24 | <othermaciej> | yeah |
| 04:25 | <Lachy> | are you going to rename cowpaths? |
| 04:25 | <Lachy> | (please don't) |
| 04:31 | <othermaciej> | I think I need to |
| 04:31 | <othermaciej> | I like the name |
| 04:31 | <othermaciej> | but if I hear one more person say "we shouldn't pave this cowpath" I will stab my eyes out |
| 04:35 | <othermaciej> | I'm not sure how else to do that |
| 04:35 | <othermaciej> | I don't think clarification or an education campaign will do it |
| 04:36 | <othermaciej> | Apparently I haven't even managed to make "Degrade Gracefully" clear, even with the detailed description and wealth of examples |
| 04:57 | <Lachy> | I just don't understand why so many people are having difficulty understanding such simple concepts like graceful degradation |
| 04:58 | <othermaciej> | nothing about the web is simple |
| 04:58 | <othermaciej> | I gotta go |
| 04:58 | <othermaciej> | later |
| 04:59 | <Lachy> | bye |
| 05:33 | <MikeSmith> | Lachy - I think some people who have posted recently about the design principles are just plain confused and lack some of the necessary background/socialization needed to understand the context for the discussion |
| 05:34 | <MikeSmith> | I'm not sure that any amunt of re-wording is going to help them understand |
| 05:39 | <Lachy> | yeah, I realise that. Though there are some people who really should have the necessary background by now, but still don't |
| 09:44 | <Hixie> | pull quotes are i think what henri was talking about |
| 09:44 | <Hixie> | but any of these would be valid uses of <aside> |
| 09:44 | <Hixie> | if html5 (the spec) were written using html5, then the examples, the issues, and some of the notes would be <aside>s. |
| 09:44 | <Hixie> | it's just anything that isn't part of the main flow |
| 09:45 | <Lachy> | would a pullquote be best marked up using <aside><blockquote>pull quote</blockquote></aside>? |
| 09:45 | <Lachy> | or maybe <q> instead. |
| 09:46 | <jgraham> | Lachy: "from article to the next" s/from/from one/ |
| 09:46 | <Lachy> | Hixie, you should probably make a note of that as a comment in the spec or something for when you get around to adding usage examples for the elements. |
| 09:48 | <hsivonen> | Hixie: has the cite attribute undergone the same kind of usefulness scrutiny as longdesc? |
| 09:48 | <Lachy> | jgraham, fixed |
| 09:49 | <Lachy> | hsivonen, I don't think it has. As far as I know, no UA does anything useful with it and <cite><a> is a better solution. |
| 09:49 | <Hixie> | Lachy: i guess |
| 09:49 | <Lachy> | every time I've used it, I've always duplicated it with a link, and that's just not worth the effort |
| 09:49 | <Hixie> | Lachy: (re blockquote) |
| 09:50 | <Hixie> | Lachy: yeah... send mail :-D |
| 09:50 | <Hixie> | hsivonen: not to my knowledge |
| 09:50 | <Lachy> | Hixie, send mail about the pullquote example? |
| 09:51 | <Lachy> | or the cite issue? |
| 09:51 | <Lachy> | oh, I see, the blockquote. I should read everything you write before responding :-) |
| 09:53 | <jgraham> | Lachy: my other comment about the article is that it's a little terse; in places it reads a bit more like a spec or a technical email than a friendly introduction (but I guess you haven't finished editing it yet) |
| 09:54 | <Hixie> | Lachy: send mail about anything you want in the spec, basically :-) |
| 09:54 | <Hixie> | feel free to mail whatwg if you don't want to cause a panic each time |
| 09:54 | <Hixie> | all the same to me |
| 09:54 | <Lachy> | jgraham, yeah, I definitely need to shorten the article a bit |
| 09:55 | <Lachy> | any particular sections? |
| 09:55 | <Lachy> | Hixie, yeah, I was sending it to whatwg |
| 09:55 | <hsivonen> | Lachy: why is "Strict XML Syntax." a benefit? :-) |
| 09:55 | <Lachy> | some authors like the immediate feedback about errors |
| 10:05 | <zcorpan> | html allows that too, you could have a setting in browsers to make parse errors fatal |
| 10:05 | <zcorpan> | hidden setting, of course |
| 10:18 | <zcorpan> | Lachy: s/XHTML serialisation/XML serialisation/ ? |
| 10:19 | <zcorpan> | Lachy: s/Browsers will/Browsers/ |
| 10:24 | <Lachy> | zcorpan, a draconian HTML5 parser is more likely to appear in an authoring tool than it is in a browser, even with a hidden pref |
| 10:24 | <zcorpan> | Lachy: indeed |
| 11:28 | <hsivonen> | Hixie: did you consider tri-state checkboxes for WF2.0? Tri-state checkboxes seems to be the most common example for demonstrating the need for ARIA. |
| 11:31 | <zcorpan> | Lachy_: shouldn't the Excerpt from A Tale of Two Cities by Charles Dickens have <br>s? |
| 11:31 | <Lachy_> | Are the new lines semantically important? |
| 11:32 | <Lachy_> | hsivonen, tri-state checkboxes were discussed in IRC a few days ago |
| 11:32 | <Lachy_> | I think it was in #html-wg |
| 11:32 | <zcorpan> | Lachy_: hmm, perhaps not, it looked like a poem |
| 11:32 | <hsivonen> | Lachy_: hmm. I thought I've read all the logs |
| 11:33 | <Lachy_> | maybe krijnh didn't log it, let me check |
| 11:34 | <hsivonen> | Lachy_: found it: http://krijnhoetmer.nl/irc-logs/html-wg/20070914#l-510 |
| 11:34 | <hsivonen> | looks like I skipped it as something that I thought was still part of the telecon and that I could digest from the minutes |
| 11:35 | <Lachy_> | yeah, there's a little bit further down with Hixie, mjs and I too |
| 11:36 | <hsivonen> | Lachy_: Re: http://krijnhoetmer.nl/irc-logs/html-wg/20070914#l-545 based on my understanding about the motivations, ARIA has been written with authoring considerations in mind--just not the ones you'd expect |
| 11:36 | <hsivonen> | Lachy_: it isn't optimized for authoring new apps. |
| 11:36 | <hsivonen> | Lachy_: it is optimized for retrofitting into legacy Ajax widgets |
| 11:38 | hsivonen | finds http://krijnhoetmer.nl/irc-logs/html-wg/20070914#l-562 |
| 11:38 | <Lachy> | I highlighted all the discussion of tri-state, just reload the logs |
| 11:39 | <hsivonen> | Lachy: thanks |
| 11:39 | <Lachy> | oh, I missed some |
| 11:57 | <hsivonen> | hmm. I wonder if getting search engines to seach embedded text-based closed captioning data in video files would be a curse or a blessing for the accessibility quality of captions. |
| 11:58 | <Lachy> | I would hope publishers realise that captions are useful for more than just the deaf. e.g. users at work without speakers or who need to keep the volume low |
| 11:59 | <Lachy> | so they would get complaints if they started using caption files for keyword stuffing |
| 12:42 | <Lachy> | looks like Rob didn't realise that jgraham's HTML5 table header implementation already makes improvements upon the spec, rather than intending to follow it as-is. http://html4all.org/pipermail/list_html4all.org/2007-September/000236.html |
| 12:44 | <Lachy> | and that what he's asking for (that we develop and refine an algorithm that actually works) is what jgraham and others are alread attempting |
| 12:45 | <Lachy> | wow, and he has a selective interpretation of the HTML4 algorithm, where it implicitly means what he wants it to mean. http://html4all.org/pipermail/list_html4all.org/2007-September/000237.html |
| 13:30 | <Lachy> | jgraham, yt? |
| 13:59 | <virtuelv> | hm, one mildly bothering thing about <object> |
| 14:00 | <virtuelv> | <object type="[unsupported]" title="Foo"> <object type="[supported]"> </object></object> |
| 14:00 | <virtuelv> | should title be rendered by a UA? |
| 15:29 | <zcorpan> | virtuelv: yes |
| 15:29 | <virtuelv> | zcorpan: what with <object type="[unsupported]" title="foo"><object type="[supported]" title="bar">? |
| 15:30 | <virtuelv> | foo or bar, or both? |
| 15:30 | <virtuelv> | also, are you sure the title for foo actually applies to bar? |
| 15:31 | <zcorpan> | bar (assuming that the supported object completely covers the unsupported) |
| 15:31 | <zcorpan> | title is inherited if title is not present (with some exceptions, e.g. <link> and <style>) |
| 15:32 | <zcorpan> | object is not an exception per the current spec, iirc. |
| 15:33 | <zcorpan> | though i can imagine that ie doesn't add the unsupported object element to the dom at all, and so drops the title |
| 15:33 | <zcorpan> | (does it?) |
| 16:13 | <gsnedders> | what's the best MathML editor? |
| 16:15 | <gsnedders> | (I know, a subjective question) |
| 16:24 | <Philip`> | Maybe a text editor and a LaTeX-to-MathML converter |
| 16:26 | <zcorpan> | perhaps the table inspector needs a heuristic to tell layout tables and data tables apart? :) |
| 16:26 | <zcorpan> | perhaps such a heuristic should be specced |
| 16:26 | <Philip`> | (Hmm, apparently TeXmacs can export XHTML+MathML too - I wonder if that actually works...) |
| 16:27 | gsnedders | is actually tempted to use a WYSIWYG editor for the first time in years (namely Amaya) |
| 16:28 | <Philip`> | (I vaguely remember TeXmacs being actually alright as a WYSIWYG equation editor, but I might just be remembering wrong) |
| 16:30 | <zcorpan> | this is interesting: <label for=foo>LABEL</label> <a href=# id=foo>LINK</a> |
| 16:30 | gsnedders | is thinking that he'll probably put only all of his Higher Maths notes online, just in case they help anyone |
| 16:30 | <zcorpan> | in ie, if you click the label, it follows the link |
| 16:30 | <gsnedders> | it'd be nicer to have them digitally organised rather than the paper mess they are now anyway |
| 16:30 | <gsnedders> | zcorpan: I doubt anything breaks from that behaviour, and it seems logical to me |
| 16:32 | <zcorpan> | safari is same as ie, in firefox and opera the label is dead |
| 16:33 | <gsnedders> | Saf3, I assume? |
| 16:33 | <zcorpan> | yeah |
| 16:35 | <zcorpan> | breaks pages that use <label>Label (<a href>help</a>): <input></label> |
| 16:37 | <Lachy> | gsnedders, http://www.w3.org/Math/iandi/ |
| 16:38 | <zcorpan> | same with <span tabindex=0>; seems in ie a label can point to anything that is in the tab order |
| 16:38 | <zcorpan> | also some things that are not in the tab order |
| 16:39 | <zcorpan> | er... ie's implicit label association seems to be very broken |
| 16:40 | <zcorpan> | it seems to point to the first element child, regardless of what that is |
| 16:40 | <zcorpan> | <label><span>foo</span><input></label> |
| 16:43 | <othermaciej> | Lachy: would you like to make an argument not renaming the "Pave the Cowpaths" principle before I start editing it? |
| 16:44 | <Lachy> | I think I've made all my arguments for keeping the name in emails I've sent on the topic |
| 16:46 | <Lachy> | what alternative name are you considering? |
| 16:46 | <othermaciej> | you've sent a *lot* of emails on the topic :-) |
| 16:47 | <Lachy> | and I expect you to read each and every one of them :-) |
| 16:47 | <othermaciej> | mostly they seemed to be about the content, not the name |
| 16:47 | <othermaciej> | my current thinking is: |
| 16:47 | <othermaciej> | "Don't Reinvent the Wheel" ==> "Consider Existing Implementations" |
| 16:47 | <othermaciej> | "Pave the Cowpaths" ==> "Study Authoring Practices" |
| 16:49 | <Lachy> | study authoring practices might be ok, if you can make it clear that it's mostly about studying *what* they're doing rather than just *how* they're doing it. |
| 16:49 | <othermaciej> | yes, that's the idea |
| 16:50 | <tantek> | you may find this distinction useful: http://microformats.org/wiki/process-faq#Why_waste_time_wading_through_flakey_HTML |
| 16:50 | <tantek> | (regarding authoring practices) |
| 16:51 | <Lachy> | tantek, thanks. I was looking for something like that in the microformats wiki a few days ago. |
| 16:52 | <othermaciej> | tantek: thanks |
| 16:53 | <othermaciej> | Lachy: as for the new wording of the actual principle I intend to base it a lot on your proposal |
| 16:53 | <Lachy> | othermaciej, just copy and paste that, and you're done :-) |
| 16:53 | <Lachy> | oh yeah, that too |
| 16:54 | tantek | needs to do more documentation and braindumping of the microformats principles, the reasoning/experience behind them, and more details. |
| 17:00 | <Lachy> | I really like the last paragraph of this post http://www.windley.com/archives/2005/07/microformats.shtml |
| 17:00 | <Lachy> | "HTML has never been about purity—it’s been about practically solving problems." |
| 17:37 | <Lachy> | good example of how wai-aria is over engineered. http://juicystudio.com/article/wai-aria-in-html.php (though he's using it as an example of how good it is) |
| 17:38 | <Lachy> | I think it would be better to make the custom slider widget (or any other type of custom control) invisible to assistive technology and let them access the <select> directly. |
| 17:38 | <Lachy> | alternatively, just use <input type=range> when supported |
| 17:40 | <Lachy> | XBL will also be good for making custom widgets like that when supported too |
| 17:43 | <zcorpan> | the main advantage of aria over basic html plus css/xbl, afaict, is that it is simpler to apply to existing apps that abuse divs for custom controls. you can slap aria in there without changing anything else, while css/xbl would probably require a rewamp of the app |
| 17:43 | <zcorpan> | however, i don't expect a significant amount of authors to actually do so, and i don't even expect those who do to do it correctly (because it's hard to test) |
| 17:43 | <zcorpan> | and it doesn't apply to new apps |
| 17:45 | <zcorpan> | (where "basic html" includes new stuff from wf2) |
| 17:48 | <Dashiva> | "A touch of class" seems to be popping up everywhere |
| 17:50 | <zcorpan> | another thing that disturbs me is that it seems to be possible to use aria in many different ways. role attribute in no namespace, role attribute in the xhtml2 namespace (?), in the xhtml namespace (?), the states being attributes in some namespace, or attributes in no namespace but the local name begin with "wairole:", or attributes in no namespace and the local name begin with "aria-"... |
| 17:50 | <zcorpan> | and browsers may have to support them all |
| 17:51 | <Lachy> | aargh! You've got to be kidding! |
| 17:51 | <zcorpan> | no |
| 17:52 | <zcorpan> | http://lists.w3.org/Archives/Public/public-pfwg-comments/2007JulSep/0000.html , https://bugzilla.mozilla.org/show_bug.cgi?id=391713 , https://bugzilla.mozilla.org/show_bug.cgi?id=395909 |
| 17:54 | <zcorpan> | perhaps we should try to make aria namespaceless completely and support only one way to do it |
| 17:54 | <Lachy> | ok, well in that case, I guess the HTMLWG will be forced to spec ARIA in HTML (even though it's a disaster), just so that it's not a complete disaster |
| 17:56 | <Lachy> | if we do it, there should be no namespace syntax at all in the HTML serialisation |
| 17:56 | <Lachy> | (or at least not the xmlns syntax and aaa:foo |
| 17:56 | <zcorpan> | indeed, e.g. role="checkbox" aria-hidden="true" |
| 17:57 | <zcorpan> | and then the same syntax can be used in xhtml as well |
| 17:57 | <Lachy> | are there any clashes between aria states and properties, and existing HTML attributes? |
| 17:57 | <zcorpan> | yeah, e.g. required |
| 17:58 | <Lachy> | if possible, just to avoid further namespace disasters like you descrbied above, the HTML parser should place the attributes into the correct namespace automatically |
| 17:58 | <zcorpan> | and i'm not sure we want to have all states and properties without any prefix, it would eat a lot of useful names we might want to use in the future |
| 17:58 | <Lachy> | just like the proposed <math> element would automatically go into the mathml namespace |
| 17:58 | <zcorpan> | why have namespaces at all? |
| 17:59 | <zcorpan> | if the parser puts them in a namespace you would have to use different syntax in xhtml |
| 17:59 | <Lachy> | because that complicates things. implementations would have to support the attributes in the aria namespace for XML and the null namespace on XHTML elements |
| 17:59 | <Lachy> | I don't mind having a different syntax in XHTML |
| 17:59 | <Lachy> | (I don't see aria ever getting used much in practice anyway, so not a big deal)_ |
| 17:59 | <zcorpan> | ok. i think it would be simpler to just have all attributes in no namespace |
| 18:00 | <Lachy> | I think it's too late for that. implementations already support them in the aria namespace |
| 18:00 | <Lachy> | and we'd need to be compatible with exisitng implementations. they don't support them in the null namespace |
| 18:01 | <zcorpan> | firefox does (iirc) |
| 18:01 | <zcorpan> | though not "aria-foo" (yet) |
| 18:02 | <Lachy> | those bugs aren't marked as fixed, I assume they currently don't (unless the patches have been checked in for testing already) |
| 18:05 | <zcorpan> | aiui firefox supports "wairole:foo" in no namespace already |
| 18:05 | <Lachy> | is it really in no namespace in the DOM? |
| 18:07 | <zcorpan> | yeah. though i could be mistaken about whether it supports it |
| 18:09 | <zcorpan> | oh, and class names. firefox supports class names (without the author using a script) i think |
| 18:17 | <Lachy> | do you mean people can do class="axs checkbox" and that works? |
| 18:19 | <zcorpan> | yeah |
| 18:19 | <zcorpan> | or so i was told; i haven't tested the implementation at all |
| 18:20 | <Lachy> | I hope that's not the case. Standardised class names didn't go down well in the HTMLWG |
| 18:20 | <Lachy> | I don't think it would fly if we had to do it with those |
| 18:21 | Lachy | wonders why the accessibility community have been messing around with HTML without consulting the real HTMLWG before implementing disastrous features (assuming what you're saying is true) |
| 18:22 | <Lachy> | (I suspect they've probably consulted the ex-HTMLWG in the past) |
| 18:23 | <zcorpan> | the role module spec is a deliverable of the ex-htmlwg |
| 18:24 | <Lachy> | I know, hence my suspicion |
| 18:24 | <zcorpan> | it seems it went like this... |
| 18:24 | <zcorpan> | 1) role was invented as a native attribute for xhtml2 elements |
| 18:25 | <zcorpan> | 2) we need role in xhtml1, so put the role attribute in the xhtml2 namespace for xhtml1 elements |
| 18:25 | <zcorpan> | 3) make the role attribute a native attribute for xhtml1 (the role module spec) |
| 18:26 | <zcorpan> | 4) the aria specs are a layer on top of role |
| 18:26 | <zcorpan> | 5) shoe-horn it in different ways to make it work in html |
| 18:27 | <zcorpan> | and then somewhere the aria specs note that xhtml 1.0 isn't modular so the role module spec can't be used for xhtml 1.0, so it suggests that the role attribute can be used in the xhtml namespace also |
| 18:28 | <zcorpan> | (how that makes any difference to the "problem" of xhtml 1.0 not being modular is unclear) |
| 18:32 | Lachy | got caught in a netsplit, missed everything you said after point 1). |
| 18:33 | <Lachy> | though it's in the logs, so not a problem |
| 19:30 | <Philip`> | http://philip.html5.org/demos/mathml/texmacs.xhtml - wow, TeXmac's XHTML export works well |
| 19:34 | <Lachy> | Philip`, it's not well formed |
| 19:40 | <Lachy> | here's a well formed version with the correct DOCTYPE http://tinyurl.com/26nyka |
| 19:43 | <zcorpan> | hsivonen: usemap is an idref in your xhtml 1.0 schema; should be url |
| 20:01 | <zcorpan> | Lachy: opera doesn't recognize the dtd, so the entities aren't replaced |
| 20:35 | <Philip`> | Why does MathML prefix all its element names with "m" when they're already in a namespace? |
| 20:37 | <Philip`> | Lachy: I don't quite understand how TeXmacs manages to generate XHTML that is always ill-formed - I thought draconian error handling was meant to result in people fixing obvious mistakes like that when they first tested it and before releasing their code |
| 20:38 | <Lachy> | Philip`, not everyone tests, unfortunately |
| 20:43 | <jgraham> | Lachy: I think I've fixed the bug you reported in the table inspector |
| 20:44 | <Lachy> | jgraham, thanks |
| 20:48 | <Philip`> | http://philip.html5.org/demos/mathml/presentational.xhtml - am I doing something stupid here, to make Firefox draw the cube-roots one line too high and with a black box inside the root sign? |
| 20:49 | <Hixie_> | http://lists.w3.org/Archives/Public/public-forms/2007Sep/0056.html para 2 |
| 20:49 | <Philip`> | (I'm not sure why TeXmacs bothers with the MathML namespace and everything, when it's just outputting HTML tables instead) |
| 20:51 | <jgraham> | Philip`: Have you got all the right fonts installed? Also there (used to be?) problems on Mac. |
| 21:01 | <Philip`> | jgraham: Ah, thanks - it works better when I install various fonts and set something in about:config |
| 21:01 | <Philip`> | Not quite as user-friendly as rendering LaTeX into a PNG, sadly :-( |
| 21:11 | <jacobolus> | Philip`: seems it would be better to just render LaTeX to SVG |
| 21:12 | <jacobolus> | even though the resulting files would likely take more space than pngs, they'd be resolution-independant |
| 21:14 | <karlUshi> | SVG Print |
| 21:14 | <karlUshi> | http://www.w3.org/TR/SVGPrint/ |
| 21:15 | <karlUshi> | This Working Draft defines features of the Scalable Vector Graphics (SVG) Language that are specifically for printing environments. |
| 21:23 | <Philip`> | http://www.ctan.org/tex-archive/systems/win32/bakoma/samples/svgtour.html fails in Opera, which is not very helpful |
| 21:57 | <Hixie_> | jgraham: yt? |
| 21:57 | <Hixie_> | or anyone based in the UK? |
| 21:59 | <jgraham> | Hixie: yep |
| 22:03 | <Dashiva> | I wonder if the group would be more productive if certain threads died out, or if they're simply on top of normal activity |
| 22:32 | <Philip`> | http://philip.html5.org/demos/mathml/svg.xhtml - not incredibly elegant code, but it looks alright |
| 22:44 | <jacobolus> | Philip`: looks pretty good. now all we need is for browsers to embed pdflatex and some pdf→svg converter! |
| 22:45 | <jacobolus> | Philip`: incidentally, isn't it possible to embed the font in the svg, instead of using outlines for everything? |
