/irc-logs / w3c / #html-wg / 2007-12-17 / end

Options:

# Session Start: Mon Dec 17 00:00:00 2007
# Session Ident: #html-wg
# [00:00] * Quits: timbl__ (timbl@209.6.134.246) (Connection reset by peer)
# [00:00] * Joins: timbl (timbl@209.6.134.246)
# [00:19] * Quits: gsnedders (gsnedders@86.135.224.200) (Quit: 404: Not Found)
# [00:19] * Joins: gsnedders (gsnedders@86.135.224.200)
# [00:20] * Joins: Lachy (Lachlan@88.91.106.219)
# [00:20] * Quits: Lachy (Lachlan@88.91.106.219) (Quit: Leaving)
# [00:20] * Joins: csarven (nevrasc@24.202.251.130)
# [00:20] * Joins: Lachy (Lachlan@84.215.41.149)
# [00:28] * Quits: gsnedders (gsnedders@86.135.224.200) (Quit: gsnedders)
# [00:29] * Joins: Facedown (HELLO@68.48.62.208)
# [00:29] * Quits: Sander (svl@86.87.68.167) (Quit: And back he spurred like a madman, shrieking a curse to the sky.)
# [00:29] <Facedown> inimino :)
# [00:36] * Parts: timbl (timbl@209.6.134.246)
# [00:39] * Quits: Facedown (HELLO@68.48.62.208) (Ping timeout)
# [00:40] * Quits: tH (Rob@87.102.85.165) (Quit: ChatZilla 0.9.79-rdmsoft [XULRunner 1.8.0.9/2006120508])
# [00:45] * Joins: Facedown (HELLO@68.48.62.208)
# [00:58] * Quits: hasather (hasather@90.231.107.133) (Quit: leaving)
# [02:01] * Joins: gsnedders (gsnedders@86.135.224.200)
# [02:29] * Joins: shepazu (schepers@128.30.52.30)
# [03:45] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
# [04:50] * Quits: sbuluf (qvrybg@200.49.132.125) (Quit: sbuluf)
# [04:51] * Joins: dbaron (dbaron@71.204.145.103)
# [04:53] * Quits: csarven (nevrasc@24.202.251.130) (Quit: http://www.csarven.ca/)
# [05:23] * Joins: adele (adele@67.170.232.64)
# [05:23] * Quits: adele (adele@67.170.232.64) (Client exited)
# [06:01] * Joins: dean (chatzilla@121.72.34.68)
# [06:01] * Quits: dean (chatzilla@121.72.34.68) (Quit: ChatZilla 0.9.79 [Firefox 2.0.0.11/2007120410])
# [06:01] * Joins: dean5 (opera@121.72.34.68)
# [06:43] * Quits: jmb (jmb@152.78.68.189) (Ping timeout)
# [06:44] * Joins: jmb (jmb@152.78.68.189)
# [07:48] <hsivonen> Philip: I think it isn't reasonable or scalable for me to provide presents for non-standard-namespace data islands
# [07:49] <hsivonen> Philip: however, you could supply your own edited schema
# [07:49] <hsivonen> Philip: or I could add a feature to filter out namespaces listed by the user
# [08:03] * Joins: tH (Rob@87.102.85.165)
# [08:10] <Facedown> gavin = phrogz.net ??!
# [08:39] <MikeSmith> Facedown - I think now.. try /whois gavin
# [08:44] * Quits: shepazu (schepers@128.30.52.30) (Quit: shepazu)
# [09:14] * Joins: shepazu (schepers@128.30.52.30)
# [09:34] * Joins: zcorpan (zcorpan@83.227.33.203)
# [09:42] * Quits: dbaron (dbaron@71.204.145.103) (Ping timeout)
# [09:58] * Joins: Sander (svl@86.87.68.167)
# [09:58] * Quits: zcorpan (zcorpan@83.227.33.203) (Ping timeout)
# [10:02] * Parts: dean5 (opera@121.72.34.68)
# [10:27] * Quits: Lachy (Lachlan@84.215.41.149) (Quit: This computer has gone to sleep)
# [10:43] * Joins: Lachy (Lachlan@213.236.208.22)
# [10:43] * Quits: Lachy (Lachlan@213.236.208.22) (Quit: Leaving)
# [11:01] * Joins: Lachy (Lachlan@213.236.208.22)
# [11:02] * Joins: ROBOd (robod@89.122.216.38)
# [11:04] * Joins: anne-mac (annevk@82.156.27.18)
# [11:04] * anne-mac reads comments on Andy's blog post
# [11:14] * Quits: Sander (svl@86.87.68.167) (Quit: And back he spurred like a madman, shrieking a curse to the sky.)
# [11:25] * anne-mac gets to the part where Maciej explains it all :)
# [11:25] <Lachy> according to my calendar, today is the deadline for DanC to present the list of requirements and binding timetable for publishing HTML5
# [11:26] <anne-mac> I'm pretty sure it's next Thursday
# [11:26] <mjs> I think he agreed to do it by Thursday
# [11:26] <mjs> of this week I think
# [11:26] <mjs> unless I counted wrong
# [11:26] <anne-mac> same here
# [11:26] * Quits: Bert (bbos@128.30.52.28) (Client exited)
# [11:27] <Lachy> hmm, yeah it must be.
# [11:27] <anne-mac> http://alex.dojotoolkit.org/?p=642 :(
# [11:27] <Lachy> I wonder why it got marked on the wrong date in my iCal
# [11:28] <mjs> I don't entirely disagree with Alex Russel's point
# [11:28] <mjs> we need a balance between innovation and consolidation through standards
# [11:28] <mjs> I think SafOperFox are striking a reasonably good balance
# [11:29] <mjs> Alex cited our animation stuff, but 3D canvas is pretty cool as well
# [11:29] <mjs> and some things are better done through standards in the first place, like <video>
# [11:30] <anne-mac> <canvas> is another example
# [11:30] <anne-mac> I wasn't sure he meant that
# [11:31] <mjs> well, <canvas> was originally a renegade nonstandard thing from Safari
# [11:31] <anne-mac> I think the current model for HTML5 features is at least quite good, early feedback from vendors about new features, and an editor to make it consistent with the rest
# [11:31] <anne-mac> fair enough
# [11:31] <mjs> although our experience with it makes us want to go to standards earlier in the process now
# [11:32] <anne-mac> with CSS there's no editor yet :(
# [11:32] <mjs> CSS doesn't really have its house in order enough
# [11:32] <mjs> for things to work the same way
# [11:35] <mjs> the problem with the HTML process right now is that we're blocked on publishing
# [11:49] <hsivonen> hmm. my W3C list mailbox in Mail.app is suddenly empty. I wonder what happened
# [11:52] <hsivonen> hmm. a slip of a mouse... it's *way* too easy to move an entire mailbox in Mail.app
# [11:53] * Joins: anne (annevk@82.156.27.18)
# [11:59] * Quits: anne-mac (annevk@82.156.27.18) (Ping timeout)
# [12:08] * Quits: ROBOd (robod@89.122.216.38) (Quit: http://www.robodesign.ro )
# [12:11] * Joins: hasather (hasather@90.231.107.133)
# [12:20] * Joins: akaroa (opera@121.72.6.99)
# [12:30] * Joins: myakura (myakura@123.225.61.58)
# [13:35] * Quits: myakura (myakura@123.225.61.58) (Quit: Leaving...)
# [13:36] * Joins: matt (matt@128.30.52.30)
# [13:40] <Philip> hsivonen: Why does http://validator.nu/?doc=http%3A%2F%2Fphilip.html5.org%2F404&schema=http%3A%2F%2Fwww.web3d.org%2Fspecifications%2Fx3d-3.1.xsd&parser=xml&laxtype=yes&showsource=yes say an HTML page validates according to the XML schema?
# [13:40] * Quits: akaroa (opera@121.72.6.99) (Ping timeout)
# [13:44] * Joins: zcorpan (zcorpan@88.131.66.80)
# [13:46] * Philip guesses it's because XSD is not supported, but it shouldn't pretend to succeed in that case
# [13:48] <hsivonen> Philip: hmm. I have no idea why that happens. I'll look into it.
# [13:49] <anne> http://www.b-list.org/weblog/2007/dec/17/standards/
# [13:49] <anne> http://www.kryogenix.org/days/2007/12/17/reigniting-the-browser-wars
# [13:49] <Philip> Incidentally, it might be nice if XSD was supported, if it isn't already and if that's easy to do :-)
# [13:52] * Joins: akaroa (opera@121.72.6.99)
# [13:54] <hsivonen> Philip: the main reason why I disabled (or more likely, thought I disabled) XSD support was that I couldn't figure out if it opened the system to security problems
# [13:55] * Joins: aaronlev (chatzilla@209.6.168.245)
# [13:56] <hsivonen> Philip: anyway, I'm willing to support the Xerces-J 2 notion of XSD if someone helps me by supplying test case schemas that use whatever inclusion/file reading points XSD has so that I can see how those file inclusions are handled
# [13:56] <hsivonen> Philip: basically, I want to be sure that Apache doesn't do any file: URL IO on its own
# [13:57] <hsivonen> Philip: and that all IO goes through HttpClient
# [13:57] <hsivonen> I don't want JDK http client to be involved, either
# [13:57] <hsivonen> I'm not willing to support xsi:schemaLocation, though
# [14:04] <hsivonen> Philip: oh yeah, I had another reason to disable XSD as well: Back when I last tested (over a year ago), then-current Xerces had a NullPointerException bug somewhere in its bowels and it was triggered by a W3C-supplied schema I used as a test case
# [14:07] * Joins: ROBOd (robod@89.122.216.38)
# [14:26] * Joins: timbl (timbl@209.6.134.246)
# [14:53] * Quits: timbl (timbl@209.6.134.246) (Connection reset by peer)
# [14:53] * Joins: timbl (timbl@209.6.134.246)
# [15:06] <Philip> hsivonen: Okay, that sounds reasonable
# [15:06] * Philip knows pretty much nothing about XSD, and so probably can't be of much help now
# [15:09] <hsivonen> Philip: fwiw, re-enabling XSD is a one-character change and recompile, so if you've already got a copy of the source tree, the line to uncomment is #com.thaiopensource.validate.xerces.XsdSchemaReaderFactory in onvdl/src/META-INF/services/com.thaiopensource.validate.SchemaReaderFactory
# [15:13] * Philip doesn't have a copy of the source tree
# [15:13] * Philip wonders if it's trivial to install
# [15:14] * Philip tries the easy-looking installation commands
# [15:14] * Joins: smedero (smedero@158.130.16.191)
# [15:24] <Philip> "Bad MD5 hash for http://mirror.eunet.fi/apache/xml/xalan-j/xalan-j_2_7_0-bin.zip"
# [15:25] <hsivonen> Philip: hmm. the apache mirrors are annoyingly unreliable. I should add a check for 4xx responses
# [15:25] <Philip> Hmm, I can't easily delete the file and try to download it again, because I've got no idea where the downloader saves it
# [15:26] <hsivonen> Philip: can you see if the downloaded file looks like a zip or like an HTML error page?
# [15:26] <hsivonen> in dependencies/
# [15:27] <Philip> There's no dependencies/xalan* - maybe it didn't save it anywhere. I get a 404 if I try that address in a web browser
# [15:27] <Philip> Looks like they only have 2.7.1 now
# [15:30] <hsivonen> it's annoying that the mirrors don't keep old versions
# [15:30] <hsivonen> I just discovered that the Inkscape docs do not completely describe Inkscape's SVG extensions
# [15:31] <Philip> I guess it's sensible to delete the old versions if the new version fixes security vulnerabilities, though I have no idea if that's relevant in this case
# [15:32] <hsivonen> I'm starting to wonder if trying to build specific support for Inkscrape extensions is a bad idea after all
# [15:32] <hsivonen> everything is harder than it first appears...
# [15:33] <hsivonen> haha. the unknown attribute that I find is documented to be INTERNAL USE ONLY; never serialized
# [15:33] <hsivonen> yeah, right
# [15:34] <Philip> I'm not sure who would find specific Inkscape support helpful - I would just want the validator to accept documents I exported from Inkscape and maybe modified by hand a little, and I wouldn't care about any details in the Inkscape-specific sections
# [15:35] <hsivonen> Philip: I'm not checking the values of Inkscape attributes or anything, but I made a whitelist instead of allowing anything
# [15:36] <hsivonen> s/anything/everything/ in the last instance
# [15:39] <Philip> A whitelist allowing inkscape:* and sodipodi:* would allow the things I'd want it to allow (i.e. documents exported from past/current/future versions of Inkscape), and wouldn't miss any errors that I would make and would want to be warned about (because I'm not going to be writing/modifying things in Inkscape's namespaces)
# [15:40] <hsivonen> Philip: would you allow those on XHTML elements and MathML, too?
# [15:40] <hsivonen> Philip: would you allow <sodipodi:namedview/> outside <defs>?
# [15:41] <hsivonen> (the way I'm rationalizing this effort is that that inkscape cruft seems to be a pretty common issue)
# [15:44] <Philip> I guess I don't see why it would hurt to allow them, since interoperability is irrelevant for product-specific extensions like these... That sounds like a potentially dangerous view, though
# [15:45] <hsivonen> aside: IDness sucks. Chemical Markup Language is smart to avoid IDness.
# [15:45] * Quits: aaronlev (chatzilla@209.6.168.245) (Ping timeout)
# [15:45] <hsivonen> I think it's a bad idea, though, that CML's non-ID is called id
# [15:45] <hsivonen> so inkscape doesn't use globally unique IDs
# [15:46] <hsivonen> so when a single XHTML page includes two Inkscape drawings, their IDs will collide
# [15:46] <hsivonen> not cool
# [15:46] <hsivonen> but I guess I should report that as an error, given specs and all
# [15:47] <Philip> I can't think of a good reason to accept Inkscape-specific elements/attributes scattered through a document, but not accept e.g. MS Office's HTML-with-namespaces exporter's elements/attributes scattered through a document
# [15:48] <anne> i think accepting either is potentially harmful
# [15:48] <Philip> but the latter sounds like something that shouldn't be accepted for documents published on the web, so I'm not sure why the former would be accepted
# [15:50] <hsivonen> anne: I don't have an exactly good feeling about allowing the Inkscape cruft by default, but is it really useful to whine to authors about it?
# [15:50] <hsivonen> anne: what's the harm?
# [15:51] <anne> the SVG would do different things in inkscape versus other editors
# [15:51] <Philip> hsivonen: Is there a standard way to combine documents with potentially-conflicting IDs?
# [15:51] <hsivonen> Philip: not to my knowledge
# [15:51] <anne> which might force other editors to support their stuff
# [15:52] <hsivonen> anne: would it be bad if the Inkscape stuff became the de facto way of storing certain editing state in SVG?
# [15:52] * Philip has had roughly that issue with X3D-plus-some-extensions, e.g. importing a file with some meshes and then referring to an ID in that file in order to animate it
# [15:52] <anne> about as bad as <apple-canvas> becoming the de facto way of doing <canvas> on the Web
# [15:53] <hsivonen> anne: I see your point
# [15:54] <hsivonen> anne: still, it seems user-unfriendly to pretend that commonly-used extensions didn't exist
# [15:54] <Philip> (Currently I've got something like <Inline DEF="a" url="some file with an element 'b'"/> and then you can use an almost-IDREF like "a b" to access the nested elements without conflicts)
# [15:57] <hsivonen> anne: what's your take on http://intertwingly.net/blog/2007/11/16/Whitelisting#c1195315674 ?
# [15:58] <anne> maybe you could make some widely used stuff a stron warning
# [15:58] * Quits: timbl (timbl@209.6.134.246) (Connection reset by peer)
# [15:59] * anne tries to fetch the page, should actually go
# [15:59] <hsivonen> and, yeah, if MS Office exported XHTML5 + namespaced extensions, I guess I should seriously consider supporting it on some level that keeps the validator useful for Office users
# [15:59] * Joins: timbl (timbl@209.6.134.246)
# [16:00] <anne> I don't think vendor-extensions are that great (long term) and all other types of extensions really should be coordinated (although it should be easy to get them in the spec if people agree it's a good thing) as we see with SQL storage, etc.
# [16:01] <anne> I have to go now
# [16:02] * Quits: matt (matt@128.30.52.30) (Quit: matt)
# [16:05] <hsivonen> anne: (writing to log in case you read it when you come back): I agree that extensions should be coordinated and done without vendor namespaces ideally. However, Inkscape is using the extension mechanism that the SVG spec blesses
# [16:06] * Parts: akaroa (opera@121.72.6.99)
# [16:06] * Joins: akaroa (opera@121.72.6.99)
# [16:06] * Parts: akaroa (opera@121.72.6.99)
# [16:09] * Joins: akaroa (opera@121.72.6.99)
# [16:20] <Philip> hsivonen: I get various messages like "./syntax/non-schema/java/src/org/whattf/checker/TextContentChecker.java:27: package org.relaxng.datatype does not exist" that look like missing dependencies
# [16:22] <Philip> (after modifying build.py to get xalan-j_2_7_1 instead of xalan-j_2_7_0)
# [16:23] <hsivonen> Philip: that's weird. did onvdl compile successfully?
# [16:23] <hsivonen> that is, is there a file at jars/onvdl-whattf.jar ?
# [16:24] * Quits: jgraham_ (james@81.86.217.3) (Quit: This computer has gone to sleep)
# [16:24] <Philip> There is a file there
# [16:24] <Philip> If I run 'build' again, the first reported error is "./onvdl/src/nu/validator/relaxng/exceptions/UnfinishedElementException.jav./onvdl/src/com/thaiopensource/validate/SchemaReaderLoader.java:7: package org.apache.xalan.processor does not exist"
# [16:25] <Philip> which probably makes sense
# [16:25] <Philip> because I manually downloaded xalan-j_2_7_1-bin.zip and it probably expects that file to be unpacked
# [16:26] <Philip> Looks a bit happier now
# [16:26] * Quits: timbl (timbl@209.6.134.246) (Connection reset by peer)
# [16:26] * Joins: timbl (timbl@209.6.134.246)
# [16:26] <Philip> (I don't know whether you can easily check that the dependency got unpacked properly)
# [16:26] * Joins: aaronlev (chatzilla@66.30.196.151)
# [16:26] <hsivonen> Philip: the build script has a Python array of dependency jars.
# [16:27] <hsivonen> Does anyone want to suggest wording for a warning about product-specific extensions?
# [16:28] * Joins: jgraham_ (james@81.86.217.3)
# [16:28] <Philip> hsivonen: It seems ro run successfully now
# [16:29] <hsivonen> Also, can someone suggest a practical content model for foreignObject in the XHTML+SVG+MathML case?
# [16:29] <hsivonen> Philip: nice
# [16:29] <Philip> s/ro/to/
# [16:29] * Joins: billmason (billmason@69.30.57.156)
# [16:30] <Philip> org.xml.sax.SAXParseException: No implementation available for schema language with namespace URI “http://www.w3.org/2001/XMLSchema”
# [16:30] <Philip> ...is what it says by default when trying to use XSD
# [16:31] <Philip> (before telling the user "The document validates according to the specified schema(s).")
# [16:31] <zcorpan> hsivonen: HTML <html> or <div>, or MathML <math>
# [16:31] <zcorpan> (though HTML5 currently says only <html> would be allowed, i think)
# [16:32] * zcorpan thinks <div> should be allowed as root
# [16:32] * zcorpan notes that Atom uses <div>
# [16:32] <gsnedders> zcorpan: yeah, currently only <html>
# [16:35] <hsivonen> zcorpan: is needing a <div> there a cargo cult of do SVG impls break if foreignObject has more than one XHTML block element as children?
# [16:37] <Philip> hsivonen: That thingy for enabling XSD seems to work correctly
# [16:37] <hsivonen> Philip: nice.
# [16:37] <Philip> though its error messages are totally useless
# [16:37] <hsivonen> Philip: then what's left is making sure that it uses the right EntityResolver
# [16:37] <hsivonen> Philip: such as?
# [16:38] <Philip> Error: http://www.w3.org/TR/xml-schema-1#cvc-complex-type.3.2.2?X3D&width
# [16:38] <Philip> From line 1, column 1; to line 5, column 52
# [16:38] <Philip> (is what it says for <X3D width>, given that width is not an allowed attribute)
# [16:39] <zcorpan> hsivonen: haven't tested multiple children
# [16:39] <hsivonen> Philip: weren't URI-identified errors supposed to be coolest thing around? :-)
# [16:39] <hsivonen> zcorpan: ok
# [16:39] <zcorpan> hsivonen: i'd be fine with other elements as root as well
# [16:40] <zcorpan> (or multiple roots for that matter)
# [16:40] <Philip> hsivonen: They tempt me to copy the URI into my address bar, and then taunt me by giving a 404
# [16:40] * Parts: akaroa (opera@121.72.6.99)
# [16:40] <hsivonen> zcorpan: the SVG 1.1 spec uses <body>...
# [16:41] <zcorpan> hsivonen: i see little reason to disallow it :)
# [16:42] <zcorpan> except that it makes the "contexts in which this element is allowed" more complex
# [16:42] <zcorpan> allowing any block or <html> as root seems like a simpler rule
# [16:42] * Philip now needs to learn enough about XML Schema to move all the elements into a namespace and add some extensions
# [16:43] <hsivonen> Philip: Jing/oNVDL doesn't interfere with those messages. There's only Xerces to blame
# [16:43] * Joins: matt (matt@128.30.52.30)
# [16:43] * hsivonen wonders if there's a human-friendly error renderer in Xerces somewhere
# [16:43] <Philip> hsivonen: I've used Xerces-C++, so I'm happy to blame it for many things :-)
# [16:46] * Quits: matt (matt@128.30.52.30) (Quit: matt)
# [16:46] <hsivonen> hmm. the W3C SVG 1.1 schema allows non-whitespace text node children in foreignObject. isn't that bogus?
# [16:46] <hsivonen> also, it allows <svg> there. isn't that bogus too as non-foreign?
# [16:46] * Quits: timbl (timbl@209.6.134.246) (Connection reset by peer)
# [16:47] * Quits: Lachy (Lachlan@213.236.208.22) (Quit: This computer has gone to sleep)
# [16:49] <hsivonen> is it just me or is foreignObject seriously underdefined in SVG 1.1?
# [16:50] <hsivonen> what's the deal with SVG 1.2 Full not having a public WD since 13 April 2005?
# [16:50] <hsivonen> I thought SVG 1.1 Full was about to go to CR
# [16:52] <hsivonen> s/1.1/1.2/
# [16:54] <zcorpan> hsivonen: i thought they worked on 1.2 Tiny
# [16:54] <hsivonen> zcorpan: Tiny is a REC
# [16:55] <zcorpan> hsivonen: oh? http://www.w3.org/TR/SVGMobile12/ says "Candidate Recommendation" for me
# [16:55] <hsivonen> zcorpan: oh right. I misread
# [16:55] * hsivonen saw the capital R and forgot to read more to the left
# [16:55] * Philip removes the code that disallows connections to localhost
# [16:58] <Philip> How are you meant to use IPv6 host addresses in HTTP URIs? http://::1/ is going to get confused with the port number syntax
# [16:59] <hsivonen> Philip: wrap in []
# [16:59] <hsivonen> http://[::1]/
# [16:59] <hsivonen> or something like that
# [17:00] <Philip> hsivonen: Ah, right
# [17:01] <Philip> hsivonen: Does that mean http://validator.nu/?doc=http%3A%2F%2F%5B%3A%3A1%5D%2F should be forbidden?
# [17:01] <hsivonen> Philip: probably.
# [17:02] <hsivonen> Philip: the localhost banning code is a broken tinfoil hat
# [17:02] <hsivonen> I tried to google for foreignObject demos but they are remarkably few in the top hits
# [17:04] <Philip> (Wow, IPv6 actually works on my local computer)
# [17:04] <hsivonen> Philip: do you get routing to the outside world as well?
# [17:05] <Philip> hsivonen: No
# [17:05] <Philip> $ ping6 ietf.org
# [17:05] <Philip> connect: Network is unreachable
# [17:05] <Philip> :-(
# [17:07] <hsivonen> if <foreignObject><div> is only a cargo cult, it is a strong one
# [17:07] <hsivonen> I don't like this. I wanted to see a proper spec for how foreignObject establishes a nested CSS rendering context
# [17:10] <Philip> (Hmm, it looks like the university network supports IPv6, but the CS department's network doesn't)
# [17:10] <Philip> (which seems a bit odd)
# [17:10] <Philip> (given that there are people here who do networky stuff)
# [17:11] <hsivonen> Jacques Distler has a use case for SVG in <annotation-xml>: http://golem.ph.utexas.edu/~distler/blog/archives/001475.html
# [17:16] <Philip> <xs:schema targetNamespace="http://www.web3d.org/specifications/x3d-namespace" xmlns="http://www.web3d.org/specifications/x3d-namespace" ...>
# [17:16] <Philip> That was unexpectedly straightforward
# [17:19] * Joins: matt (matt@128.30.52.30)
# [17:19] * Quits: hsivonen (hsivonen@130.233.41.50) (Ping timeout)
# [17:19] * Joins: hsivonen (hsivonen@130.233.41.50)
# [17:25] * Joins: Lachy (Lachlan@84.215.41.149)
# [17:46] * Joins: timbl_ (timbl@209.6.134.246)
# [17:53] <anne> I think <foreignObject> should allow anything <body> allows
# [17:53] <anne> I don't think allowing <html> makes sense personally
# [17:55] <hsivonen> anne: does allowing anything that body allows work in browsers without a wrapper body or div?
# [17:55] <hsivonen> anne: what about <math>?
# [17:56] <anne> if it doesn't that would be buggy
# [17:56] <zcorpan> mathml elements need a <math> ancestor to work correctly in firefox, iirc
# [17:56] <hsivonen> zcorpan: but
# [17:56] <anne> makes sense
# [17:57] <hsivonen> should <math> be allowed as a child of foreignObject without a <p> wrapper?
# [17:57] <hsivonen> did CDF ever spec this stuff?
# [17:57] <anne> no
# [17:58] <anne> they haven't gotten this far yet
# [17:58] * Quits: aaronlev (chatzilla@66.30.196.151) (Ping timeout)
# [17:59] <zcorpan> hsivonen: it should
# [17:59] <hsivonen> anne: mmkay. I would have thought this was what CDF was supposed to spec since 2004 or so
# [18:00] <anne> cdf focused on <object> instead
# [18:00] <zcorpan> why would you want to have html in between? it's verbose as it is to embed mathml in svg :)
# [18:00] <hsivonen> zcorpan: <math> is semantically inline
# [18:01] <zcorpan> hsivonen: ok. then i'd say allow inline html in foreignObject as well :)
# [18:01] <anne> yeah, though it's already pretty unclear what "semantically" means in terms of content inside an image
# [18:02] * anne thinks <body> should allow either inline or block
# [18:02] * anne hopes that solves it
# [18:02] <zcorpan> the common case for mathml in svg is probably going to be short labels
# [18:03] <hsivonen> anne: would you allow a wrapper <body>, though, like zcorpan and SVG 1.1 suggest?
# [18:04] <anne> probably not
# [18:04] <anne> but i guess there's no real harm if it's there anyway...
# [18:07] <Philip> "Error: http://www.w3.org/TR/xml-schema-1#src-redefine.6.2.2?ChildContentModelInteractive&rcase-RecurseLax.2" - I wonder if that's bad
# [18:08] <hsivonen> And here I thought the Jing RELAX NG messages were bad
# [18:09] <Philip> I'm not sure why my lax is recursive
# [18:09] <anne> why support XSD?
# [18:11] <zcorpan> just tested multiple roots in foreignObject in opera, worked fine
# [18:11] <zcorpan> also firefox
# [18:11] <Philip> anne: I want to validate X3D content, and X3D provides XSD and DTD, and everyone hates DTD so I want to make the XSD work :-)
# [18:14] <anne> just convert the XSD to a RelaxNG
# [18:14] <Philip> Is that trivial?
# [18:14] <hsivonen> XSD is relatively hated, too
# [18:15] <anne> Philip, in my XML days I had an XML editor that did it out of the box
# [18:15] * anne forgot the name
# [18:16] * zcorpan finds http://paste.lisp.org/display/21388 which uses <foreignObject><xhtml:img/>
# [18:17] <hsivonen> Philip: a converter is supposedly available from https://msv.dev.java.net/ but the server is misbehaving at the moment
# [18:19] <hsivonen> Philip: https://msv.dev.java.net/files/documents/61/31333/rngconv.20060319.zip
# [18:20] <hsivonen> though the truthiness over at wikipedia is that XSD misfeatures are "advantages over RELAX NG"
# [18:20] <hsivonen> as if any feature bullet point were an advantage
# [18:21] <Philip> rngconv says "warning: unsupported whiteSpace facet is ignored" a lot
# [18:22] <hsivonen> http://en.wikipedia.org/wiki/XML_Schema_Language_Comparison#Advantages_over_RELAX_NG
# [18:22] * Philip wonders how to try validating with RNG
# [18:22] <Philip> Oh, I forgot I've got validator.nu
# [18:23] <Philip> "Error: Both operands of interleave contain text. At line 5680, column 35 in resource http://localhost/.../x3d-3.1.rng"
# [18:24] * Joins: dbaron (dbaron@71.204.145.103)
# [18:24] <Philip> Oddly, there is no element named "interleave"
# [18:25] <hsivonen> interleave has some mighty annoying restrictions
# [18:25] * Quits: billmason (billmason@69.30.57.156) (Connection reset by peer)
# [18:27] * Joins: billmason (billmason@69.30.57.156)
# [18:27] <Philip> If I delete the offending element, I get ten errors like "Conflicting ID-types for attribute USE of element GeoOrigin."
# [18:29] <hsivonen> Yeah, that's very annoying. I've considered turning DTD compat off, but I left it on in case someone actually wants it
# [18:29] * Philip thinks this is not entirely trivial, and it seems doubtful that it will match precisely the same documents as the original Schema
# [18:29] <hsivonen> moral of the story, if you don't want DTD compat, say xsd:NCName instead of xsd:ID
# [18:30] <Philip> Does that mean it can't automatically detect ID/IDREF consistency?
# [18:31] <hsivonen> if you use xsd:NCName, then, no
# [18:31] <hsivonen> I left DTD Compat on exactly in case someone actually wanted to do ID/IDREF checking in RELAX NG despite it sucking horribly
# [18:32] <zcorpan> why does it suck?
# [18:32] <Philip> Does xsd:IDREFS map onto something trivial?
# [18:32] <hsivonen> Philip: list { xsd:NCName+ } or something like that. I forget the syntax details
# [18:33] <Philip> "Error: Bad value 0 0 0 for attribute centerOfRotation on element Viewpoint from namespace http://www.web3d.org/specifications/x3d-namespace."
# [18:33] <Philip> Hmm...
# [18:33] <hsivonen> zcorpan: it has the annoying restriction Philip is seeing
# [18:33] <hsivonen> zcorpan: and it doesn't check that an IDREF points to an element of the right type
# [18:34] <hsivonen> zcorpan: the annoying restriction being that the IDness is applied to an element,attribute pair
# [18:34] <hsivonen> zcorpan: instead of the IDness getting applied to a particular attribute instance when the schema otherwise matches
# [18:34] <Philip> attribute centerOfRotation
# [18:34] <Philip> {
# [18:34] <Philip> xsd:string
# [18:34] <Philip> {
# [18:34] <Philip> pattern = """((\+|\-)?(0|[1-9][0-9]*)?(\.[0-9]*)?((E|e)(\+|\-)?[0-9]+)?
# [18:34] <Philip> (\+|\-)?(0|[1-9][0-9]*)?(\.[0-9]*)?((E|e)(\+|\-)?[0-9]+)? (\+|\-)?(0|[1-9][0-9]*)?(\.[0-9]*)?((E|e)(\+|\-)?[0-9]+)?)?"""
# [18:35] <Philip> }
# [18:35] <Philip> }?
# [18:35] <Philip> Yay, regexps
# [18:35] <hsivonen> the XSD regexps are peculiar
# [18:35] <hsivonen> for example, \s is mostly useless
# [18:35] <Philip> Looks like line-wrapping in the XSD->RNG converter messed up the whitespace
# [18:35] <hsivonen> because it matches anything in Zs
# [18:36] <hsivonen> IIRC
# [18:36] <Philip> <xs:restriction base="xs:string">
# [18:36] <Philip> <xs:whiteSpace value="collapse"/>
# [18:36] <Philip> <xs:pattern value="((\+|\-)?(0|[1-9][0-9]*)?(\.[0-9]*)?((E|e)(\+|\-)?[0-9]+)? (\+|\-)?(0|[1-9][0-9]*)?(\.[0-9]*)?((E|e)(\+|\-)?[0-9]+)? (\+|\-)?(0|[1-9][0-9]*)?(\.[0-9]*)?((E|e)(\+|\-)?[0-9]+)?)?"/>
# [18:36] <Philip> </xs:restriction>
# [18:36] <Philip> is what the XSD says
# [18:36] <Philip> which I guess is what "warning: unsupported whiteSpace facet is ignored" means
# [18:37] * Philip is unsure of the least painful solution
# [18:38] <Philip> Annoyingly, these patterns are repeated for every attribute in the RNG file
# [18:38] <Philip> (whereas it was a single <xs:simpleType name="SFVec3f"> originally)
# [18:40] <zcorpan> hsivonen: ok
# [18:48] * Joins: aaronlev (chatzilla@209.6.168.245)
# [18:49] * Philip goes back to the Schema because that seems less likely to have so many small bugs
# [18:49] <Philip> "Error: src-resolve: Cannot resolve the name 'ChildContentModelInteractive_fn3dktizrknc9pi' to a(n) 'type definition' component." - so, uh, where did it get that fn3dktizrknc9pi from?
# [18:52] <hsivonen> Philip: unlike Trang, the Sun converter isn't designed to retain the documentary human-perceived structure of the schema :-(
# [18:52] <hsivonen> and Trang doesn't support XSD as an input format
# [18:55] * Philip copies and pastes bits of XSD found on Google until the errors go away
# [18:58] <Philip> Ooh, it works
# [19:05] <Philip> but I need a Schema for XSLT
# [19:08] * Philip uses the XSLT 2.0 one, because it exists, even though it's probably not what web browsers support
# [19:09] <hsivonen> Hixie: do you have a current wild guess about what the XHTML5-in-SVG integration point in foreignObject should look like?
# [19:10] * Quits: timbl_ (timbl@209.6.134.246) (Quit: timbl_)
# [19:12] <Philip> Hooray, now the validator says my document is valid, except for a few bits which are not valid and need to be redesigned
# [19:12] <Philip> I just hope it still says invalid documents are invalid
# [19:13] * Joins: hober (ted@68.101.220.172)
# [19:42] <hsivonen> I'm pretty sure naming a cat "Astrophy" is an intertextual reference of some kind
# [19:42] <hsivonen> Hixie: what's the story of the Astrophy cat?
# [19:42] <hsivonen> reference to an old TV show combined with astrophysics somehow?
# [19:46] <gsnedders> hsivonen: Astrophy? where?
# [19:47] <hsivonen> gsnedders: http://www.hixie.ch/resources/images/astrophy/original
# [19:51] * Quits: aaronlev (chatzilla@209.6.168.245) (Ping timeout)
# [19:51] * Joins: tH_ (Rob@87.102.85.140)
# [19:51] * Quits: tH (Rob@87.102.85.165) (Ping timeout)
# [19:51] * tH_ is now known as tH
# [19:55] * zcorpan came to think of a new name for void elements
# [19:55] <zcorpan> leaf elements
# [19:55] <zcorpan> leafs in a tree don't have children
# [19:56] * zcorpan read "(A node with no children is sometimes called a leaf.)" in http://www.hixie.ch/commentary/astrophy-interface , googling for astrophy
# [19:56] <hsivonen> zcorpan: void elements is a parsing algorithm term
# [19:56] <hsivonen> leaf is a tree term
# [19:57] <zcorpan> hsivonen: #writing talks about void elements
# [19:59] <zcorpan> http://www.whatwg.org/specs/web-apps/current-work/multipage/section-writing0.html#elements0
# [19:59] <zcorpan> however, leaf probably has the same problem as "empty"
# [19:59] <hsivonen> it does
# [20:01] <hsivonen> anne: btw, I was in a bar the other day and a guy asked me if XHR access-control is vulnerable to DNS spoofing
# [20:01] <hsivonen> anne: is it?
# [20:06] * Joins: timbl (timbl@128.30.5.98)
# [20:06] <anne> example?
# [20:07] <hsivonen> anne: site A trusts site B to read its data, but site B is DNS-spoofed and is actually site C
# [20:07] <hsivonen> anne: data leaks form A to C
# [20:10] * Parts: zcorpan (zcorpan@88.131.66.80)
# [20:12] * Joins: aaronlev (chatzilla@66.30.196.151)
# [20:19] <anne> hsivonen, yeah, you'd have to use TLS or something to prevent that from happening
# [20:29] <hsivonen> hmm. I wonder if XHTML 1.0 + SVG 1.1 + MathML 2.0 should allow <svg> and <math> as blocks. that seems bogus to me
# [20:30] * Joins: adele (adele@17.255.110.63)
# [20:32] <shepazu> why?
# [20:32] <shepazu> or I guess I mean, what do you mean as "blocks"
# [20:32] <shepazu> ?
# [20:33] <hsivonen> shepazu: I mean same level as <p> and <ol>
# [20:34] <hsivonen> shepazu: <svg> is analogous to <img>
# [20:34] <shepazu> where do you think they should be?
# [20:34] <hsivonen> shepazu: and <math> is semantically inline
# [20:34] <shepazu> isn't <img> allowed at the same level as <p> and <ol>?
# [20:35] <hsivonen> shepazu: not in Strict
# [20:35] <shepazu> huh.... how do they differ?
# [20:36] <hsivonen> shepazu: "they"?
# [20:36] <shepazu> <img> and <p>, in terms of where they can be located?
# [20:38] <hsivonen> shepazu: <img> is inline and <p> is block
# [20:39] <hsivonen> shepazu: so <p> can appear as child of <body>
# [20:39] <hsivonen> shepazu: but <img> can't
# [20:39] <hsivonen> (or it *can* but it would be invalid Strict)
# [20:39] <shepazu> hmm
# [20:40] <shepazu> well, in general, I'd agree that <svg> should follow the model of <img>
# [20:41] <shepazu> but I don't understand the rationale for the different location models... how does this help authors or browsers?
# [20:41] <hsivonen> shepazu: I'm not sure, but presumably the definers of HTML 4 felt is was important
# [20:43] <shepazu> is it useful to keep that distinction?
# [20:43] <hsivonen> shepazu: probably not
# [20:44] <hsivonen> shepazu: there's a Google spreadsheet about this
# [20:44] <shepazu> hsivonen:
# [20:44] <shepazu> hsivonen: I'm sure there is :)
# [20:44] <shepazu> (sorry, still getting used to my mac IRC client)
# [20:44] <hsivonen> shepazu: http://spreadsheets.google.com/ccc?key=pkNVM1HEQs-wsHB7s1M5Lbw
# [20:47] <hsivonen> shepazu: (if you have a Google id, I can add you to the user list)
# [20:48] <shepazu> hsivonen: I do, but I'll have to find it :)
# [20:48] <hsivonen> Hixie: the spreadsheet should probably be made world-readable
# [20:50] <smedero> weee, the spreadsheet crashed Safari 3.
# [20:50] * Joins: Sander (svl@86.87.68.167)
# [20:51] * Quits: Lachy (Lachlan@84.215.41.149) (Ping timeout)
# [20:52] <hsivonen> shepazu: the protocol for disputing a containment decision is putting a question mark after yes/no
# [20:52] * Quits: timbl (timbl@128.30.5.98) (Connection reset by peer)
# [20:52] <smedero> not reproducible though, 2nd time it loaded fine.
# [20:52] <shepazu> ok
# [20:52] * Joins: timbl (timbl@128.30.5.98)
# [20:55] * Quits: smedero (smedero@158.130.16.191) (Quit: smedero)
# [20:55] <hsivonen> James Clark's XHTML 1.0 schema and the W3C SVG 1.1 schema have conflicting production names :-(
# [20:58] * Joins: Lachy (Lachlan@88.91.106.219)
# [21:07] * Quits: dbaron (dbaron@71.204.145.103) (Quit: 8403864 bytes have been tenured, next gc will be global.)
# [21:10] * Quits: jmb (jmb@152.78.68.189) (Ping timeout)
# [21:11] * Joins: jmb (jmb@152.78.68.189)
# [21:31] * Quits: hober (ted@68.101.220.172) (Quit: ERC Version 5.3 (devel) (IRC client for Emacs))
# [21:37] * Quits: trackbot-ng (trackbot-n@128.30.52.30) (Client exited)
# [21:38] * Joins: zcorpan (zcorpan@83.227.33.203)
# [21:39] <Hixie> hsivonen: it is
# [21:41] <hsivonen> Hixie: the spreadsheet world-readable?
# [21:42] <hsivonen> Hixie: the publish button tells me it has not been published
# [21:42] <hsivonen> Hixie: and at least it does seem to require login
# [21:44] <Hixie> i checked the box on the sharing page that says "Anyone can view this document at:
# [21:44] <Hixie> http://spreadsheets.google.com/ccc?key=pkNVM1HEQs-wsHB7s1M5Lbw"
# [21:44] <Hixie> oh i guess i need to publish as well
# [21:44] <Hixie> huh
# [21:45] <Hixie> done
# [21:45] <hsivonen> Hixie: ah. now it says it is public, yes
# [21:45] <hsivonen> Hixie: thanks
# [21:45] <hsivonen> Hixie: did you see the SVG foreignObject content model question?
# [21:46] * Joins: trackbot-ng (trackbot-n@128.30.52.30)
# [21:46] * trackbot-ng is loading HTML Issue Tracking data...
# [21:46] * trackbot-ng found 16 users
# [21:46] <trackbot-ng> Tracking ISSUEs and ACTIONs from http://www.w3.org/html/wg/tracker/
# [21:50] * Quits: trackbot-ng (trackbot-n@128.30.52.30) (Client exited)
# [21:51] <Hixie> hsivonen: *looks*
# [21:51] <jgraham_> hsivonen: I think <math> makes sense as a block-level element
# [21:51] <Hixie> hsivonen: i'll think about it
# [21:52] <Hixie> hsivonen: note that on the block vs inline stuff, my current plan is to remove the distinction.
# [21:52] <hsivonen> jgraham_: in what context? dislay math (LaTeX $$...$$) is semantically inline (can be inline a sentence) even if display:block
# [21:52] <gsnedders> in LaTeX there are both block and inline maths content. I think it is sensible to have both.
# [21:52] <jgraham_> I was thinking of LaTeX also
# [21:53] <hsivonen> what's block math in LaTeX?
# [21:53] * gsnedders doesn't know the syntax off the top of his head
# [21:53] <Philip> $...$
# [21:53] <jgraham_> \begin{equation} \end{equation}
# [21:53] <Philip> I think
# [21:53] * Quits: Lachy (Lachlan@88.91.106.219) (Quit: Leaving)
# [21:54] <jgraham_> $ $ is inline math
# [21:54] <Philip> Oh, is $$...$$ block?
# [21:54] <hsivonen> jgraham_: ok. I haven't used it
# [21:54] <hsivonen> Philip: no, $$...$$ is semantically inline but display:block
# [21:54] <jgraham_> It might be the same as $$...$$; I'm not sure
# [21:55] <jgraham_> (there are other things like \begin{align} \end{align} for multiple expressions aligned at a point and so on
# [21:55] <jgraham_> )
# [21:55] <Philip> http://www-h.eng.cam.ac.uk/help/tpl/textprocessing/teTeX/latex/latex2e-html/ltx-421.html
# [21:55] <gsnedders> $…$ is inline
# [21:56] * Quits: jane (j@76.170.65.146) (Client exited)
# [21:56] <gsnedders> and \begin{math}…\end{math} is too
# [21:56] <Philip> vs http://www-h.eng.cam.ac.uk/help/tpl/textprocessing/teTeX/latex/latex2e-html/ltx-224.html which does labelled equations
# [21:56] * Joins: jane (j@76.170.65.146)
# [21:58] * Joins: trackbot-ng (trackbot-n@128.30.52.30)
# [21:58] * Quits: zcorpan (zcorpan@83.227.33.203) (Ping timeout)
# [21:58] * trackbot-ng is loading HTML Issue Tracking data...
# [21:58] * trackbot-ng found 16 users
# [21:58] <trackbot-ng> Tracking ISSUEs and ACTIONs from http://www.w3.org/html/wg/tracker/
# [21:59] <jgraham_> ftp://ftp.ams.org/pub/tex/doc/amsmath/short-math-guide.pdf
# [21:59] <jgraham_> Note that recommends not using $$...$$
# [21:59] <Philip> Looking in some nearby .tex files, one author uses \[ and another uses \begin{equation*}
# [22:00] <jgraham_> The * just means "don't number"
# [22:01] * Philip has always used $$ because it's what he learned first and it has always worked well enough
# [22:01] <Philip> (I still copy-and-paste the same margin-setting commands that I found on the web somewhere when writing my very first .tex file)
# [22:04] * trackbot-ng is loading HTML Issue Tracking data...
# [22:04] * jgraham_ finds LaTeX is often hard work
# [22:04] * trackbot-ng found 16 users
# [22:04] <trackbot-ng> Tracking ISSUEs and ACTIONs from http://www.w3.org/html/wg/tracker/
# [22:05] <gsnedders> Philip: \[ is block
# [22:06] <gsnedders> Philip: so is \begin{equation}
# [22:07] <hsivonen> gsnedders: how did you find out?
# [22:07] * Quits: timbl (timbl@128.30.5.98) (Quit: timbl)
# [22:07] <gsnedders> hsivonen: http://ctan.tug.org/tex-archive/info/lshort/english/lshort.pdf
# [22:07] <Philip> LaTeX doesn't seem to care about semantics, except to the extent that it permits easy styling and changes between styles
# [22:08] <hsivonen> I think the right test is to ask what happens to a comma immediately outside the equation
# [22:08] <gsnedders> Philip: LaTeX cares enough about semantics to make educated guesses about how content should be typeset :P
# [22:09] <jgraham_> gsnedders: I think the extent to which LaTeX has a notion of semantics is highly debatable
# [22:09] <gsnedders> jgraham_: that's true.
# [22:10] <hsivonen> gsnedders: which page? in lshort?
# [22:10] <gsnedders> There again, it isn't overly relevant: HTML has never been intended to create the exact pixel perfect rendering in future version or implementation
# [22:10] <gsnedders> hsivonen: dunno. closed it now.
# [22:11] <hsivonen> ok
# [22:11] <jgraham_> hsivonen: I think the difference you're looking for might depend on whitespace
# [22:12] <gsnedders> hsivonen: just searched for "math" till I got it
# [22:12] <hsivonen> hmm. can CSS do display block with the next punctuation character doing the LaTeX thing?
# [22:12] <jgraham_> \begin{equation} with a newline above it starts a new block
# [22:13] <hsivonen> jgraham_: but are the semantics then <p><math>? :-)
# [22:13] <jgraham_> hsivonen: Well arguably. But I don't see the point of typing the extra <p>
# [22:14] <jgraham_> (the LaTeX whitespace thing is something I always get wrong because it makes the source harder to read)
# [22:14] * Joins: Lachy (Lachlan@84.215.41.149)
# [22:15] <hsivonen> I never got beyond simple cargo cult in LaTeX
# [22:15] <jgraham_> See pg 18 of TNSSGTL2e
# [22:15] <hsivonen> when it was time to do serious stuff, I jumped to XHTML and Prince
# [22:16] <Hixie> i need the opposite of the -a argument to ping
# [22:17] <Hixie> something that beeps whenever a ping takes more than 100ms
# [22:17] <Hixie> (as soon as that 100ms passes)
# [22:18] <gsnedders> hsivonen: I've never found the flexibility in typesetting of maths in XHTML/MathML/Prince to be good enough
# [22:18] <gsnedders> hsivonen: and I've never found anything brilliant at typesetting graphs…
# [22:18] <hsivonen> gsnedders: fortunately, there are areas in applied computer science that are devoid of math :-)
# [22:19] <gsnedders> hsivonen: and I do more writing of the areas that are, and intend on doing a physics degree at uni :P
# [22:19] <hsivonen> reinventing LaTeX in CSS takes lots and lots of work and time
# [22:20] * Quits: Dashiva (noone@84.48.60.15) (Ping timeout)
# [22:20] <gsnedders> (if anyone knows any good way of typesetting graphs please do let me know, though)
# [22:20] * Quits: hasather (hasather@90.231.107.133) (Ping timeout)
# [22:20] <Philip> gsnedders: Gnuplot?
# [22:20] <Philip> plus some magic so it outputs text that LaTeX formats
# [22:20] * Joins: Dashiva (noone@84.48.60.15)
# [22:21] <jgraham_> gsnedders: I don't understand what you mean typesetting
# [22:21] <jgraham_> graphs
# [22:21] <hsivonen> isn't there a LaTeX package for plotting graphs (in the graph theory sense) right in LaTeX?
# [22:21] <Hixie> MikeSmith: btw, rev="" is gone altogether from html5 at the moment
# [22:21] <gsnedders> I more than anything else mean creating graphs from equations and any text (like labels on axis) is actually typeset by TeX
# [22:22] * jgraham_ uses Veusz to plot graphs
# [22:22] * Quits: ROBOd (robod@89.122.216.38) (Quit: http://www.robodesign.ro )
# [22:22] <jgraham_> gsnedders: You could do that with python+matplotlib
# [22:22] <gsnedders> (ideally something that can take LaTeX equations as input)
# [22:22] <jgraham_> you can almost get there (much more easilly) with Veusz
# [22:23] <gsnedders> I'd also like EFS and SVG output
# [22:23] <jgraham_> I don't think anything will take LaTeX as input because I don't think it's unambiguous enough
# [22:23] <jgraham_> efs? do you mean eps?
# [22:24] <gsnedders> yeah
# [22:25] <jgraham_> I think both my suggestions will do that. I'm not entirely sure about Veusz->SVG
# [22:25] * jgraham_ goes to check
# [22:26] <gsnedders> site on lists EPS/PNG/PDF
# [22:26] <gsnedders> *only
# [22:26] * Quits: deltab (deltab@82.36.30.34) (Ping timeout)
# [22:26] <jgraham_> That might be related to the Qt3 version
# [22:27] <anne> hmm, people still believe in conneg
# [22:27] <anne> http://www.w3.org/TR/cooluris/
# [22:27] * jgraham_ has the primary author in his research group
# [22:28] <gsnedders> changelog's latest mention of SVG is: "Diable svg output, and enable pdf output"
# [22:28] * Joins: hasather (hasather@90.231.107.133)
# [22:29] * Joins: timbl (timbl@128.30.5.98)
# [22:29] <anne> hmm
# [22:30] <anne> " One good practice is to include the current year in the URI path, so that you can change the URI-schema each year without breaking older URIs."
# [22:30] <anne> I so not agree
# [22:30] <hsivonen> that's so sad
# [22:31] <hsivonen> the WD that is. not that you disagree with it
# [22:33] <anne> it's unfortunate that it tries to compete with a great resource on the topic by re-using the name
# [22:35] * Quits: Hixie (ianh@129.241.93.37) (Quit: brb)
# [22:35] * Joins: Hixie (ianh@129.241.93.37)
# [22:37] * Quits: Hixie (ianh@129.241.93.37) (Quit: brb again)
# [22:37] * Joins: Hixie (ianh@129.241.93.37)
# [22:38] * Quits: shepazu (schepers@128.30.52.30) (Quit: Core Breach)
# [22:38] * Joins: shepazu (schepers@128.30.52.30)
# [22:40] * Quits: Hixie (ianh@129.241.93.37) (Quit: brb again again)
# [22:40] * Joins: Hixie (ianh@129.241.93.37)
# [22:42] * Quits: hasather (hasather@90.231.107.133) (Quit: leaving)
# [22:43] * Quits: Hixie (ianh@129.241.93.37) (Quit: still trying to make irssi pretty)
# [22:43] * Joins: Hixie (ianh@129.241.93.37)
# [22:44] * Quits: matt (matt@128.30.52.30) (Quit: matt)
# [22:46] <anne> is everyone proposing radical changes to standardization these days because browsers haven't given them enough features?
# [22:46] <anne> or maybe because the idea is that standards have to be finished before they are implemented and therefore everything seems to go so "slow"?
# [22:46] <gsnedders> No: I'm not.
# [22:46] * Quits: Hixie (ianh@129.241.93.37) (Quit: let's try blue...)
# [22:46] * Joins: Hixie (ianh@129.241.93.37)
# [22:47] <gsnedders> anne: in part it seems to be because things like CSS 2.1 are still CRs and not RECs, and therefore not complete.
# [22:47] * Joins: deltab (deltab@82.36.30.34)
# [22:47] <anne> well, that's a direct result of not having done enough testing
# [22:51] <jgraham_> anne: I think it's because browsers haven't given them enough features
# [22:52] <jgraham_> but, also, perhaps because they see some of what is proposed for the future and it doesn't match what they want
# [22:53] <anne> so browsers not giving enough features is a) IE holding things back and for designers it's that CSS isn't doing much to address the application space
# [22:53] <anne> (there should be a b) somewhere in there)
# [22:53] <jgraham_> my impression is that CSS is the biggest problem area in this regard. AFAICT (and this is well away from anything I know about) designers really want something like the XUL box model
# [22:54] <jgraham_> but the attempts to integrate it with the CSS legacy seem to be floundering
# [22:54] <Hixie> the biggest problem is lack of editors
# [22:55] <hsivonen> is that spec so big a task that everyone on their right stays away or is it just a general lack of editors?
# [22:55] <Hixie> but that may be caused, in the csswg case, by the underlying attitude problems
# [22:55] <anne> CSS is a big task
# [22:55] <Hixie> css is no bigger than html5, imho
# [22:55] <anne> case in point
# [22:55] <jgraham_> is lack of editors just a lack of people qualified o edit?
# [22:56] <gsnedders> from what I can see the number of editors is more or less constant, just the length and detail of specs have increased over the years
# [22:56] <Hixie> jgraham_: i don't know
# [22:56] <jgraham_> or are there enough qualified people but they're chosing not to do it?
# [22:56] <Hixie> jgraham_: it's hard to know if someone is qualified to edit until they edit :-)
# [22:56] * Joins: dbaron (dbaron@63.245.220.241)
# [22:56] <Hixie> gsnedders: yeah, probably
# [22:56] * Hixie goes to shower and go to work, bbiab
# [22:56] <hsivonen> I guess a qualified person also needs someone who thinks paying for the editing is worthwhile. or so I'd expect
# [22:56] <gsnedders> which has the effect of there being less specs being put out, as more time is spent on each
# [22:57] <hsivonen> my local copy of Validator.nu now validates http://golem.ph.utexas.edu/~distler/blog/files/xhtml-mathml-svg.xhtml
# [22:57] <gsnedders> in the case of tolerant-http-parsing it's definitely time that is the limit (but, there again, in my case, pay doesn't get around the legal obligation to go to school)
# [23:01] <hsivonen> yay. now it validates http://golem.ph.utexas.edu/~distler/blog/files/xhtml-mathml-ruby.xhtml too
# [23:01] * Quits: gsnedders (gsnedders@86.135.224.200) (Quit: gsnedders)
# [23:01] <anne> you're doing <ruby> already?
# [23:02] <hsivonen> anne: well, I'm just including Petr Nalevka's ruby module
# [23:03] <hsivonen> anne: this is an ad hoc XHTML 1.0 + Ruby combo
# [23:05] * Joins: dbaron_ (dbaron@63.245.220.241)
# [23:06] * Joins: hober (ted@68.101.220.172)
# [23:07] * Quits: dbaron (dbaron@63.245.220.241) (Ping timeout)
# [23:19] * Philip doesn't like random internet outages
# [23:20] <Philip> gsnedders, in case you're reading the log: You can run Gnuplot with
# [23:20] <Philip> set term epslatex color solid
# [23:20] <Philip> set output "whatever.tex"
# [23:20] <Philip> then do epstopdf whatever.eps
# [23:21] <Philip> then
# [23:21] <Philip> \begin{figure}\begin{center}\input{whatever.tex}\end{center}\end{figure}
# [23:21] <Philip> and then run pdflatex and it ought to come out looking pretty and TeXified
# [23:22] <Philip> (You have to be a bit careful with escaping, like doing set ylabel "CPU usage (\\%)" to avoid confusing TeX)
# [23:22] * Quits: aaronlev (chatzilla@66.30.196.151) (Quit: ChatZilla 0.9.79 [Firefox 3.0b3pre/2007121405])
# [23:43] * Quits: shepazu (schepers@128.30.52.30) (Quit: shepazu)
# [23:53] <Philip> & is really the worst possible character to use for splitting arguments in URIs, given that most URIs are written in HTML which has special handling for &
# [23:54] <Philip> Why can't everyone just use ; instead?
# [23:54] * Joins: shepazu (schepers@128.30.52.30)
# [23:57] <hsivonen> Philip: to answer the rhetorical-looking question, e.g. Java servlets don't parse ;-separated parameters out of the box
# [23:57] <hsivonen> supporting ; is an outstanding feature request for Validator.nu
# [23:58] <Yudai> Philip: almost all browsers generate an url with "&" as the splitter in forms
# [23:59] <Philip> Why don't e.g. Java servlets support ';'? CGI.pm has been doing it for decades (maybe), and it seems like an easy way to eliminate the single most common HTML syntax error
# Session Close: Tue Dec 18 00:00:00 2007

The end :)