Options:
- # Session Start: Wed Oct 24 00:00:00 2007
- # Session Ident: #html-wg
- # [00:09] <Hixie> anne: i don't think you meant UMT
- # [00:15] * Joins: aroben (aroben@17.203.12.236)
- # [00:15] * Quits: Sander (svl@86.87.68.167) (Quit: And back he spurred like a madman, shrieking a curse to the sky.)
- # [00:35] <anne> UMT is two hours before Amsterdam time right?
- # [00:35] <Hixie> UMT is metric time
- # [00:35] <Dashiva> UTC, you mean?
- # [00:35] <anne> crap
- # [00:35] * Joins: timbl (timbl@146.115.66.146)
- # [00:35] <Hixie> 1000 UMT is sometime in 3 years.
- # [00:36] <Hixie> i mean, if you _want_ to delay the forms task force, i understand, but 2010 years is pushing it :-P
- # [00:44] * Quits: mjs (mjs@17.255.109.173) (Quit: mjs)
- # [00:59] * Joins: mjs (mjs@17.203.15.161)
- # [01:02] * Quits: billmason (billmason@69.30.57.156) (Quit: .)
- # [01:14] * Joins: sbuluf (pywk@200.49.140.202)
- # [01:20] * Joins: kingryan (rking3@208.66.64.47)
- # [01:30] * Quits: gavin_ (gavin@99.227.30.12) (Ping timeout)
- # [01:31] * Quits: tH (Rob@87.102.76.132) (Quit: ChatZilla 0.9.78.1-rdmsoft [XULRunner 1.8.0.9/2006120508])
- # [01:35] * Joins: gavin_ (gavin@99.227.30.12)
- # [02:47] * Joins: heycam` (cam@130.194.72.84)
- # [02:59] * Quits: kingryan (rking3@208.66.64.47) (Client exited)
- # [03:00] * Joins: mjs_ (mjs@17.255.109.173)
- # [03:38] * Quits: gavin_ (gavin@99.227.30.12) (Ping timeout)
- # [03:43] * Joins: gavin_ (gavin@99.227.30.12)
- # [03:44] * Quits: mjs_ (mjs@17.255.109.173) (Quit: mjs_)
- # [03:47] * Quits: mjs (mjs@17.203.15.161) (Quit: mjs)
- # [04:03] * Joins: kingryan (rking3@64.81.240.149)
- # [04:04] * Quits: kingryan (rking3@64.81.240.149) (Quit: kingryan)
- # [04:07] * Joins: mjs (mjs@17.255.109.173)
- # [04:07] * Joins: mjs_ (mjs@17.203.15.161)
- # [04:20] * Joins: mjs__ (mjs@17.203.14.158)
- # [04:22] * Quits: mjs (mjs@17.255.109.173) (Ping timeout)
- # [04:27] * Quits: mjs_ (mjs@17.203.15.161) (Quit: mjs_)
- # [04:35] * Quits: aroben (aroben@17.203.12.236) (Ping timeout)
- # [04:49] * Joins: karl (karlcow@128.30.52.30)
- # [04:52] * Joins: olivier (ot@128.30.52.30)
- # [05:02] * Joins: aroben (aroben@17.203.12.236)
- # [05:08] * Quits: mjs__ (mjs@17.203.14.158) (Quit: mjs__)
- # [05:21] * Quits: dbaron (dbaron@63.245.220.241) (Quit: 8403864 bytes have been tenured, next gc will be global.)
- # [05:45] * Quits: gavin_ (gavin@99.227.30.12) (Ping timeout)
- # [05:48] * Joins: aroben_ (aroben@17.255.96.238)
- # [05:49] * Quits: aroben (aroben@17.203.12.236) (Ping timeout)
- # [05:50] * Joins: gavin_ (gavin@99.227.30.12)
- # [05:51] * Joins: aroben__ (aroben@17.203.12.236)
- # [05:52] * Quits: aroben_ (aroben@17.255.96.238) (Ping timeout)
- # [05:58] * Quits: marcos (chatzilla@131.181.148.226) (Quit: ChatZilla 0.9.78.1 [Firefox 2.0.0.8/2007100816])
- # [06:08] * Joins: marcos (chatzilla@131.181.148.226)
- # [06:25] * Quits: olivier (ot@128.30.52.30) (Quit: This computer has gone to sleep)
- # [06:27] * Joins: olivier (ot@128.30.52.30)
- # [06:29] * Quits: jane (j@76.170.65.146) (Client exited)
- # [06:29] * Joins: jane (j@76.170.65.146)
- # [06:37] * Joins: mjs (mjs@64.81.48.145)
- # [06:53] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [07:19] * Quits: jgraham (jgraham@81.86.210.220) (Ping timeout)
- # [07:47] * Joins: jgraham (jgraham@81.86.210.220)
- # [07:50] * Quits: marcos (chatzilla@131.181.148.226) (Ping timeout)
- # [07:53] * Quits: gavin_ (gavin@99.227.30.12) (Ping timeout)
- # [07:58] * Joins: gavin_ (gavin@99.227.30.12)
- # [08:39] * Joins: marcos (chatzilla@131.181.148.226)
- # [08:39] * Quits: heycam` (cam@130.194.72.84) (Quit: bye)
- # [08:44] * Quits: mjs (mjs@64.81.48.145) (Quit: mjs)
- # [08:47] * Quits: anne (annevk@81.68.67.12) (Ping timeout)
- # [09:03] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Client exited)
- # [09:12] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [09:16] * Quits: paullewis (paullewis@82.242.109.217) (Quit: paullewis)
- # [09:27] * Joins: mjs (mjs@64.81.48.145)
- # [09:35] * Quits: olivier (ot@128.30.52.30) (Quit: Leaving)
- # [09:36] * Quits: karl (karlcow@128.30.52.30) (Quit: Where dwelt Ymir, or wherein did he find sustenance?)
- # [09:36] * Quits: Thezilch (fuz007@68.54.228.249) (Connection reset by peer)
- # [09:43] * Quits: aroben__ (aroben@17.203.12.236) (Ping timeout)
- # [09:49] * Joins: paullewis (paullewis@81.255.115.137)
- # [10:00] * Quits: gavin_ (gavin@99.227.30.12) (Ping timeout)
- # [10:05] * Joins: gavin_ (gavin@99.227.30.12)
- # [10:08] * Quits: Lachy (Lachy@213.236.208.22) (Quit: Leaving)
- # [10:18] * Joins: tH_ (Rob@87.102.76.132)
- # [10:18] * tH_ is now known as tH
- # [10:42] * Quits: mjs (mjs@64.81.48.145) (Quit: mjs)
- # [10:49] * Joins: mjs (mjs@64.81.48.145)
- # [11:09] * Joins: ROBOd (robod@89.122.216.38)
- # [11:10] * Quits: mjs (mjs@64.81.48.145) (Quit: mjs)
- # [11:10] * Joins: mjs (mjs@64.81.48.145)
- # [11:13] * Quits: sbuluf (pywk@200.49.140.202) (Quit: sbuluf)
- # [11:26] * Quits: mjs (mjs@64.81.48.145) (Quit: mjs)
- # [11:40] * Joins: mjs (mjs@64.81.48.145)
- # [11:40] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Quit: Less talk, more pimp walk.)
- # [12:06] * Quits: gavin_ (gavin@99.227.30.12) (Ping timeout)
- # [12:11] * Joins: gavin_ (gavin@99.227.30.12)
- # [12:18] * Joins: anne (annevk@81.68.67.12)
- # [12:31] <anne> "Gigantoredesignorrific"
- # [12:31] <anne> lol
- # [12:34] <anne> Hixie, lots of </dfn> missing in the error code table
- # [12:34] <Hixie> crap
- # [12:34] <Hixie> thanks
- # [12:35] <Hixie> fixed
- # [12:35] <anne> would be nice if Bert's script indicated errors somehow
- # [12:38] <anne> so db.transaction(function(t){ t.executeSql("SELECT ...", [...], function(t2,r) { alert(r.insertId }) }) ...
- # [12:38] <anne> yikes
- # [12:39] <Hixie> you can s/t2/t/
- # [12:39] <anne> Hixie, if SQLStatementCallback can be implemented as a function accepting a single argument I think it makes more sense to let the results be first and the transaction be second
- # [12:39] <anne> ok, wasn't sure
- # [12:40] <Hixie> yeah i considered that but the problem then is that you never know what the first argument will be
- # [12:40] <Hixie> generally making sure your arguments all line up to be hte same order is better api design
- # [12:40] <Hixie> less surprising
- # [12:40] <anne> it would always be of type SQLResultSet no?
- # [12:40] <Hixie> well it would be SQLResultSet for hte first callback and SQLError for the second
- # [12:41] * Joins: Lachy (Lachy@213.236.208.22)
- # [12:41] <Hixie> and SQLTransaction for the transaction()'s first callback's first argument
- # [12:41] <anne> yeah, but you wouldn't implement those as the same function anyway
- # [12:41] <Hixie> and SQLError for the transaction()'s second callback's first argument
- # [12:41] <Hixie> sure but as an author you'd get really confused as to which was which
- # [12:41] <Hixie> i'm not saying it would be technically a problem
- # [12:41] <Hixie> just that it would be a pain for authors
- # [12:41] <Hixie> at least as much of a pain as having to declare two arguments
- # [12:42] <Hixie> possibly more
- # [12:42] <Hixie> as it would lead to more bugs
- # [12:42] * Quits: Lachy (Lachy@213.236.208.22) (Quit: Leaving)
- # [12:42] * Joins: Lachy (Lachy@213.236.208.22)
- # [12:42] <anne> ok, would DOM-Bindings handle these double argument handleEvent methods too btw?
- # [12:43] <anne> also, why are transactions so important?
- # [12:43] * Quits: deltab (deltab@82.36.30.34) (Ping timeout)
- # [12:44] <Hixie> i hope so (re DOM bindings)
- # [12:44] * Joins: deltab (deltab@82.36.30.34)
- # [12:44] <Hixie> how do you mean? (re transactions)
- # [12:45] <anne> well, I had to read up on them to learn what they exactly did, I've never used them anywhere... then again, I don't write banking applications and such
- # [12:45] <Hixie> heh
- # [12:46] <Hixie> they're needed whenever you have multiple threads manipulating the same data, basically
- # [12:47] <Hixie> e.g. if you want to update a row in one table to equal the results of processing another table
- # [12:47] <Hixie> you don't want another thread editing the second table in between the SELECT and the UPDATE
- # [12:47] <Hixie> otherwise they'll get out of sync
- # [12:48] <anne> hmm, but we don't have threads
- # [12:51] <Hixie> multiple browser windows = threads
- # [12:51] <anne> ah yeah, and everything is done async as well...
- # [12:51] <Hixie> and we will have threads relatively soon, i imagine, people are clamouring for WorkerPool in html5
- # [12:51] <anne> can you still implement that using a single thread model?
- # [12:52] <Hixie> WorkerPool would introduce true threads
- # [12:52] * anne also wonders how it relates to roc's "fantasy browser"
- # [12:53] <Hixie> (threads wouldn't be able to access any objects from the "UI" layer, but they could communicate using async messages)
- # [12:53] <anne> so you'd have a thread mostly for calculating and such?
- # [12:57] <Hixie> yeah
- # [12:57] <Hixie> or for communicating to the server
- # [12:58] <Hixie> e.g. so that an offline app could send updates to and get updated from the server while online
- # [12:58] * drry is now known as {|}
- # [12:58] * {|} is now known as {|}0
- # [12:58] * {|}0 is now known as {|}_
- # [12:58] * Hixie replies to 29 e-mails on the SQL stuff in one go
- # [12:58] <anne> ok, so you'd still have access to XMLHttpRequest?
- # [12:58] <Hixie> sure
- # [12:58] * Quits: {|}_ (drry@222.225.140.32) (Quit: {|}_)
- # [12:59] <Hixie> and TCPConnection and so on
- # [12:59] <anne> so what's the "UI" layer exactly?
- # [12:59] <anne> the DOM?
- # [12:59] <Hixie> right
- # [12:59] <Hixie> and anything it touches
- # [12:59] <Hixie> basically you'd have two completely separate GC roots
- # [12:59] <anne> this might be problematic for resolving URIs with XMLHttpRequest...
- # [13:00] <Hixie> (well, n completely separate GC roots)
- # [13:00] <Hixie> why?
- # [13:00] <anne> although it depends on how things are defined I suppose
- # [13:00] <anne> if one thread changes the base URI in some way and the other does a request
- # [13:00] <Hixie> the two threads couldn't affect each other
- # [13:00] <Hixie> they have absolutely no objects in common
- # [13:01] <anne> yeah, but XMLHttpRequest needs document.baseURI
- # [13:01] <Hixie> it would have a completely separate document
- # [13:01] <anne> interesting
- # [13:01] <Hixie> (probably one that === null)
- # [13:01] <Hixie> (so XHR would have to be tweaked to handle that case)
- # [13:01] <anne> k
- # [13:01] <Hixie> (but that's no biggie)
- # [13:01] * Joins: drry (drry@222.225.140.32)
- # [13:07] <anne> you made a mistake in your e-mail
- # [13:08] <anne> saying that there's no way to a transaction as opposed to not having a way to not do a transaction
- # [13:14] * Quits: timbl (timbl@146.115.66.146) (Quit: timbl)
- # [13:17] <Hixie> heh oops
- # [13:17] <Hixie> where?
- # [13:19] <anne> "I have not supplied a way to do a statement with a transaction in this version of the API"
- # [13:19] <anne> or does that refer to COMMIT/BEGIN/ etc.?
- # [13:19] * anne thought s/with/without/
- # [13:20] <anne> http://www.w3.org/mid/471EF747.8030107@cisra.canon.com.au is funny
- # [13:20] <Hixie> er oops, yeah, shoulda said "without"
- # [13:21] <Hixie> well hopefully people will understand what i meant
- # [13:21] <Hixie> Anthony is right, many parts of "the world" have moved to XML
- # [13:22] <Hixie> of course, they're moving _from_ SGML and other languages, not from HTML.
- # [13:22] <Hixie> anyway, bed time
- # [13:27] * Joins: zcorpan (zcorpan@88.131.66.80)
- # [13:38] <anne> the new API prolly made http://glazkov.com/blog/html5-gears-wrapper/ obsolete too, and WebKit's implementation
- # [13:47] <anne> http://burningbird.net/technology/dash-it-all/
- # [13:57] * Joins: timbl (timbl@128.30.7.216)
- # [13:58] * anne comments
- # [14:14] * Quits: gavin_ (gavin@99.227.30.12) (Ping timeout)
- # [14:19] * Joins: gavin_ (gavin@99.227.30.12)
- # [14:41] <Philip> By the way, Mozilla had a fixed non-premultiplied getImageData for about three minutes yesterday, so that's good progress towards interoperability
- # [14:41] <Philip> (but then it was backed out because they got the approval process wrong, or something)
- # [14:42] <Philip> shape();transform();fill() is probably the most significant issue remaining, as far as I can remember
- # [15:02] * Joins: matt (matt@128.30.52.30)
- # [15:10] <anne> heh
- # [15:10] <anne> you fixed some other issue in Firefox though?
- # [15:25] * Quits: Lachy (Lachy@213.236.208.22) (Quit: Leaving)
- # [15:25] * Joins: Lachy (Lachy@213.236.208.22)
- # [15:33] * Joins: myakura (myakura@122.18.197.237)
- # [16:21] * Quits: gavin_ (gavin@99.227.30.12) (Ping timeout)
- # [16:26] * Joins: gavin_ (gavin@99.227.30.12)
- # [16:38] * Joins: billmason (billmason@69.30.57.156)
- # [16:48] * Joins: Thezilch (fuz007@68.54.228.249)
- # [16:57] <Philip> anne: I fixed the thing with fillRect/drawImage/etc clobbering the path in Firefox, if that's what you're referring to
- # [16:58] <Philip> I wrote the patch for premultiplied getImageData too, hence noticing that it almost got applied properly :-)
- # [17:02] <anne> nice
- # [17:04] <anne> it's too bad the actual Mozilla developers are busy with adding more features though...
- # [17:05] <gavin> which features?
- # [17:05] <anne> mozDrawText() or something in that direction
- # [17:05] <gavin> we're past feature freeze for 1.9, so most developers are in fact not working on features right now
- # [17:06] <anne> I should've used past tense I suppose
- # [17:06] * Philip wrote a patch for shadows too, which almost counts as a new feature :-)
- # [17:06] <gavin> ok
- # [17:06] <Philip> (but it seems it won't be applied in 1.9)
- # [17:07] <gavin> I don't think that's true
- # [17:07] <gavin> you mean 389366, right?
- # [17:08] <gavin> it has approval1.9+ and is blocking1.9+, so it will land for M10 (beta 2)
- # [17:08] <Philip> By shadows, I mean 310682
- # [17:08] <gavin> oh, ok
- # [17:08] <gavin> yeah, that's unlikely to land for 1.9
- # [17:10] <Philip> 380338 is the other one I remember as being reasonably significant for interoperability, but there doesn't seem to be much activity in that bug
- # [17:10] <anne> seems someone else provided the text draw extension and vlad merely reviewed it
- # [17:10] * anne retracts his statement he made earlier
- # [17:11] <gavin> Philip: you could always ask for blocking1.9? if you think it's important...
- # [17:12] <gavin> a comment in the bug explaining why you think it's important would help, if you do
- # [17:12] <Philip> gavin: Okay, that sounds like it might be worthwhile
- # [17:13] <Philip> I'll try to look through the other bugs I know of to see if there are other problems that people might actually care about in practice
- # [17:13] <gavin> great
- # [17:16] <Philip> (I'm currently ignoring all the issues like "lines which intersect their own miter are drawn with even-odd filling" which are different between browsers but which nobody actually cares about :-) )
- # [17:16] <gavin> anne: maybe I'm just a bit sensitive to people saying "Mozilla developers" as though that's a group that acts cohesively :)
- # [17:17] * Joins: aaronlev (chatzilla@66.31.86.217)
- # [17:19] <anne> gavin, granted, it's a bit more complex than that :)
- # [17:21] * Quits: myakura (myakura@122.18.197.237) (Quit: Leaving...)
- # [17:22] <Philip> The system is complex enough that whenever I submit a patch, I just hit random checkboxes and shuffle dropdown lists until it gets checked in
- # [17:23] <gavin> yeah
- # [17:23] <Philip> If I change things enough times, eventually someone who knows what they're doing will come along and sort it out
- # [17:23] <gavin> there is some documentation available at http://developer.mozilla.org/en/docs/Getting_your_patch_in_the_tree , but there are details that aren't covered
- # [17:23] <gavin> like our nearing-release approval process, which can be complcicated and something changes a lot
- # [17:24] <gavin> (currently is documented at http://wiki.mozilla.org/TreeStatus )
- # [17:25] <gavin> but I suppose for that to be relevant you need to have some context about what our rmilestone release process is like
- # [17:25] <anne> I suppose that if you just cc yourself to every bug like you you'll be fine :p
- # [17:26] <gavin> I follow things rather closely, yes :)
- # [17:26] <gavin> Philip: anyways, I'm always glad to answer questions should you have them :)
- # [17:27] * Quits: zcorpan (zcorpan@88.131.66.80) (Ping timeout)
- # [17:44] * Quits: paullewis (paullewis@81.255.115.137) (Quit: paullewis)
- # [17:47] * Joins: Sander (svl@86.87.68.167)
- # [18:12] * Joins: zcorpan (zcorpan@83.227.34.9)
- # [18:22] * Joins: paullewis (paullewis@82.242.109.217)
- # [18:28] * Quits: gavin_ (gavin@99.227.30.12) (Ping timeout)
- # [18:33] * Joins: gavin_ (gavin@99.227.30.12)
- # [18:44] * Quits: zcorpan (zcorpan@83.227.34.9) (Ping timeout)
- # [19:30] * Quits: mjs (mjs@64.81.48.145) (Quit: mjs)
- # [19:32] * Joins: aroben__ (aroben@17.203.12.236)
- # [19:52] * Joins: kingryan (rking3@208.66.64.47)
- # [19:56] * Joins: dbaron (dbaron@63.245.220.241)
- # [19:59] * Joins: ChrisWilson (cwilso@131.107.0.103)
- # [20:06] <DanC> hi Chris
- # [20:06] <DanC> I just sent mail to public-html about test suite licensing; can you take a look?
- # [20:07] <DanC> hmm... is GR around? (oedipus [sp?]) or Ben Millard?
- # [20:08] <DanC> jgraham, are you around to chat about issue tracking? the W3C systems elves did their magic and we have http://www.w3.org/html/wg/tracker/ now
- # [20:19] <anne> It would be nice if the licensing implications were made more explict as IANAL
- # [20:22] <DanC> http://www.w3.org/2007/10/18-html-wg-minutes
- # [20:22] <anne> I thought I read those...
- # [20:22] <ChrisWilson> That's the goal; we're just trying to make sure the licensing implications are essentially "open source" if possible, to enable cross-pollination with html5-lib.
- # [20:22] <hsivonen> I wonder why the W3C felt a need to develop its own license instead of reusing 3-clause BSD or MIT. I guess the rationale for creating a new license would explain the difference.
- # [20:23] <ChrisWilson> (Dan's pointer was not relevant to your question, Anne)
- # [20:23] <DanC> oh... I wasn't responding to you when I dropped that URI, anne
- # [20:23] <anne> so it seems, doesn't not contain "license" or "test"
- # [20:23] <ChrisWilson> I'd recommend not trying to chart those waters, Henri. Or at least not blocking on that. :)
- # [20:23] <anne> euh s/not //
- # [20:24] <anne> Moving of code.google.com would be annoying
- # [20:24] <anne> s/of/away from/
- # [20:24] <DanC> the difference between the MIT license and the W3C software license is basically that the W3C software license uses a URI to identify the work that is licensed
- # [20:24] <ChrisWilson> Anne, can you elaborate on that statement?
- # [20:24] <ChrisWilson> W3C specs and test suites need to be captured on W3.org.
- # [20:25] * DanC looks to see if this is covered in the W3C copyright faq...
- # [20:26] * Joins: jgraham_ (jgraham@81.86.210.220)
- # [20:27] <jgraham_> DanC: re: issue tracking; if I can have ~25 minutes I can go home and be avaliable
- # [20:27] <DanC> that should be fine, jgraham
- # [20:27] <jgraham_> OK
- # [20:27] * Quits: jgraham_ (jgraham@81.86.210.220) (Quit: Ex-Chat)
- # [20:27] * Joins: IanJ (Ian@128.30.52.30)
- # [20:27] <anne> ChrisWilson, just reading what Sam Ruby just posted
- # [20:27] <DanC> IanJ, do you remember why W3C made a new software license rather than just using the MIT license?
- # [20:28] <anne> ChrisWilson, I'm fine with W3C hosting a copy
- # [20:28] <anne> ChrisWilson, that's allowed per the current license afaict
- # [20:28] <IanJ> I don't remember. I think that predates my attention to these things.
- # [20:29] <DanC> I'm pretty happy to set up a copy of the code.google stuff in w3.org; I'd rather use bzr/mercurial ... in fact, I might use hgimportsvn for the w3.org copy
- # [20:29] <anne> code.google.com does not allow the W3C license btw
- # [20:29] <anne> Sam Ruby pointed to http://groups.google.com/group/codesite-discuss/msg/56b51d094e762976
- # [20:29] <ChrisWilson> I'm not sure what the end sum is here. I don't mind, obviously, if html5lib continues to be developed as a test suite and feeds into the HTML 5 test suite at the W3C on a continuing basis.
- # [20:29] <DanC> anne, google considered and rejected the W3C software license? or just hasn't been asked yet?
- # [20:29] <ChrisWilson> I don't think it's a good idea to host W3C projects at 3rd party sites.
- # [20:30] * hsivonen wasn't aware that W3C supported anything but CVS for version control
- # [20:30] <anne> DanC, the latter, I think
- # [20:30] * DanC reads from Daniel Berlin...
- # [20:31] <anne> ChrisWilson, html5lib is not really a W3C project though
- # [20:31] <anne> ChrisWilson, it's something I set up together with jgraham a year ago or so
- # [20:31] <DanC> well, I'd appreciate it if you'd submit it to the HTML WG for review and such, under open source terms.
- # [20:32] <anne> anyway, I suppose moving it could be done as well
- # [20:32] <anne> I just want the licensing implications clear first
- # [20:32] <DanC> IanJ, do you know of any substantive differences between the MIT license and the W3C software license? maybe I should get back to rigo?
- # [20:33] <ChrisWilson> Anne, I know - I'm talking about "official test suite for W3C HTML 5", which is in our charter to produce as a working group.
- # [20:33] <IanJ> I don't know the MIT license.
- # [20:33] <ChrisWilson> Sure, I understand wanting the licensing implications clear - Dan was looking for something compatible with the current html5lib license, so content could move back and forth with no legal restrictions.
- # [20:34] <ChrisWilson> (i.e., in IANAL terms, "free".)
- # [20:34] <ChrisWilson> and I'm not a lawyer.
- # [20:34] <anne> IanJ, http://www.opensource.org/licenses/mit-license.php
- # [20:34] * IanJ thanks anne
- # [20:35] <DanC> IANAL either, so I asked one of the lawyers in W3C, and he advised using the W3C software license. If that's good enough for you, anne (and other html5lib contributors), that would be great. Otherwise, I have to go back and talk with him more.
- # [20:35] <DanC> but I think Rigo is likely to be OK with the MIT license.
- # [20:35] <DanC> any particular licensing implications that I could clarify, anne?
- # [20:36] <anne> well, what does using the W3C software license mean?
- # [20:36] <IanJ> difference #1: w3c requires "notice" (I read: "Explanation") of modificationz.
- # [20:36] * Quits: gavin_ (gavin@99.227.30.12) (Ping timeout)
- # [20:36] <IanJ> Diference #2: Any pre-existing intellectual property disclaimers, notices, or terms and conditions must be included in any derivative works.
- # [20:37] <IanJ> The only thing MIT lic. seems to require is a copyright notice and reference to permission notice.
- # [20:37] <IanJ> (also required by w3c license)
- # [20:38] <IanJ> So, on first, read, w3c requires more info about changes in derivative works and also inclusion of any other relevant ipr statements.
- # [20:38] <IanJ> seems copyright is held by original copyright holders for both licenses.
- # [20:40] <anne> we went with MIT as it it's the most permisse license out there and because Google didn't allow Public Domain (iirc, maybe I was the only one who wanted PD though)
- # [20:41] <DanC> anne, I figured you could change to the W3C software license by editing http://html5lib.googlecode.com/svn/trunk/LICENSE , with the knowledge and consent of the people listed there... but as Sam Ruby points out, I guess we'd have to get code.google to add it as an option
- # [20:41] * Joins: gavin_ (gavin@99.227.30.12)
- # [20:42] <DanC> ok, looks like I should take Sam Ruby's msg back to Rigo.
- # [20:42] * Joins: mjs (mjs@17.255.109.173)
- # [20:43] <anne> hmm, so if someone takes html5lib and adds an option he would have to file a bug or something on our project and contribute that option back per the W3C license?
- # [20:43] <anne> that seems annoying
- # [20:43] <DanC> Ian, is that what "notice" means?
- # [20:43] <anne> what's wrong with MIT for the W3C exactly?
- # [20:44] <IanJ> I believe "notice" here means "explanation", not email.
- # [20:44] <DanC> I don't know of anything wrong with the MIT license, anne.
- # [20:44] <DanC> explanation where, Ian?
- # [20:44] * IanJ gets text to paste...one moment
- # [20:44] <IanJ> http://www.w3.org/Consortium/Legal/2002/copyright-software-20021231
- # [20:44] * Quits: mjs (mjs@17.255.109.173) (Quit: mjs)
- # [20:44] <hsivonen> This is a rather unusual situation. In my experience, the MIT license has always been the easiest one to get people to buy into
- # [20:44] <IanJ> "....provided that you include the following on ALL copies ......"
- # [20:45] <IanJ> ".....Notice of any changes or modifications to the files,...."
- # [20:45] <DanC> (the only thing "wrong" with the MIT license is that I don't have W3C management/legal approval to do HTML test suite work using it... at least, not yet)
- # [20:45] * Joins: mjs (mjs@17.255.109.173)
- # [20:45] <IanJ> Thus, I read "Notice" to be "please document changes"
- # [20:45] <IanJ> I do not read "Notice" here to mean "You have to tell w3c"
- # [20:45] <DanC> likewise, we could ask what's wrong with the W3C software license for code.google. nothing, as far as we know, but we'd have to ask.
- # [20:45] * Quits: mjs (mjs@17.255.109.173) (Quit: mjs)
- # [20:45] * Joins: mjs (mjs@17.255.109.173)
- # [20:45] <anne> that's not exactly the same question
- # [20:46] <anne> you could ask what's wrong with the W3C software license for html5lib, and that would be one the reasons
- # [20:46] <DanC> right
- # [20:46] <anne> other reasons might be that the W3C license is more restrictive
- # [20:46] <mjs> I'd much prefer if the test suite was under a permissive open source license
- # [20:46] <DanC> the W3C software license is pretty permissive
- # [20:46] <IanJ> I read it as saying that modifications are ok, but documentation of changes is required.
- # [20:47] <IanJ> Such documentation does not seem to be required by the MIT license.
- # [20:47] <ChrisWilson> mjs, that was our goal.
- # [20:47] <mjs> w3c software license might be ok
- # [20:47] <mjs> w3c documentation license would be quite problematic
- # [20:47] <DanC> quite
- # [20:47] <mjs> we often use test suites directly in our publicly redistributed test suite
- # [20:47] * DanC wonders if mjs has seen the email thread
- # [20:47] <mjs> I have
- # [20:48] <mjs> anyway, we often have to make small changes to work with our regression test framework
- # [20:48] <mjs> and having to document those to whatever degree the license is supposed to require is a hassle
- # [20:48] <mjs> especially if we have to make a small change to every individual file
- # [20:48] <mjs> plus I don't like redistributing stuff that I'm not sure is truly open source compatible
- # [20:49] <DanC> the W3C software license is OSI certified
- # [20:49] <mjs> so MIT license would make me happiest
- # [20:49] <DanC> but OK, I'll run the MIT license up the W3C legal flagpole and see what I get
- # [20:50] <ChrisWilson> Can someone (mjs, henri, anne?) sum up what's wrong with the W3C software license for this purpose? My understanding from Dan is that it permits all the things you've stated.
- # [20:50] <mjs> w3c software license is probably not unreasonable but I'm not sure it is compatible with the LGPL, and I'm sure the MIT license is
- # [20:52] <DanC> "Software which is free from any claims beyond W3C terms and conditions are compatible with the GPL and may be redistributed under the GPL." -- http://www.w3.org/Consortium/Legal/IPR-FAQ-20000620#GNU
- # [20:52] <DanC> IANAL and all that.
- # [20:53] <anne> ChrisWilson, it's more restrictive, basically, for users of the code
- # [20:53] <ChrisWilson> yeah, what he said. :)
- # [20:53] <hsivonen> ChrisWilson: the MIT license is well understood and already used in this case. Changing the license to something less understood creates concern and changing it at all creates inconvenience to everyone involved who may have to get legal approval for the new license (from whomever it is they need legal approval for doing what they do).
- # [20:53] <ChrisWilson> how so? It's explicitly compatible with GPL.
- # [20:53] <anne> they have to provide notices for changes, etc.
- # [20:54] <mjs> DanC: whose legal opinion is it that the w3c license is compatible with the GPL?
- # [20:54] <ChrisWilson> Sigh. I'm not suggesting html5lib change their license. I'm pointing out that having a W3C deliverable that is NOT licensed with a W3C license is going to be a pain in the ass, and is going to attract lawyers.
- # [20:54] <mjs> DanC: is that a w3c opinion or an fsf opinion?
- # [20:54] <anne> this is going to attract lawyers anyhow
- # [20:54] <anne> either at W3C or at code.google.com
- # [20:54] <jgraham> DanC: I'm around now
- # [20:55] <DanC> I think it's an FSF opinion, mjs, but that's a good question.
- # [20:55] <DanC> I'm pretty sure Joe Reagle asked the FSF
- # [20:55] <IanJ> http://www.fsf.org/licensing/licenses/
- # [20:55] <IanJ> " GPL-Compatible Free Software Licenses"
- # [20:55] <IanJ> ... W3C Software Notice and License
- # [20:56] <hsivonen> (IANAL, but I would assume dual licensing to be an option, albeit silly when the licenses are so similar)
- # [20:57] <DanC> jgraham, my attention is a bit divided, but please take a look at http://www.w3.org/html/wg/tracker/
- # [20:57] <anne> ChrisWilson, also, I believe DanC suggested publishing snapshots under some restrictive W3C license
- # [20:57] <DanC> yes
- # [20:57] <anne> ChrisWilson, which would likely satisfy that particular crowd of people (whichever that is)
- # [20:57] <hsivonen> I'm more interested if that the more restrictive snapshot would require copyright *assignment*
- # [20:58] * Joins: briansuda (briansuda@85.220.118.144)
- # [20:58] <hsivonen> also, why would anyone want to use the more restrictive snapshot instead of checking out MIT stuff from Google Code
- # [20:58] <ChrisWilson> I'm going to stop talking, because I think I'm repeating myself. <ChrisWilson> I'm not sure what the end sum is here. I don't mind, obviously, if html5lib continues to be developed as a test suite and feeds into the HTML 5 test suite at the W3C on a continuing basis.
- # [20:58] <DanC> jgraham, the idea is to add issues to http://www.w3.org/html/wg/tracker/ from public-html, from hixie's list, etc. ; Julian and a few others offered to help triage too
- # [20:59] <jgraham> OK
- # [20:59] * DanC is content to take the licensing discussion back to email too
- # [20:59] * Quits: mjs (mjs@17.255.109.173) (Quit: mjs)
- # [21:04] * anne wonders what Mark Birbeck is trying to say
- # [21:04] * anne leaves it
- # [21:04] * anne also won't reply to the licensing thread for now
- # [21:04] * IanJ waves...
- # [21:04] * Parts: IanJ (Ian@128.30.52.30) (Client exiting)
- # [21:10] * jgraham gets the last hours worth of email all in one go, for some reason
- # [21:15] * hsivonen is confused about the relationship of the spec markup and styling
- # [21:15] <anne> Hixie uses negative margins for that
- # [21:15] <anne> if I assume correctly what you're talking about, that is
- # [21:15] <hsivonen> anne: for the element definition boxes?
- # [21:15] <anne> yeah
- # [21:16] <hsivonen> anne: thanks
- # [21:16] <hsivonen> I had hoped to see a container element to ease scraping.
- # [21:16] <hsivonen> oh well
- # [21:17] <hsivonen> looks like I need to remember the previous heading so that I can act on it when the parser reports a dl with class element
- # [21:18] * Joins: mjs (mjs@17.255.109.173)
- # [21:19] * Quits: mjs (mjs@17.255.109.173) (Quit: mjs)
- # [21:28] <jgraham> So I'm wary of changing the html5lib tests to be under the w3c license for reasons that have mostly been stated already, notably that getting the agreement of all contributors may be non-trivial, that it is a less well known license which may affect perception of when the software can be used, that it requires us to give up google code, which has been pretty reliable, and that the condition requiring documentation of modifications in the W3C license seems
- # [21:28] <jgraham> somewhat onerous.
- # [21:28] <jgraham> Dumping snapshots to w3c license presumbly prevents bugs fixed in the snapshot being reincorporated into the main tree
- # [21:29] <anne> hmm, good point
- # [21:29] <jgraham> However I am prepared to work around all these issues if a license change is the least painful way to get the tests used
- # [21:30] <hsivonen> jgraham: unless the creators of bug fixes agree to release their fixes under the MIT license
- # [21:30] <jgraham> hsivonen: Of course, but it creates overhead
- # [21:30] <hsivonen> yes
- # [21:32] <jgraham> I can't even imagine what would happen if they were unwilling to MIT license their fix. Would we just have to remove the whole test?
- # [21:33] <jgraham> (and test the same thing a different way)
- # [21:34] <ChrisWilson> wow, I'm amazed you're that worried about test suite bug fix licensing, but were fine with moving the HTML5 effort to W3C license.
- # [21:35] <hsivonen> ChrisWilson: fwiw, when I copied bits of the spec to my code comments, I used the WHATWG copy of the spec in order to avoid any W3C document license issues.
- # [21:57] * DanC tunes back in...
- # [21:59] <DanC> hsivonen, on Rigo's behalf, I should note that you don't necessarily avoid W3C license issues by doing that; HTML 5 is, arguably, a derivative work of HTML 4, which is copyright W3C.
- # [22:01] * Quits: aroben__ (aroben@17.203.12.236) (Ping timeout)
- # [22:06] <hsivonen> DanC: I copied the tokenization section. IANAL, but I don't buy the notion that was a derivative work of HTML 4 in the copyright sense
- # [22:07] <DanC> well, you've been warned.
- # [22:08] <DanC> Rigo and Danny, who are laywers, convinced me.
- # [22:09] <DanC> I think excerpting into code is fair use, in any case.
- # [22:22] * Quits: briansuda (briansuda@85.220.118.144) (Quit: briansuda)
- # [22:24] * Joins: aroben__ (aroben@17.203.12.236)
- # [22:26] <DanC> hmm... I wonder if that's new information that motivates re-considering the 9 May decision. ugh.
- # [22:33] <Hixie> btw i'm on the google code / project hosting team at google, so if there are any questions about what licenses are allowed or whatever, i can answer them
- # [22:34] <Hixie> (i work directly with greg stein, daniel berlin, and others)
- # [22:35] * Quits: ROBOd (robod@89.122.216.38) (Quit: http://www.robodesign.ro )
- # [22:41] <anne> shepazu, http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/,access
- # [22:43] * Quits: gavin_ (gavin@99.227.30.12) (Ping timeout)
- # [22:48] * Joins: gavin_ (gavin@99.227.30.12)
- # [22:55] * Quits: matt (matt@128.30.52.30) (Quit: matt)
- # [22:55] <DanC> the pointers that Sam Ruby gave made google's position clear enough for me, hixie, but thanks.
- # [23:07] <Philip> It's slightly annoying how http://burningbird.net/technology/dash-it-all/ loses all its HTML entities in Opera (at least 9.2)
- # [23:13] <hsivonen> Philip: XML on the Web is so easy
- # [23:17] <anne> I fell in that trap once
- # [23:19] <jgraham> Using XML on the web?
- # [23:20] <anne> Yes, and that it was easy
- # [23:22] <anne> Took me quite a while to get out of it too
- # [23:23] <jgraham> My website is still XML but I haven't updated it for years. I should really ditch the whole thing and start again
- # [23:23] <jgraham> Perhaps with some sort of CMS. But I always think I should write one nd it just hasn't happened yet
- # [23:24] * jgraham is currently trying to make html5lib work with recent lxml releases instead
- # [23:24] * Quits: timbl (timbl@128.30.7.216) (Quit: timbl)
- # [23:27] * Philip wonders how to store data for something like http://canvex.lazyilluminati.com/survey/2007-07-17/analyse.cgi/index in a way that's much more efficient to display
- # [23:27] <Philip> (Currently it's just got tables like (page, tag, count) and (page, tag, attribute, count) and then does some SQL to get all the aggregated numbers, which isn't great)
- # [23:30] * Joins: mjs (mjs@17.255.109.173)
- # [23:34] * Quits: mjs (mjs@17.255.109.173) (Quit: mjs)
- # [23:34] * Joins: mjs (mjs@17.255.109.173)
- # [23:35] <anne> Hixie, I'd be fine with naming it aria-role=
- # [23:44] <kingryan> jgraham: btw, I make the ruby version of html5lib actually test the errors
- # [23:44] <kingryan> the changes to the tests shouldn't break the python version, though
- # [23:44] <jgraham> So it checks what?
- # [23:44] <jgraham> The name of the error
- # [23:45] <jgraham> Wow, that looks good
- # [23:45] <jgraham> Did you have to do the while thing by hand?
- # [23:46] <jgraham> s/while/whole/
- # [23:46] <jgraham> s/thing/patch/
- # [23:59] * anne thought the python version did that too
- # [23:59] <anne> thanks to markp
- # Session Close: Thu Oct 25 00:00:01 2007
The end :)