Semantic Web Interest Group IRC Chat Logs for 2002-08-23

00:29:14 <danbri> dc_rdfig:view

00:29:14 <dc_rdfig> C: Accessibility Valet - enhanced EARL (http://valet.webthing.com/access/online.html)

00:29:14 <dc_rdfig> D: Topic Maps Published Subjects - Definitions, Requirements and Examples (http://www.oasis-open.org/committees/tm-pubsubj/docs/recommendations/general.htm)

00:29:15 <dc_rdfig> E: Layering LX on RDF (http://www.w3.org/2002/08/LX/RDF/layering)

00:29:16 <dc_rdfig> F: Gary Frederick's message about converting ical to an xml format (http://lists.w3.org/Archives/Public/www-rdf-calendar/2002Jul/0001.html)

00:29:18 <dc_rdfig> G: Face-To-Face communications (http://www.economist.com/printedition/displayStory.cfm?Story_ID=1292480)

00:30:32 <danbri> F:see also recently [http://slashdot.org/articles/02/08/22/1454227.shtml?tid=99|slashdotted] [http://www.desktoplinux.com/news/NS8491802052.html|announcement] about Oeone and OpenOffice.org collaboration re Mozilla/calendar.

00:30:32 <dc_rdfig> Added comment F10.

02:12:15 <Dan_C>http://slashdot.org/articles/02/08/22/2213208.shtml?tid=155

02:12:15 <dc_rdfig> A: http://slashdot.org/articles/02/08/22/2213208.shtml?tid=155 from Dan_C

02:12:25 <Dan_C> A:|BT Loses Case Over Hyperlink Patent

02:12:25 <dc_rdfig> Titled item A.

03:25:38 <jordan> jordan is now known as j|out

04:13:39 <DanCon> crud... bug in uripath.refTo

04:13:47 <DanCon> refTo input file:/some/dir/foo -> file:/some/dir/#blort

04:13:47 <DanCon> result: #blort =?= ./#blort

04:17:36 * DanCon waves

04:17:58 <AaronSw> A:It was a relatively close call, though. The judge struck it down because the patent specified a "central computer" and the Internet is clearly not a central computer...

04:17:58 <dc_rdfig> Added comment A1.

04:18:01 * AaronSw waves

04:18:39 * DanCon debugs uripath.refTo

04:18:41 <DanCon> refTo input file:/some/dir/foo -> file:/some/dir/#blort

04:18:41 <DanCon> result: #blort =?= ./#blort

04:19:12 * DanCon wonders if there's emacs support for python debugging

04:20:07 <AaronSw> A:Oops, never mind. It's also because HTML isn't made up of "blocks of information" and URLs aren't "a complete address".

04:20:08 <dc_rdfig> Added comment A2.

04:20:29 * DanCon wonders if AaronSw has been studying law

04:20:52 <AaronSw> I've been hanging around with lawyers... It does something to your brain, I think.

04:21:13 <AaronSw> I guess reading lots of legal briefs hasn't helped much either.

04:21:54 <DanCon> do you just read them assuming they're written in english? or do you have some training?

04:22:15 <AaronSw> they parse pretty well as english

04:22:49 <DanCon> yeah, but I suspect certain words have more than their usual meaning.

04:23:17 <DanCon> like "friend of the court"

04:23:33 <AaronSw> Well, they normally say that in latin...

04:24:52 <AaronSw> amici curiae

04:26:12 <AaronSw> I'm surprised how well this Judge understand the technology.

04:26:34 <AaronSw> Other things I've read start off like "the internet is a network of computers from which a 'web page' can be 'downloaded'"

04:28:06 <DanCon> I was surprised too.

04:28:23 <DanCon> gives one hope about things like the tag issue on deep linking

04:28:51 * DanCon was just catching up on useit.com and found that Nielsen wrote recently on deep linking

04:35:35 * DanCon fixes uripath.refTo() bug

04:36:17 <AaronSw> have you considered submitting uripath for the python std lib?

04:36:35 <DanCon> yup

04:36:51 <DanCon> I talked with guido about that at python10

04:37:03 <DanCon> (python10 was before I wrote the code)

04:37:38 <AaronSw> what was the conclusion?

04:38:09 <DanCon> he didn't doubt my claim that urlparse() was buggy, and he was pretty open about patches or whatever

04:38:41 <DanCon> at this point, I owe a little more work to see if urlparse() can be salvaged, or if we just need a new module

04:39:05 <DanCon> and I should make the test cases fit the pyunit way of doing things

04:39:20 <DanCon> and I should get review from uri@w3.org

04:39:37 <DanCon> I think there might be one or two cases where this code disagrees with Roy/RFC2396

04:40:15 <DanCon> RFC2396 doesn't agree with the invariant that join(x, refTo(x,y)) == y

04:40:39 <AaronSw> what's refTo?

04:40:51 <DanCon> ... cuz it allows wierd URIs like http://x/foo/../bar

04:41:11 <DanCon> refTo computes a relative URI reference from here to there

04:41:53 <DanCon> e.g. when cwm is writing out URIs

04:42:02 <AaronSw> ah, split?

04:42:15 <DanCon> no, split just whacks off the #fragid bit

04:42:55 <AaronSw> That's an interesting uriref sublety I hadn't thought of...

04:43:17 <DanCon> when cwm writes <http://xyz/abc/def/ghi> to http://xyz/abc/def/foo.rdf, it just writes <ghi>

04:43:37 <AaronSw> right, i know what you mean

04:44:08 <DanCon> it really sucks that XSLT doesn't grok relative URIs at all

04:44:21 <deltab> imho "join" isn't a good name for "absolutize relative URL in relation to absolute URL"

04:44:34 <deltab> but I'm not sure what is

04:44:47 <DanCon> I'm not completely happy either

04:45:06 <DanCon> pydoc uripath shows it: [[

04:45:07 <DanCon> The name of this module and the functions are somewhat

04:45:07 <DanCon> arbitrary; they hark to other parts of the python

04:45:07 <DanCon> library; e.g. uripath.join() is somewhat like os.path.join().

04:45:08 <DanCon> ]]

04:45:47 <AaronSw> rfc2396 calls it resoltuion

04:45:58 <DanCon> yeah, like that won't confuse anybody.

04:46:18 <AaronSw> :-)

04:46:53 <deltab> someone was complaining recently about os.path.join's behaviour when the second path starts with /

04:47:17 <AaronSw> Hm. Is there anyway to do CVS repository access-control without giving everyone an account on my server?

04:47:42 <deltab> AaronSw: pserver

04:48:21 <DanCon> at w3c, we use ssh accounts that have a dummy shell

04:48:43 <AaronSw> what kind of dummy shell? i tried to give people rbash but then cvs doesn't work

04:48:53 <AaronSw> can you commit with pserver?

04:48:58 <deltab> yes

04:49:08 <DanCon> hmm... I do seem to recall some funnybusiness with the cvs shell we use

04:49:15 <DanCon> dunno the details

04:49:41 * DanCon popped back from uripath debugging to getting my quicken data in RDF... is pretty happy with it now... maybe ready to start converting to gnucash's format with XSLT

04:50:10 <DanCon> this form of RDF for my quicken data is compatible with dbview

04:50:29 <DanCon> hmm... well, aims to be, anyway

04:55:19 <DanCon> is uuidgen in the standard Java API? anybody know?

04:55:30 * DanCon would like to call it from XSLT

04:58:35 <mnot> hmm. or.w3c.util.UUID is there, and it refers to (doesn't subclass, tho) java.rmi.server.UID

04:58:55 <mnot> s/or./org//

04:59:08 * DanCon will probably just post-process using python

04:59:24 <DanCon> target output format:

04:59:24 <AaronSw> you need to stop torturing yourself with xslt

04:59:25 <DanCon> <gnc:transaction version="2.0.0">

04:59:26 <DanCon> <trn:id type="guid">f1ab62efca8cc08830eb957f14300af3</trn:id>

04:59:57 <DanCon> hey, this time I'm not doing something crazy like URI parsing with XSLT...

05:00:06 <DanCon> I'm doing an XML->XML transform.

05:00:09 * mnot groans

05:00:18 <AaronSw> progress...

05:00:24 <DanCon> input format:

05:00:25 <DanCon> <rdf:Description rdf:about="trans#_1">

05:00:25 <DanCon> <fin:txAcct rdf:resource="account#Commerce_B_Sav"/>

05:04:06 <AaronSw> hm, having my CVSROOT passwd file on the web probably isn't very smart, i sit

05:05:35 * mnot is reminded of yet another unfinished project... RSP (RDF Server Pages ;)

05:05:48 <DanCon> RDF Server pages... sounds like redfoot

05:06:23 <mnot> never looked too deeply into redfoot, was more interested in redland... will go back and take a look

05:06:43 * DanCon might be confusing redland and redfoot

05:06:52 <DanCon> no, I think it's redfoto

05:06:54 <DanCon> redfoot

05:06:55 <DanCon> phpht

05:07:03 <mnot> cool... redfoto

05:11:40 <AaronSw> ugh, CVS is a security nightmare

05:11:58 <AaronSw> paging djb ;)

05:12:48 <DanCon> yes, CVS is a security nightmare. v. painful here at W3C

05:13:15 <deltab> considered using something else?

05:13:20 <DanCon> well, nothing in CVS is actively a security problem, but you naturally want to do more than it supports

05:13:33 <DanCon> else: not seriously. We track subversion at a distance.

05:13:44 <AaronSw> a lot of things seem like an active security problem

05:13:56 <AaronSw> commit access gives users access on the machine, anonymous checkout gives users some access to the machine

05:14:25 * DanCon doesn't see how commit gives any more access than is obvious

05:16:05 <AaronSw> well it may be obvious but it's still annoying that i have to give them user-level accounts...

05:17:37 <DanCon> that seems to go under the "you naturally want to do more than it supports" category. It was built to handle multiple users on a unix box, not zillions of users over the Internet

05:17:50 <AaronSw> Ah, I suppose.

05:22:13 <DanCon> ok, I've got this far:

05:22:14 <DanCon> <gnc:transaction version="2.0.0">

05:22:15 <DanCon> <trn:id type="guid">trans#_1</trn:id>

05:29:49 <DanCon> hmm... this isn't as straightforward as I was thinking...

05:30:11 <DanCon> in quicken, each account has N transactions which have N splits.

05:30:29 <DanCon> In gnucash, each transaction has N splits, and each split has one account.

05:30:52 <DanCon> so if you transfer $100 from checking to savings, that's 2 quicken transactions, but one gnucash transaction

05:32:29 <DanCon> so I have to make 2 passes: first, go over the quicken splits, creating one gnucash transaction for each group of related splits

05:32:49 <DanCon> then go over that list of transactions and fill in the details. Hmm... maybe it's one pass after all.

05:33:14 <DanCon> nope; 2 passes

09:39:12 <dajobe>http://www.w3.org/2000/04/maillog2rdf/log2rdf.pl

09:39:13 <dc_rdfig> B: http://www.w3.org/2000/04/maillog2rdf/log2rdf.pl from dajobe

09:39:21 <dajobe> B:|maillog2rdf - email log to RDF

09:39:46 <dc_rdfig> Titled item B.

09:39:47 <dajobe> B:dug out of my bookmarks; more info for rss-dev discussions on rss & rdf?

09:39:47 <dc_rdfig> Added comment B1.

09:55:17 <dajobe> B1:dug out of my bookmarks; more info for rss-dev discussions on rss 1.0/rdf and email?

09:55:18 <dc_rdfig> Replaced comment B1.

10:03:27 * chaalsNCE gets closer to making my flight data available in SVG.

10:03:45 <chaalsNCE> So far: reads short form output

10:04:08 <chaalsNCE> ( 9. UA 863 H 12AUG SFOMEL HK1 2305 *0930 O* E MO

10:05:02 <chaalsNCE> and figures out the date (ISO 8601) and whether I arrive on a later day (that's crude at the moment - it would happily let me arrive on the 32nd of december 2002)

10:05:24 <chaalsNCE> then it goes to Jim Dean's thing and gets airport lat/long.

10:05:45 <chaalsNCE> (That needs to be optimised so it doesn't ask for data it has already)

10:06:11 <chaalsNCE> next thing to do is put it in some publicly visible place so people can complain about my coding :(

10:08:34 * JibberJim waits for that... and will likely re-implement it all in js (which no-one elses uses so I don't get any coding complaints...)

10:15:35 * chaalsNCE notes that "that's the nest way to code it in ruby" (and if that doesn't work I can always re-write it all into Logo...

10:15:39 <chaalsNCE> ;-)

10:41:39 <rillian> newbie n3 question. how does one do qualifiers, e.g. for dublincore elements?

10:43:22 * sandro looks up dc qualifiers

10:44:04 <sandro> do you know how to do them in RDFXML?

10:45:07 * sandro looks at http://logicerror.com/dcrdfDraft

10:45:36 <sandro> sorry, it's a little over my busy head.

11:18:19 <chaalsNCE>http://www.w3.org/2002/08/flight/crunchflights.rb

11:18:19 <dc_rdfig> C: http://www.w3.org/2002/08/flight/crunchflights.rb from chaalsNCE

11:18:37 <chaalsNCE> C:| A ruby script to process flight information

11:18:38 <dc_rdfig> Titled item C.

11:19:35 <chaalsNCE> C: Makes RDF from data in format

11:19:35 <dc_rdfig> Added comment C1.

11:19:42 <chaalsNCE> C: 9. UA 863 H 12AUG SFOMEL HK1 2305 *0930 O* E MO

11:19:42 <dc_rdfig> Added comment C2.

11:20:09 <libby> chaals, you got a sample output?

11:20:18 <chaalsNCE> C: and then searches Jim Deans stuff at http://www.daml.org/cgi-bin/airport? to get location of airports

11:20:18 <dc_rdfig> Added comment C3.

11:21:16 <libby> Mike dean?

11:21:41 <chaalsNCE> that's a question of ontology ;-)

11:22:28 <chaalsNCE> C3: and then searches Mike Dean's stuff at http://www.daml.org/cgi-bin/airport? to get location of airports

11:22:28 <dc_rdfig> Replaced comment C3.

11:24:40 <chaalsNCE> C: [sample output|http://www.w3.org/2002/08/flight/out.rdf] for CVS version 1.2 is available

11:24:40 <dc_rdfig> Added comment C4.

11:24:56 <libby> :) cool

11:25:26 <chaalsNCE> C: uses DanC's schema but possibly incorrectly, and has some stuff I made up I think.

11:25:26 <dc_rdfig> Added comment C5.

11:25:50 <chaalsNCE> C: fair number of todo's listed in the source, along with the one-line documentation

11:25:50 <dc_rdfig> Added comment C6.

11:26:52 * chaalsNCE wonders if jibberjim has read the todo list yet...

11:27:18 <pixel-away> pixel-away is now known as pixel

11:27:47 * chaalsNCE is really not a ruby programmer. I started trying to be one yesterday, so this is what you get ;_)

11:27:52 <JibberJim> yep, fortunately it's at the end of the todo's, so I don't have to spend 25 minutes of my life yet...

11:28:03 <chaalsNCE> lol

11:32:57 <chaalsNCE> danbri?

11:33:21 <chaalsNCE> (anyone here who knows a litttle ruby? Looking for XML modules...)

11:33:43 <JibberJim> sample output in C4 has undefined Namespace NS1 and NS0

11:34:57 <chaalsNCE> yep. That's why I am looking for XML stuff. It gets those from Mike Dean, and their namesapce is the one declared as airport in the output.

11:35:11 <chaalsNCE> so I am working on a way to fix it ;-)

11:35:28 <libby> there's a ruby channle right? alsogoogle ruby pragmatic programmer xml works well

11:36:33 <shellac>http://www.geocities.com/danbfan/index.html

11:36:33 <dc_rdfig> D: http://www.geocities.com/danbfan/index.html from shellac

11:37:02 <shellac> D:|A Dan Brickley fan website

11:37:03 <dc_rdfig> Titled item D.

11:37:40 <JibberJim> D: "Dan was seen cavorting with Jennifer Loove Hewitt at the Hannibal premiere

11:37:40 <dc_rdfig> Added comment D1.

11:37:48 <libby> A:not _the_ dan brickley, unfortunately

11:37:48 <dc_rdfig> Added comment A3.

11:38:04 <chaalsNCE> s/unfortunately/fortunately/

11:39:00 <shellac> D:Found by accident using Deja

11:39:00 <dc_rdfig> Added comment D2.

11:40:16 <deltab> D: "Dan was spotted dancing with [http://www.google.com/jobs/brtiney.html|Britney Sppears] at SPA after her appearance on Saturday Night Live"

11:40:17 <dc_rdfig> Added comment D3.

11:40:38 <deltab> D3: "Dan was spotted dancing with [http://www.google.com/jobs/britney.html|Britney Sppears] at SPA after her appearance on Saturday Night Live"

11:40:38 <dc_rdfig> Replaced comment D3.

11:40:42 <shellac> I wondered why he was so keen to get to the states...

11:40:46 <deltab> oh the irony

11:41:40 <shellac> D:Explains why foaf:Name is not a daml:UnambiguousProperty

11:41:40 <dc_rdfig> Added comment D4.

11:41:55 <shellac> there - that makes it more edumacated

11:45:20 <JibberJim> Rats, Geocities don't let you link direct to a photo, so we can't tell foafbot it's a picture of Dan Brickley...

11:46:17 <Talliesin> hmm.

11:47:27 <libby> bummer, save it somewhere

11:48:08 <JibberJim> I don't want pictures of half naked Dan Brickley's on my site, people will think me weird..

11:48:31 <deltab> heh, there's a guestbook:

11:48:32 <deltab> Comments : i read on rosie.com that dan modeled for calvin klein.....is

11:48:32 <deltab> that true?

11:48:33 <chaalsNCE> could be worse. Apparently sometimes he takes all his clothes off...

11:49:38 <libby> erg, jim, see what you mean

11:51:01 * chaalsNCE goes the stupid learner's path and writes his own routines for things that probably exist already.

11:51:43 * shellac gives in and looks at photots

11:51:48 <JibberJim> chaalsNCE, can you stick up a map of the world xplanet just gives me an error in cygwin, it seems the download expects a different version...

11:52:12 <shellac> y'know , he's been working out...

11:52:19 * chaalsNCE doesn't have xplanet yet...

11:52:40 <chaalsNCE> hair dirty blonde?!?!!? that's another way of saying salt and pepper?

11:53:47 <chaalsNCE> Hmmm. anyone got xplanet?

11:54:20 <larsbot> not sure I would trust this http://www.daml.org/cgi-bin/airport?FBU thing

11:54:35 <larsbot> anyone trying to land at that location would get a nasty surprise

11:54:57 * chaalsNCE is perpared to trust stuff until proven wrong - it is quicker than making it ;-)

11:55:02 <JibberJim> If the default xplanet image is mercator, I think it's okay as I've got one at http://jibbering.com/imgs/earth.jpg

11:56:12 <chaalsNCE> what's the surprise?

11:56:53 <chaalsNCE> go on, spoil it for us!!

11:57:58 <chaalsNCE> jim: looks to me like a mercator projection

12:00:14 <JibberJim> so latitude/longitude should be in the airport namespace?

12:02:05 <larsbot> or at least they should only refer to existing airports

12:02:26 <larsbot> FBU was closed years ago, and the place is now taken over by IT companies...

12:02:56 <chaalsNCE> some might argue that it isn't such a bad idea to land in airports that don't exist anymore ;-)

12:03:40 * chaalsNCE is relying on it to get data for a given airport - the airports being landed at are from commercial airline timetables.

12:03:53 <larsbot> in that case it should be ok :)

12:04:05 <chaalsNCE> jim: right. But now I am factoring out the whole lookup of the airport...

12:11:55 <deltab> A3:""

12:11:56 <dc_rdfig> Deleted comment A3.

12:12:03 <deltab> sorry Libby

12:13:22 <libby> ?

12:13:51 <deltab> your comment was attached at A, not D

12:16:56 <JibberJim> hmm... Chaals is either flying to Baghdad or my lat/lon calcs are somewhat off... http://jibbering.com/2002/8/flight.svg

12:22:43 * chaalsNCE wonders what the weather is like in baghdad...

12:25:24 <chaalsNCE> you mean I have to read all the scripts and do a dry-run to get the data out of them :(

12:26:42 <chaalsNCE> your viewbox outh to be -180 90 180 -90 or some permutation of that

12:27:29 <JibberJim> Ah, yeah I can play with the viewbox...

12:27:37 <Talliesin> A:You know once upon a time patents were intended to actually encourage innovation!

12:27:37 <dc_rdfig> Added comment A3.

12:27:54 <chaalsNCE> (this is how I am going to do the SVG thing - saves doing maths ;-)

12:28:42 <chaalsNCE> why are you adding the xlink declaration in yur source? is it necesary for spec or for tools?

12:29:46 <JibberJim> There's a bug in the 1.0 dtd, which makes it non validateable without it.

12:30:04 <chaalsNCE> OK. better copy that then ;-)

12:39:49 <JibberJim> Is the RDF okay? I've got AC419 departsFrom _:n2 which is of type airport, but _:n2 doesn't have a lat long?

12:41:16 <deltab> looks to me like only the destination airports have lat/long

12:41:20 <azaroth> (How do you get the bot to read you back all the comments in a thread?)

12:42:00 <deltab> C:

12:42:01 <dc_rdfig>http://www.w3.org/2002/08/flight/crunchflights.rb

12:42:02 <dc_rdfig> A ruby script to process flight information

12:42:03 <dc_rdfig> (1:chaalsNCE) Makes RDF from data in format

12:42:04 <dc_rdfig> (2:chaalsNCE) 9. UA 863 H 12AUG SFOMEL HK1 2305 *0930 O* E MO

12:42:05 <dc_rdfig> (3:chaalsNCE) and then searches Mike Dean's stuff at http://www.daml.org/cgi-bin/airport? to get location of airports

12:42:06 <dc_rdfig> (4:chaalsNCE) [sample output|http://www.w3.org/2002/08/flight/out.rdf] for CVS version 1.2 is available

12:42:07 <dc_rdfig> (5:chaalsNCE) uses DanC's schema but possibly incorrectly, and has some stuff I made up I think.

12:42:08 <dc_rdfig> (6:chaalsNCE) fair number of todo's listed in the source, along with the one-line documentation

12:42:08 * JibberJim doesn't trust his little javascript parser so couldn't say for sure...

12:42:46 <deltab> azaroth: that what you meant?

12:42:51 <azaroth> Yep, thanks :)

12:43:16 <azaroth> Cool script.

12:44:55 * azaroth grins. You could link it to one of the airport weather services for further amusement.

12:45:27 <chaalsNCE> (oh yeah, not every airport has data yet. But I am about to check in a new version along with a new script. Then no airport will have lat-long

12:45:37 <JibberJim> no airport :-(

12:45:52 <chaalsNCE> but you willl be able to copy the new script to hunt those down. By and by I'll factor again so you can get the right pieces ;-)

12:48:56 <DanCon> C:hmm... goesunto isn't in the travelTerms vocab. I use cyc:toLocation

12:48:56 <dc_rdfig> Added comment C7.

12:51:08 <DanCon> C:perhaps it's time to document the travelterms a little better

12:51:08 <dc_rdfig> Added comment C8.

12:55:56 <chaalsNCE> OK, will edit to use cyc:toLocation (and about to commit new sample outputs, so good timing ;-)

12:58:01 <chaalsNCE> DanC: the opencyc.sourceforge namespace in swap/pim/travelterms?

13:01:17 * chaalsNCE notes fromLocation as well...

13:21:21 <DanCon> no, the opencyc namespace is in opencyc.org

13:21:36 <DanCon> my($kNS) = "http://opencyc.sourceforge.net/daml/cyc.daml#";

13:22:13 <JibberJim> chaals, changing the viewbox like that hugely hurts performance.

13:57:21 * chaalsNCE waits for my computer to stop complaining about being full and start working again :(

13:59:48 <azaroth> Computers always work best when hungry.

13:59:57 <chaalsNCE> jim: OK, well, all I need to do is make the lat/long relate to a 1000x2000 box centered at 1000,500

14:00:18 <chaalsNCE> Or is it fast enough to use a transform (scale and translate)

14:01:26 <JibberJim> any time it scales or translates raster images it's slow, so I guess you could translate the airport vector bits on top of an untranslated image

14:06:34 <JibberJim> Ah, yeah, that works, just define the viewbox for the vectors.

14:14:58 <j|out> j|out is now known as jordan

14:15:41 * chaalsNCE gets over a stupid process problem...

14:22:12 <chaalsNCE> C: added [an SVG output|http://www.w3.org/2002/08/flight/mapflights.rb] that doesn't really work yet, and moved the bit about fetching latitudes/longitudes into it

14:22:12 <dc_rdfig> Added comment C9.

14:23:12 <chaalsNCE> C: [sample output of the map|http://www.w3.org/2002/08/flight/map.svg] (currently only useful to see where I am going - everything is in a comment) also available

14:23:12 <dc_rdfig> Added comment C10.

14:24:37 <chaalsNCE> DanCon I'm confused. Shoould I use xmlns:cyc='http://opencyc.sourceforge.net/daml/cyc.daml#' or something else?

14:25:01 <jordan> jordan is now known as jordan|mva

14:25:58 <chaalsNCE> C7:""

14:25:58 <dc_rdfig> Deleted comment C7.

14:26:02 <JibberJim> shouldn't the RDF be in metadata elements, rather than comments?

14:26:21 <chaalsNCE> Yeah, and the path data should make paths. I'm working on it ;-)

14:26:52 <chaalsNCE> Some stuff should be in title/descs, too. And it would be nice to have lots of other cool stuff ;-)

14:27:20 <JibberJim> and a little plane which tracks along the paths...

14:27:48 <chaalsNCE> that was already mentioned at lunch

14:28:00 * JibberJim doesn't get to have lunch :-(

14:28:17 <chaalsNCE> But it would be good to have information about the destinations linkable

14:28:42 <JibberJim> Once the RDF format is stable, it'll be easy enough for me to do that client-side...

14:29:12 <chaalsNCE> (Hey! I'm going to LPL! Where is that, are there any people I know there, and what's fun to do?)

14:29:46 <chaalsNCE> This is the sort of thing I do mentally as I sit in the travel agents office looking at 40 different possible flight paths. It would be nice to have it laid out in a friendly way ;-)

14:30:31 <JibberJim> hmm.. LPL - well you could hire a car and watch people nick the wheels...

14:30:45 <chaalsNCE> My travel agent wanted to show people this. I think if she thought about it she would want to be able to get variants on routes available - a sort of graphic interface to SABRE

14:31:03 <chaalsNCE> Nah, I can do that at LON and I go there all the time

14:40:31 <JibberJim> So is there a vocabulary for describing where a particular airline flies between?

14:40:54 <AaronSw> lol. PFPS to sbp: "if you count amongthe key principles of the Semantic Web a lack of both syntax and semantics,then I'll go along with this statement, but I would hope that the SemanticWeb would not continue to espouse these principles."

14:40:56 <chaalsNCE> well, I am now using the cyc:fromLocation and cyc:toLocation

14:41:19 <chaalsNCE> (possibly from the wrong namespace ;-)

14:41:32 <chaalsNCE> they are used by the Flight term in DanCons travel terms

14:41:59 <JibberJim> but is travelTerms:Flight defined just for a single flight, or can I use it for just saying some airline flies from A to B ?

14:42:40 <chaalsNCE> I think it is defined for a single flight. You could always define a route term if there isn't one already (or even if there is....)

14:45:24 * chaalsNCE likes jim's new map...

14:45:43 <chaalsNCE> and goes to shamelessly copy it ;-)

14:45:53 <JibberJim> it works in Batik now to!

14:47:14 <chaalsNCE> Cool!

14:47:33 <chaalsNCE> what is it using for source?

14:47:33 <AaronSw> jang: "I agree in principle that informal language and anthropomorphisation of the global pulsating hyperbrain or any part thereof should be avoided in specs."

14:47:43 <chaalsNCE> (still my out.rdf file which I have now changed?

14:47:47 <JibberJim> http://jibbering.com/2002/8/sample-air.rdf

14:47:48 <dc_rdfig> E: http://jibbering.com/2002/8/sample-air.rdf from JibberJim

14:47:52 <JibberJim> oops...

14:48:26 <JibberJim> E:| Sample RDF output from Chaals's Ruby flight information script

14:48:26 <dc_rdfig> Titled item E.

14:50:16 * chaalsNCE curses mozilla 1.1 for being too smart about viewing source

14:53:12 * chaalsNCE heads back into programming land...

14:55:14 <Talliesin> yes the view menu needs a "the bloody bytes, I'll work it out myself" option.

15:10:25 <sandro> How do you get cwm to work (pass its regression tests) without a /devel filesystem or putting it in the root filesystem?

15:13:45 <AaronSw> you manually inspect such errors

15:14:43 <sandro> <bzzzt> wrong answer. manually inspecting for errors is verboten. (but thanks for playing.)

15:14:55 * sandro find an old unused filesystem and makes it /devel.

15:15:05 * sandro should fix the test suites instead.

15:15:09 <AaronSw> this game show is hard

15:15:22 <AaronSw> the test suites can't really be fixed because of xmlns relative uri sillyness, i think

15:15:39 <AaronSw> i guess you could add some intelligence to detect such things...

15:16:03 <AaronSw> but cwm test suites always need manual inspection because of whitespace changes and such, no?

15:16:14 <sandro> I think it would be right for the test runner to know the current directory string and replace that text in the reference and generated outputs with *CURRENTDIRECTORY* or some such.

15:16:33 <sandro> the diff is done without looking at whitespace changes, I think.

15:16:56 <sandro> The "no manually inspecting for errors" is an Extreme Programming rule, btw.

15:17:10 <AaronSw> ok, ordering changes then

15:17:23 <sandro> Yes, indeed. That's an issue.

15:17:47 * bijan has on his list a sane N3 test suite

15:17:56 <sandro> I was looking at passing with the test suite with XSB and it looks like a nightmare because of such idiosynchracies.

15:17:58 <bijan> suite/mechanism

15:18:08 <sandro> Indeed, bijan.

15:18:20 <sandro> cwm is not really the right tool, IMNSHO.

15:18:34 <sandro> but I guess that's an asside.

15:18:48 <bijan> Starting place is some graph diff/comparison tool

15:18:57 <AaronSw> I was thinking of starting a group to do some N3 spec and test suite writing

15:19:17 <rillian> hi AaronSw

15:19:19 <bijan> At least introduce some preliminary sanity to what you output.

15:19:22 <AaronSw> rillian!

15:19:37 <AaronSw> what brings you to our merry land?

15:19:40 <bijan> Part of the problem, of course, is that you want to verify "logical" correctness, not just actual output.

15:19:49 <AaronSw> did ogg ever pick a metadata format? I need to research this for my creativecommons work

15:20:16 <bijan> But graph isomorphism will do for a start :)

15:20:26 <rillian> AaronSw: we never did, but I've been thinking about it again

15:20:35 <rillian> 's why I'm in your merry land :)

15:20:40 <bijan> (needs to be a spec :()

15:20:47 <bijan> (i.e., there needs to be one)

15:20:55 <AaronSw> An N3 spec?

15:21:05 <bijan> Yep.

15:21:11 <DanCon> AaronSw, you seem to be "back". I got wind of some sort of chat-break for you, but I couldn't trace it to any notice from you.

15:21:30 <AaronSw> i'm not back until 9/11

15:21:40 <AaronSw> but i keep cheating :/

15:21:50 <AaronSw> sbp dared me to take two months off of IRC

15:21:58 <DanCon> ah.

15:22:02 <JohnE> bijan - I agree, an "official" n3 spec is needed

15:22:03 * bijan knows it's experimental, et al, but given that we have multiple implemenations it might be nice to have some way of stating agreement

15:22:12 <DanCon> "sane N3 test suite"? the existing test suite isn't sane?

15:22:31 <AaronSw> can I add JohnE and bijan to my list of people willing to contribute to the n3 specwriting effort?

15:22:37 * DanCon has another obligation, and really should leave, but finds it hard to resist #rdfig when it's busy like this

15:22:41 <sandro> The cwm test suite has a lot of cwm specific stuff, like the order of output triples.

15:22:58 <JohnE> please do

15:23:02 <jang> dancon: "Datatyping - is it a verb?"

15:23:05 <jang> it's a gerund.

15:23:15 <AaronSw> ba dum bum - chish

15:23:22 <sandro> Is that like strong typing -- when you nearly break your keyboard?

15:23:36 <DanCon> order of output triples... that's sort of a limitation of the testing technology.

15:23:37 <bijan> DanCon: I haven't trolled through it in a while, but last time I tried to use it for CWMClone development it wasn't 1) usefully organized, 2) clear what passing it meant

15:23:56 <DanCon> re 1: did you find retest.sh?

15:23:57 <bijan> DanCon: right, hence my amendation to sane test mechanism"

15:24:06 <sandro> yah.

15:24:10 <bijan> I believe so (it has been a while)

15:24:20 <DanCon> er... diff is limited, but I don't see why you call it insane

15:24:43 <DanCon> retest.sh is a pretty obvious enumeration of the tests, no? that's not sufficiently useful?

15:25:46 <bijan> I didn't find it to be, ymdv

15:26:01 <AaronSw> I think bijan's complaint is that the tests are a bunch of "this should work but didn't once" rather than "these are all the things a good processor should do"

15:26:13 <DanCon> I mentioned bijan's suggestion to redo retest.sh in python to timbl; his response was "I'll probably do it in N3". That seems to meta for my engineering tastes.

15:26:24 <sandro> mine, too.

15:26:27 <bijan> Grouped by feature, preferably.

15:26:32 <bijan> Oh yeah, with y'all there.

15:26:58 <DanCon> aaron, the test suite includes feature tests as well as regression tests. That's a problem?

15:27:15 <sandro> It would be nice if most of the tests were in terms of input formula and output formula, instead of input text and output text. With some formula-equivalence (isomorphism) rules.

15:27:27 <bijan> Also, as I said, i'd prefer a graph isomorphism check than something serialization specific, but I guess a taste/mechanism issue

15:27:37 <bijan> heh.

15:27:42 <DanCon> er... formula-equivalence is what many of the tests test.

15:28:16 <bijan> Hmm. We've all looked at these things at one point or another...interesting that our impressions are so divergent.

15:28:20 <bijan> Must be a documentation problem :)

15:28:22 <DanCon> yeah, well, if there were a tool with the maturity of diff that did graph isomorphism testing, we'd use it.

15:28:27 * sandro was reading XP Explained last night -- the second on how to transition to XP and adding tests as needed. The current tests do meet that criterion. They just are not factored, so if we improve the n3 pretty-printer, all the tests break.

15:28:44 <chaalsNCE> well, my map works excpt that I got lat/long confused....

15:28:49 <bijan> DanCon: right, that's what I'd lke, such a tool

15:28:51 <DanCon> cool, chaalsNCE

15:29:04 <bijan> I've been casually investigating such things; just not had time to push on it.

15:29:37 <DanCon> ok, but much of the machinery in cwm *is* such a tool; i.e that's what we're testing. I suppose the tests for the graph-equivalence code could be done using diff and the other tests could use the graph-matching code.

15:30:54 * JibberJim Is generating some RDF of the routes from LON for the various low cost carriers to put on his map.

15:30:56 <sandro> Note that cwm does not input or output graphs, in general, it inputs and outputs formulas. A graph matching tool could not handle formulas, so it wouldn't work on many of the tests.

15:31:06 <DanCon> that too

15:31:32 <DanCon> I'm not convinced the cwm/n3 tests are insane, but I agree they're immature and less than ideal.

15:32:04 <sandro> ... and it would probably be insane to use them as your only spec in implementing a clone. :-)

15:32:15 <bijan> heh.

15:32:28 <DanCon> hmm... I'm not sure.

15:32:42 * bijan isn't attached to calling them insane or anything other than DanCon's description

15:32:50 <DanCon> the N3 pretty-printer actually does have some rationale for the way it prints things. It's designed to be canonical, not arbitrary.

15:33:12 <AaronSw> Yes, TimBL has an interesting recursive spec for it i heard him explain to jjc at www2002

15:33:40 * DanCon really must dash now; might be back in a bit

15:34:21 <AaronSw> rillian, can i suggest a requirement that whatever metadata format you pick supports assertions of the type "this file is licensed under license <URL>"?

15:39:56 <JibberJim> Chaals, that's a very fat line you're flying on...

15:40:23 * JibberJim also wonders why the pilot doesn't go the other way, it'll be a lot quicker...

15:41:20 <bijan> Are xlink link types extensible?

15:42:29 <chaalsNCE> C: Worky!! [a live SVG map of the sample data|http://www.w3.org/2002/08/flight/map.svg] does the right thing!

15:42:30 <dc_rdfig> Added comment C10.

15:42:57 * chaalsNCE tries to feed it new data...

15:43:46 <JibberJim> Live?

15:44:06 <danbri> bijan, there's supposed to be a mapping of XLink to RDF (see Ron Daniel note on harvesting rdf from xlink) so yes...

15:44:14 <danbri> .google xslt rdf harvesting ron

15:44:16 <datum> xslt rdf harvesting ron: http://www.w3.org/TR/xlink2rdf/

15:44:41 * AaronSw compiles new version of the media-type draft

15:45:14 <bijan> Ooo, that's harverstign *from* xlink

15:45:30 <bijan> That's doesn't immediate suggest that you can go from arbitrary rdf to Xlink

15:45:43 <chaalsNCE> well, it probably still doesn't really reflect the flight path. But what the heck...

15:45:46 <bijan> And even if so, such an encoding needn't use extendsible types, right?

15:46:02 <Talliesin> what mean by "extensible" here?

15:46:07 * chaalsNCE hammers daml.org for more aiport data. Must optimise that bit soon

15:46:08 <bijan> I.e., add new types.

15:46:19 <bijan> AFAICTT, Xlink has a fixed set of link types.

15:46:22 <danbri> i thought it explained where in xlink syntax the predicate URI lives

15:46:22 <bijan> but I want to confirm.

15:46:43 <danbri> no, should be various app specific link types, qname'd I thought

15:46:46 <deltab> The value of the type attribute must be supplied. The value must be one of

15:46:46 <deltab> "simple", "extended", "locator", "arc", "resource", "title", or "none".

15:46:59 <deltab> -- http://www.w3.org/TR/xlink/#link-types

15:47:08 <bijan> Right. Looks like arc role is what the mapping is using.

15:47:13 <danbri> they're types of the link syntax, not of the relationship expressed...

15:47:18 <JibberJim> does daml.org provide the airport data for download?

15:47:22 <danbri> yeah, i'm confused easily by their terminology

15:47:28 <Talliesin> Well it depends on what you mean by "type".

15:47:32 * JibberJim is also hammering them.

15:48:00 <Talliesin> At one level of abstraction 2 xlinks with different role URIs are different "types".

15:48:09 <Talliesin> but that's not what xlink:type means.

15:48:37 <bijan> Right.

15:49:18 <mhgrove> mhgrove is now known as mike-lunch

15:49:23 <bijan> arcrole does it for me.

15:49:24 <deltab> html:rel is a different sort of type, right?

15:49:38 <Talliesin> loose words like "type" quickly lead to situations where I'm not sure if I'm misunderstanding, misunderstood, or just plain dumb

15:50:04 <chaalsNCE> C: [a second dataset that is real flight info|http://www.w3.org/2002/08/flight/2map.svg] shows there are still problems - when there is a return leg on the same path it isn't clear...

15:50:04 <dc_rdfig> Added comment C11.

15:50:27 <Talliesin> okay, if you are considering html:link with different rel to mean different type, then different xlink:role or xlink:arcrole makes different type as well.

15:51:39 <JibberJim> make it a curveTo chaals, then a->b looks different to b->a

15:51:48 <Talliesin> chaalsNCE, your pilots flying along the great circle?

15:52:44 * Talliesin would be interested in how easy calculating the geodesic path would be in SVG

15:52:58 <chaalsNCE> Great Circle: Probably Not. They are flying around politicaly sensitive boundaries, so the actual lie is just between the intersting points

15:53:10 * danbri sees some red SVG lines; a map downloads some time later. coool :)

15:53:27 <chaalsNCE> jim: I think I'll actually mark that there is a return flight (different type of line...)

15:54:00 <chaalsNCE> But to do that I need some real program stuff - at the moment it is all horrid regexp hacks

15:54:12 * chaalsNCE likes to get instant gratification every now and then...

15:54:15 <JibberJim> have you got the RDF?

15:54:37 <JibberJim> Ryanair's destinations from STN: http://jibbering.com/2002/8/ryanair.svg - from http://jibbering.com/2002/8/ryanair.rdf

15:55:01 <rillian> AaronSw: that's certainly an option. why should it be required?

15:55:35 <AaronSw> er, I meant a requirement for the format that you choose to support/use

15:55:42 <AaronSw> not a requirement on the users

15:55:57 <rillian> ah. sure

15:56:06 <rillian> does musicbrainz have a tag for that?

15:57:14 * rillian also notes that even the built-in vorbis comment structure has canonized tags for COPYRIGHT and LICENSE

15:57:16 <AaronSw> I don't think so, but we do: http://creativecommons.org/metadata/

15:57:28 <rillian>http://xiph.org/ogg/vorbis/doc/v-comment.html

15:57:29 <dc_rdfig> F: http://xiph.org/ogg/vorbis/doc/v-comment.html from rillian

15:57:42 <chaalsNCE> there is an rdf file at http://www.w3.org/2002/08/flight/2out.rdf that was used to generate the second map.

15:57:46 <AaronSw> Ooh, it even mentions us. :-)

15:58:10 <AaronSw> F:|Ogg Vorbis I format specification: comment field and header specification

15:58:10 <dc_rdfig> Titled item F.

15:58:33 <rillian> F: of particular interest is the list of canonized tags

15:58:33 <AaronSw> Are you looking for a metadata format for vorbis or tarkin or that new thing they have or all three or what?

15:58:33 <dc_rdfig> Added comment F1.

15:58:44 <rillian> all three

15:59:24 <rillian> ideally anyway

15:59:39 <rillian> AaronSw: so cc is your thing?

15:59:53 <AaronSw> I'm the Creative Commons metadata guy

16:00:45 <AaronSw> Hm, I know that Rob Kaye of musicbrainz wanted to reflect all the ID3 tags into RDF, I wonder what happened to that.

16:01:25 <rillian> dunno

16:01:34 <rillian> the id3 system isn't especially well designed

16:02:24 <rillian> but there are fairly obvious mappings in the common case

16:03:45 <rillian> so what's the status of RDF/XML. I was reading about n3 this morning

16:03:59 <rillian> and consensus seemed to have moved toward ignoring it :)

16:04:31 <AaronSw> RDF/XML is an ugly bloated piece of junk. N3 is a reasonably nice syntax for humans. RDF/XML is the standard.

16:05:00 <AaronSw> [speaking only for myself]

16:05:13 <rillian> (and serveral others, almost word-for-word :)

16:05:38 <rillian> one of the reasons I've been wary of adopting something for Ogg is the flux in the standard

16:05:48 <dajobe> "N3 is the line noise of rdf syntaxes" - how about that AaronSw

16:05:49 <rillian> no one likes rdf/xml, so it might well go away

16:05:56 <rillian> everyone likes n3, but it's not mature

16:05:59 <dajobe> oh, more FUD

16:06:04 <dajobe> bring it on

16:06:09 <AaronSw> Well, there's one thing that hasn't changed over time which is the (URI, URI, Literal) pattern

16:06:17 <Talliesin> I like rdf/xml!

16:06:41 <AaronSw> "and in the left corner we have Dave Beckett from the no-fun-thorugh-backwards-compatibility-squad!"

16:06:47 <rillian> AaronSw: well, yes. I've never had a problem with RDF

16:06:49 * azaroth casts his uninformed support for rdf/xml from what he's seen of it

16:07:02 <rillian> the question was what representation and interpretation conventions to specify

16:07:19 <dajobe> yes, to thorough backwards compatibility

16:08:00 * DanCon thinks these handspring treo 90 PDAs look pretty cool

16:08:20 <Talliesin> last night I was mulling about the idea of a rdf-in-xml that used xpointer so that anonymous nodes could be referenced from anywhere. It would probably be insane though.

16:08:53 <AaronSw> i noticed that Ogg uses some declare-length-before-bits stuff that's probably motivated by requirements not in other formats...

16:09:04 * chaalsNCE votes rdf/xml ...

16:09:28 <chaalsNCE> interpreting RDF is something I want machines to do. SO I don't care how they represent it.

16:09:41 <AaronSw> if you replaced the vectors with a <URI><space><URI or literal> thing, i think that'd be RDF enough for me...

16:09:47 <chaalsNCE> If they show it to me I want it to be in a nice format, not some almost-readable thing

16:09:53 * Talliesin votes for digital bio-diversity - let them all be.

16:10:11 <AaronSw> Yeah, the problem with RDF/XML is that machines have a worse time interpreting it than we do, I think...

16:10:42 * JibberJim finds a simple rdf/xml parser easy to write, okay it doesn't do the obscure bits of the spec, but it's rare for me to run into anything I can't parse.

16:11:19 <AaronSw> I ran into at least 12 documents I couldn't parse with a fully spec-compliant parser the other day.

16:11:28 <AaronSw> just going through the list of foaf files

16:11:39 <danbri> what was the problem?

16:11:48 <danbri> ns prefixing etc?

16:11:51 <AaronSw> people like you weren't qualifying their attributes :)

16:12:08 * chaalsNCE claims to be totally unlike danbri

16:12:24 <rillian> AaronSw: that's more the direction I've been leaning lately

16:12:29 <danbri> in June 2000, that was acceptable...

16:12:42 <rillian> one of the objections is that the data should be 'streamable'

16:12:49 * chaalsNCE wonders if he was acceptable in June 2000 ...

16:12:51 <danbri> So don't blame RDF/XML for being fixed recently, that's a terrible argument

16:12:52 <rillian> which is more of a hack with xml

16:12:55 * danbri doubts it

16:13:15 * chaalsNCE :P

16:13:19 <AaronSw> what's streamable mean?

16:13:51 <AaronSw> you might find a good compromise between RDF/XML and N3 in N-Triples by the way. it's both standardized and easy to parse.

16:13:57 <rillian> chopped up in little bits, sent piecewise, then reassembled and parsed as-best-we can on the other end even if there are drop-outs

16:14:22 <AaronSw> N-Triples is line-based, one statement per line

16:14:35 <AaronSw> <URI> <URI> "literal" .

16:14:41 <rillian> the arguments *for* xml is mostly that I'd really like to have markup for the lyrics/transcript/subtitle format, and if we've got an xml parser anyway...

16:15:07 <rillian> n3 does seem overcomplicated for what we need

16:15:13 <dajobe> lol

16:15:45 <rillian> most of the stuff I was reading on it this morning was about how to do reflexivity for logic, which is way more complicated that we need

16:16:02 <AaronSw> Heh, yeah

16:16:28 <chaalsNCE> Jim: It's not clear what actual routes there are...

16:16:47 <chaalsNCE> on the other hand I want to put cities onto mine (and nicer lines ;-)

16:17:07 <rillian> AaronSw: which brings me back to the question I came here to ask: how does one do qualifiers in n-triples/n3

16:17:07 <JibberJim> in mine - all the routes are from STN to those destinations...

16:17:25 <AaronSw> What do you mean by qualifiers?

16:17:37 <rillian> in the dublincore sense

16:18:08 <chaalsNCE> oh yeah, you said that.

16:18:15 <chaalsNCE> what are the different types of dots?

16:18:31 <JibberJim> yellow for RyanAir, blue for Buzz.

16:19:17 <chaalsNCE> cool.

16:19:28 <AaronSw> dublin core has two types of qualifiers... formatting qualifiers (the value is in the media type format) and constraining qualifiers (this is a dc:date that is the date of creation)

16:19:39 <AaronSw> the second is easy, you just use dc:created instead of dc:date

16:20:22 <rillian> AaronSw: no dumbing-down that way though

16:20:35 <rillian> but perhaps that's not a problem if we're rolling our own representation

16:20:39 <AaronSw> well, the dumb down rules are specified in the schema, which states:

16:20:40 <azaroth> AaronSw: Really ? That seems to defeat the purpose of qualified dc

16:20:49 <AaronSw> dc_rdfig:created rdfs:subPropertyOf dc:date .

16:20:50 <dc_rdfig> Not understood: created rdfs:subPropertyOf dc:date .

16:20:56 <AaronSw> argh, stupid expansion

16:21:06 <AaronSw> azaroth, how so?

16:21:22 <rillian> you need the schema to know how to dumb things down

16:21:32 <rillian> if you have the schema, you don't need to dumb things down?

16:21:58 <AaronSw> you could also include supPropety info in the instance data, to be safe

16:22:00 <rillian> (not really for machinability I guess)

16:22:00 <azaroth> Say I have the same name for a qualifier that can qualify multiple DC simple fields?

16:22:02 <azaroth> Then what?

16:22:29 <AaronSw> um. it's a subPropertyOf each of them?

16:22:38 <rillian> dc:date:creation?

16:23:38 <azaroth> Perhaps I'm misunderstanding from your example then, but how would you express dc:(date):created vs dc:(other):created ?

16:24:01 <AaronSw> they'd have to have different names, if they were different

16:24:21 <rillian> AaronSw: so why not dc:date:creation?

16:24:43 <AaronSw> why not name it that? because then only DC could reate new qualifiers of dc:date

16:24:46 <azaroth> rillian, AaronSw: Exactly. So why bother with the dumbing down to the qualified name? Why not dc:date:creation as per rillian?

16:25:28 <AaronSw> subPropertyOf lets any URI be a qualifier

16:26:06 <rillian> byzantine

16:26:22 <AaronSw> I feel I'm missing something.

16:26:48 <rillian> AaronSw: I'm suggesting we add structure to the uri

16:27:01 <rillian> s/structure/additional structure/

16:27:09 <rillian> not that dc define longer names

16:27:44 <AaronSw> Yeah. That's not really the RDF way.

16:28:14 <AaronSw> structure should usually be explicit in the RDF, not inferred from implementation details like the URIs chosen

16:28:46 <rillian> hmm. different format then:

16:29:02 <rillian> <uri> <uri> [<qualifier uri>] literal

16:29:17 <rillian> n-tuples instead of n-triples :)

16:29:20 <AaronSw> well then it's no longer RDF. ;-)

16:29:47 <rillian> AaronSw: I must have misunderstood then

16:30:01 <rillian> <creator qualifier="corporate">IBM</creator>

16:30:04 <AaronSw> Oh, you mean have that syntax represent the same RDF statements?

16:30:34 <AaronSw> Yeah, that bit of XML isn't RDF either.

16:31:02 <AaronSw> equivalent RDF/XML would probably be like "the creator is a something whose qualifier is corporate and whose value is IBM"

16:31:14 <rillian> <thing> dc:date my:teatime "200208211615"

16:31:21 <AaronSw> <music> dc:creator <something> .

16:31:31 <AaronSw> <something> rdf:value "IBM" .

16:31:41 <AaronSw> <something> <qualifier> "corporate" .

16:32:14 <rillian> AaronSw: ok. I'd seen my xml example written down. was that ever RDF?

16:32:30 <AaronSw> Not to my knowledge.

16:32:53 <rillian> (the funny thing here is that I disagreed with Rob about the musicbrainz schema because it didn't do enough of that. He wanted everything to be a tree)

16:33:16 <rillian> AaronSw: ok, thanks

16:33:22 <AaronSw> didn't do enough of what?

16:33:47 <rillian> web-like structure

16:35:50 <rillian> cdmp:cd contains mm:trackList, which contains mm:Track which contains track-specific metadata

16:36:01 <rillian> in a strict hierarchy

16:37:15 <AaronSw> Well, when you parse it it falls down into triples...

16:37:23 <AaronSw> cdmp:cd mm:trackList <something>

16:37:37 <AaronSw> <something> rdf:_1 <track1>

16:37:45 <AaronSw> <track1> rdf:type mm:Track

16:37:53 <rillian> I think he didn't see how to write a parser that used a graph instead of tree

16:37:56 <AaronSw> <track1> mb:artist "Danding Sombreros"

16:38:22 <rillian> or maybe just didn't want to :)

16:38:22 <AaronSw> Probably. RDF wasn't very natural to his userbase, i think

16:38:33 <rillian> anyway, I guess the answer to my question

16:38:43 <rillian> is that you do qualifiers the same way you do in rdf/xml

16:38:53 <rillian> I'd just misunderstood how you do it in rdf/xml :)

16:39:07 <AaronSw> :-)

16:41:15 <rillian> oops. n-triples specify us-ascii

16:42:10 <AaronSw> dajobe, why is that?

16:42:18 <dajobe> oh, the i18n argument

16:42:29 <dajobe> if we do it at all, we have to do it completly

16:42:42 <AaronSw> which would mean...?

16:42:45 <dajobe> and what's the content-encoding of n3?

16:42:51 <rillian> utf8

16:42:56 <dajobe> says you

16:43:15 <rillian> dajobe: so you're making the 'it's to hard to do it perfectly so we must do it wrong' argument? :-)

16:43:19 <dajobe> no

16:43:31 <AaronSw> there's only one place where utf8 would go, right? In the literals

16:43:31 <dajobe> you can encode all unicode chars in n-triples

16:43:39 <dajobe> well, what about IRIs?

16:43:49 <danbri> "N3 files are encoded in UTF-8 (See RFC2279), in normalized in Normalization Form C. The language is defined in terms of a sequence of Unicode characters. (Implementations may chose to implement using 8-bit bytes, passing bytes >7F transparently, but this will not allow them to check the validity of embedded non-ASCII characters.

16:43:49 <danbri> "

16:43:56 <danbri> in n3 home page

16:43:57 <AaronSw> well, either RDF uses IRIs or it doesn't.

16:44:04 <dajobe> is that new?

16:44:33 <rillian> what's Normalization Form C?

16:44:54 <dajobe> ok, so ntriples uses the us-ascii subset of utf-8 ;)

16:45:11 <rillian>http://www.unicode.org/unicode/reports/tr15/

16:45:11 <dc_rdfig> G: http://www.unicode.org/unicode/reports/tr15/ from rillian

16:45:25 <rillian> G:|Unicode normalization forms

16:45:25 <dc_rdfig> Titled item G.

16:45:30 <AaronSw> G:|Unicode Normalization Forms

16:45:31 <dc_rdfig> Titled item G.

16:45:57 <dajobe> I don't think [[The exact set of unicode character to be allowed in (a) literal strings or (b) identifiers is not yet clear.)]] is much help

16:46:20 <rillian> grove: N3 (now?) uses a content encoding of utf-8 with Normalization Form C

16:46:33 <rillian> G:N3 (now?) uses a content encoding of utf-8 with Normalization Form C

16:46:33 <dc_rdfig> Added comment G1.

16:46:34 <AaronSw> I think Normal Form C is that all combining characters should be replaced with single characters when possible

16:46:58 <rillian> looks like it

16:47:14 <rillian> makes it easier for the renderer, I guess?

16:49:07 <AaronSw> yeah. it also makes it easier to compare unicode strings

16:49:18 <AaronSw> Unicode is a mess with details lke this though

16:49:48 <AaronSw> I mean think of domain names. Can one person get Angstrom Sign OL.com and another get LATIN CAPITAL LETTER A WITH RING ABOVE OL.com?

16:50:28 <AaronSw> They're identical, but different unicode bytes

16:51:02 <rillian> ah, makes identity testing the same as bytesequence testing

16:51:02 <rillian> right

16:51:29 <JibberJim> They're not the same!

16:51:46 <AaronSw> Er, they look identical

16:52:35 <rillian> of course, if the web were civilized, that wouldn't be a problem...

16:53:13 <JibberJim> In a particular font you have they have the same visual representation, they are distinct characters though, just like the "Dan Brickley" in D: isn't the same as the "Dan Brickley" here.

16:54:23 <sandro> Is there an operation in llyn to just empty the store completely?

16:54:56 <AaronSw> civilizing the web might be harder than fixing unicde ;)

16:55:11 <AaronSw> JibberJim, but the dan brickleys don't look identical under proper lighting

16:55:42 * rillian has a well known impractical desire to focus on 'fixing what's broken'

16:56:00 <JibberJim> the words "Dan Brickley" do, but they represent different things, one has fansites with half naked pictures of him on, the other's a model in the US...

16:56:25 <AaronSw> which one does RDF work?

16:57:19 <JibberJim> I'm not too sure...

16:58:19 <sbp> LATIN CAPITAL LETTER A WITH RING ABOVE OL.com: check out http://www.ietf.org/html.charters/idn-charter.html

16:58:53 <sbp> the RACE Unicode encoding seems to be the leading contender at the moment for IDNs

16:59:25 <AaronSw> check out http://cr.yp.to/proto/idnc3.html

16:59:41 <AaronSw> you're not one of them 7-bits-foreve people, are you?

16:59:45 <sbp> oh, you know about DJB?

17:00:10 <sbp> argh, 6PM already

17:00:12 <sbp> Gotta run

17:07:31 <dajobe> ah, the simpsons

17:08:25 <azaroth> &quot;hell yeah&exclamationMark; who needs more than alphanumerics&questionMark; ;)

18:03:39 <AaronSw> sbp?

18:03:56 <AaronSw> I'm thinking of doing a hypertextual spec for N3

18:04:48 <AaronSw> i.e. An N3 document is made up of a series lines, each of which is a <a href="directive">directive</a> or a <a href="statement">statement</a>.

18:15:37 <AaronSw> PaulP: 'RDF/XML: "the machine efficiency of XML with the readability of ASN.1"'

18:18:18 <dajobe> yawn zzzz

18:49:14 * sandro replies to PFPS with an example of RDF layering -- talk about lack of readability!

19:22:55 <AaronSw>http://notabug.com/2002/n3/spec

19:22:56 <dc_rdfig> H: http://notabug.com/2002/n3/spec from AaronSw

19:23:02 <AaronSw> H:|Notation3 Specification [DRAFT]

19:23:02 <dc_rdfig> Titled item H.

19:23:35 <dajobe> "utf-8 bytes" hmm

19:23:48 <AaronSw> Yeah, I'm not sure about the terminology there. deltab?

19:24:04 <dajobe> the tokens are in unicode characters, encoded as utf-8 octets?

19:25:41 <dajobe> add pointers to the yapp, other grammars

19:25:53 <dajobe> and ask the grammar writers to comment :)

19:26:51 <AaronSw> oops, I need to put in the semantics.

19:27:06 <dajobe> I need a phrase for the language that is "the parts of n3 that I'd consider impl. at this point I.e not {} and funky stuff like [ := a; b c ] etc"

19:27:35 <AaronSw> :- is implementable

19:27:41 <dajobe> I didn't say it was

19:27:43 <dajobe> n't

19:28:18 <AaronSw> As I understand it, [ :- a; b c] is syntactic sugar for a b c .

19:28:27 <dajobe> sort of

19:28:34 <dajobe> it also passes back 'a' to the left hand sentence

19:28:54 <AaronSw> right

19:29:07 <AaronSw> why wouldn't you implement that?

19:30:25 <dajobe> oh, I might. Just a lots of sugar in n3

19:30:45 * danbri returns from a chair-buying expedition; Pine cheerfully tells me "[41 new messages! Most recent from Peter F. Patel-Schneider]"

19:31:00 <AaronSw> yeah, i'm worried about the weirder stuff like :-, $ and %. it is looking like perl

19:31:05 <dajobe> he he

19:31:22 <dajobe> perl (6) is refactoring it's tokens, it has run out

19:31:39 <AaronSw> there are always more unicode chaacters!

19:31:44 <dajobe> (s/it's/its)

19:32:02 <AaronSw> i sat in on a bunch of perl6 talks at OSCON, it's looking a lot more like python

19:32:54 <danbri> danbri.possessions.add( ResourceClass.new( "http://xmlns.com/wordnet/1.6", "Dan's 2nd cheapo office-chair purchase of the month"))

19:33:21 <AaronSw> "Dan's 2nd cheapo office-chair

19:33:25 <AaronSw> purchase of the month" is in wordnet now?

19:33:34 <danbri> nope, just chairs

19:35:09 <AaronSw> PFPS: "If you punt to naturallanguage descriptions of the meaning of the vocabulary, you may as wellmake an entire KB be a string."

19:35:09 <dc_rdfig> Label PFPS not found.

20:03:00 * danbri disagrees; sometimes natural language is the only tool to hand that can capture subtle corners of meaning

20:05:01 * sbp wonders what a forumla is in http://notabug.com/2002/n3/spec

20:05:27 <sbp> the literal syntaxes are questionable, too

20:05:47 <AaronSw> forumla: syntax or semantics?

20:06:00 <sbp> otherwise, good from a cursory review

20:06:06 <sbp> just the name

20:06:13 <sbp> Gotta run

20:06:25 <AaronSw> are they named something else now?

20:06:31 <AaronSw> are you ever on IRC for more than 60 seconds?

20:06:43 <AaronSw> will another episode contain the thrilling conclusion of...

20:28:16 * dajobe wonders about adding more ':)'s to emails

20:29:48 * danbri wonders about adding more >:-/ to emails

20:30:08 <dajobe> had a haircut?

20:30:09 <AaronSw> I wonder if smileys would change my sour reading of dave's tone

20:30:30 * dajobe just amused at prescod etc

20:31:35 * danbri wonders what re prescod

20:31:54 <danbri> hmm, I was just writing to the REST list before I crashed my PC; wonder if the msg was sent or rescued

20:33:02 <danbri> hmm2: mozilla upgrade time: run 1.1 snapshot or a nightly? anyone tried a nightly in last day or few?

20:33:49 <dajobe> yeah

20:33:55 <dajobe> last nights ok

20:34:18 <dajobe> what crashed the pc?

20:34:39 * danbri hitting the reset button when moving the keyboard cable <blush/>

20:34:45 <danbri> moz: ta

20:35:26 <Talliesin> dajobe, thanks for that post to DC-ARCH, that situation had been irritating me for the last week, but my concerns seemed to be ignored.

20:36:26 <AaronSw> Talliesin, what's your RL name?

20:36:47 <AaronSw> (if you don't mind sharing)

20:37:04 <dajobe> Talliesin: np

20:37:46 <Talliesin> hehe Talliesin is one of my RL names, but I'm also known as Jon Hanna.

20:39:45 * danbri took a while to realise that!

20:39:57 <AaronSw> aha

20:40:36 * danbri reads/deletes mail *almost* as fast as it arrives...

20:52:28 * danbri replies to pps (realising I've rarely gotten into those debates, just watched from the side)

20:52:39 <danbri> well, still scribbling. will send shortly.

20:52:55 <dajobe> well me too normally, just felt in the mood

20:53:04 <dajobe> wry amusement that is

20:53:20 <dajobe> or is this the rdf fol debate?

20:56:09 <danbri> yep

20:57:27 <sbp> Aaron: the forumla thing was just me pointing out a typo

20:57:48 <sbp> also, I dunno where you got this BNF from, but '' literals have never been in the N3 canon, AFAIK

20:58:15 <sbp> [[[

20:58:16 <sbp> prefixname

20:58:16 <sbp> (that is, a prefix followed immediately by a name)

20:58:16 <sbp> ]]]

20:58:18 <AaronSw> they were in rdfn3.g, i think

20:58:35 <sbp> odd way to put it

20:58:44 <AaronSw> how would you put it?

20:58:51 <sbp> prefix name

20:59:00 <AaronSw> that would imply a space

20:59:14 <sbp> "prefix ' ' name" would imply a space

20:59:24 <AaronSw> not in the scanner syntax

20:59:49 <sbp> what scanner syntax is this, anyway? there are no references

21:00:00 <AaronSw> i felt it would be sufficiently obvious

21:00:22 <sbp> "obvious" doesn't cut it in specifications :-)

21:00:40 <sbp> and as for "sufficiently obvious"... [sucks in air]

21:01:33 <AaronSw> I think I'm going to take the policy that I will only add documentation like that if someone honestly wants to know and doesn't know what it means.

21:01:43 <AaronSw> You're free to write the documentation yourself, of course.

21:01:57 <sbp> also, you use parentheses for comments *and* as part of the syntax (to group)

21:01:58 <AaronSw> Also, I hope that the examples section will make many things more clear.

21:02:21 <Talliesin> Talliesin is now known as TallieSupper

21:02:42 <AaronSw> fixed

21:02:56 * sbp refreshes

21:04:15 <sbp> parens still used under "qname"

21:05:12 <sbp> I notice that it doesn't allow @prefix _: <...> . override

21:05:38 <sbp> CWM does, but I'm not sure whether that's intentional. personally, I think you're right to disallow it

21:05:41 <AaronSw> Neither does afon.

21:06:29 <sbp> right

21:06:59 <sbp> on "This resolves to the universally quantified variable with the document-scope name following the "?"." - I think that TimBL said to quantify ?vars over their parent formulae, i.e. one level up. I'm not at all sure on that, though

21:07:26 <sbp> typo: URI-referene

21:07:58 <sbp> "list1 term+ list0": should be "list1 term* list0", I think

21:08:13 <sbp> same goes for formula

21:08:35 <AaronSw> ignore. fixed. fixed. fixed

21:08:35 <dajobe> couldn't you somehow generate that stuff from an annotated grammar. I always worry about introducing problems with hand-written grammars

21:08:48 <dajobe> even simple N-Triples missed quotes etc. now and then

21:09:10 <AaronSw> The annotated grammar is handwritten too, isn't it?

21:09:30 <dajobe> but it could be an annotated machine processable grammar

21:09:36 <dajobe> i.e html in the comments or somesuch

21:09:46 <dajobe> as long as the rules come from checked code

21:09:50 <sbp> perhaps fiddle about with DanC's yapps-to-html code

21:09:53 <dajobe> yeah

21:10:00 * dajobe would like to have got that far

21:10:17 <sbp> since Yapps, unfortunately, appears to be the way for the future

21:10:27 <AaronSw> So what you're really saying, is that the machine should check my grammar.

21:10:36 <dajobe> would be nice

21:10:40 <AaronSw> Why unfortunately?

21:10:41 <dajobe> then people could take the grammar too

21:11:06 <sbp> I'm not a fan of Yapps. I do think that CWM should use a generated parser, but I don't think that that generation should be done by Yapps

21:11:16 <AaronSw> Why not?

21:11:37 <dajobe> do tell

21:12:05 <AaronSw> ooh: http://www.microsoft.com/PressPass/features/2002/aug02/0821PalladiumFAQ.asp

21:12:11 <sbp> I just didn't like yapps as I was using it. IIRC there were some bugs (which I didn't note down--argh!), and other weirdnesses. I form code prejudices pretty fast based upon intuitive experience

21:12:29 <sbp> and then there's the optional period bug

21:12:52 <sbp> I think that the consensus on that was that N3 isn't LL(1) because of it

21:13:04 <sbp> but the hack that they introduced to deal with it is awful

21:13:26 <sbp> the Yapps grammar was meant to make the grammar readable, and the optional period thing has thrown it right off. terrible

21:14:14 <sbp> summary: someone needs to write a better LL(1) metaparser for Python :-)

21:15:24 * sandro fights with finishing up the blindfold grammar for n3.....

21:15:41 * sandro gets -lx working for output in cwm.

21:17:22 * sandro does cwm test/rules12.n3 --lx | otter [ more or less] and proves rdf(granpa,ancestor,bill).

21:18:22 <AaronSw> what's --lx do?

21:19:06 <sbp> and tell us quickly so that I can document it

21:19:11 <JohnE_> dajobe - you following this prescod thread?

21:19:14 <sbp> unless it's going to start off another flame war...

21:19:27 <JohnE_> strange point hes raising

21:19:35 * dajobe shrugs

21:19:58 <sandro> outputs in LX (a language of mine, which happens to run in otter). I have python code to turn LX into (flat) ntriples. When I link that in, cwm should have a working --flatten.

21:21:08 <sandro> But I need to run off and clean the kitchen, so that'll have to wait for monday. And perhaps Tim to get back from vacation and take a look at it before I check it into the main branch.

21:23:31 <AaronSw> Argh, WebDAV locking is biting me.

21:23:39 <AaronSw> Could not MOVE /notabug/swhack/._2002-08-23.html due to a failed precondition on the source (e.g. locks). [423, #0]

21:23:42 <AaronSw> (25)Inappropriate ioctl for device: This resource is locked and an "If:" header was not supplied to allow access to the resource. [423, #0]

21:23:56 <dajobe> JohnE: (or one of you) I also like RNC which is why it's in the spec

21:24:32 <JohnE___> sorry - my connection just bounced and reconnected

21:40:15 <Code_Poet> Code_Poet is now known as Talliesin

21:49:58 <jordan|mva> jordan|mva is now known as jordan

21:58:13 * danbri wonders if http://lists.w3.org/Archives/Public/www-rdf-interest/2002Aug/0175.html works for anyone here (re shared-meaning etc.)

21:58:55 <AaronSw> you could have been a bit stronger

22:00:01 * danbri chose not too; it's not a fight.

22:00:06 <danbri> s/too/to/

22:00:38 <AaronSw> I didn't mean in the fight sense.

22:01:00 <danbri> you chose a fighty metaphor :)

22:01:16 * danbri remembers Rule 1

22:01:17 <AaronSw> Do you pick fights with the <strong> tag?

22:01:24 <AaronSw> ;-)

22:01:26 <danbri> all the time...

22:01:34 <danbri> grrrrraaarghzzzzz

22:01:39 * AaronSw makes a note to keep danbri away from his HTML

22:04:42 <Talliesin> At least he doesn't use <b>

22:07:21 <AaronSw> i pick fights with <b>

22:08:04 <AaronSw> although the "landgrab" subject did make it sound like a fight

22:08:29 <danbri> true

22:08:41 * danbri never claimed to be consistent; and if he did, he was lying

22:09:30 <AaronSw> he was telling the truth until you came along

22:09:39 <Seth> danbri, /www-rdf-interest/2002Aug/0175.html worked for me :))

22:19:50 <danbri> ta

22:47:27 <danbri>http://www.mozillazine.org/articles/article2409.html

22:47:27 <dc_rdfig> I: http://www.mozillazine.org/articles/article2409.html from danbri

22:47:46 <danbri> I:|Support for Remote Calendar Data in Mozilla Calendar (mozillazine)

22:47:47 <dc_rdfig> Titled item I.

22:48:25 <danbri> I:In iCalendar .ics format, via subscription to HTTP/FTP documents

22:48:26 <dc_rdfig> Added comment I1.

22:48:31 <AaronSw> I:This sounds just like [iCal|http://www.apple.com/ical/]. I think I will try to download the calendar to my [iPod|http://www.apple.com/ipod/].

22:48:31 <dc_rdfig> Added comment I2.

22:48:59 <danbri> I:An online RDFcal2Ical service would be handy for prototyping an RDF version of this... Any volunteers?

22:48:59 <dc_rdfig> Added comment I3.

22:51:30 <Talliesin> Talliesin is now known as tallifone

22:53:50 <AaronSw> I:Postscript: It sort of worked. The iPod loaded the calendar fine, but apparently decided to repeat the holidays every *day* rather than every *year*.

22:53:50 <dc_rdfig> Added comment I4.

22:55:25 <danbri> I:Aren't these day-long holidays? ouch...

22:55:25 <dc_rdfig> Added comment I5.

22:56:20 <Seth> any rdf files of that for my collection?

22:56:42 * danbri doesn't know of a convertor (Libby might...)

22:58:30 <deltab> I:"With an iPod it's like it's Christmas every day!"

22:58:31 <dc_rdfig> Added comment I6.

22:58:34 <tallifone> tallifone is now known as Talliesin

22:58:35 <AaronSw> Hm, RDF->iCal looks doable.

22:58:37 <AaronSw> lol!

22:59:12 <Seth> n3 files are just as good :)

22:59:33 <AaronSw> Is RDFcal specified anywhere, danbri?

22:59:51 <danbri> RDFCal shorthand for "the 4 or 5 experimental RDF calendar formats"

23:01:10 <AaronSw> Well gee, with that kind of standardization who needs iCal?

23:01:58 <danbri> one is a pretty straight rdfization of icalendar (libby and co on www-rdf-calendar list); another is DanC's RDFization of Palm's cal format; also RSS+events (nice modest little vocab); prolly some others too, variants on the 1st...

23:02:43 <AaronSw> Well, which one do *you* use?

23:03:17 <danbri> rss/events, for www docs. Playing with palm import via palm vocab. There's a simple mapping rule in the SWAP tests dir that worked

23:03:47 <danbri> 2-3 weeks ago, we got XHTML event listing into RSS+events into cwm and out as Palm RDF and into DanC's tools and into my Palm.

23:03:56 <danbri> that was fun... though could do w/ streamlining

23:04:15 <AaronSw> That's pretty cool. Did you announce it somewhere and I missed it?

23:04:29 <danbri> #rdfig log'n'blog, that's all

23:04:44 * danbri doesn't have url handy nor can think of easy way to find it :(

23:04:50 <AaronSw> http://www.w3.org/2001/sw/Europe/events/view/

23:05:01 <danbri> that's libby's java viewer...

23:05:21 <danbri> hmm, how to get RPMs onto a Debian box?

23:05:30 <AaronSw> alien

23:05:34 <danbri> I just apt-got 'rpm'... seems to need init'ing somehow

23:06:26 <AaronSw> I recall that not working. the packages page says to use http://packages.debian.org/stable/admin/alien.html instead

23:06:45 * danbri installed alien, exploring

23:06:46 <deltab> this? http://rdfig.xmlhack.com/2002/08/04/2002-08-04.html#1028498627.780524

23:07:36 <danbri> thanks; but no, later than that...

23:07:46 * danbri wishes for week/fortnight views of chump data

23:08:10 <danbri> :)

23:08:11 <danbri> alien libical-0.23.1.20020423-oe1.i386.rpm

23:08:11 <danbri> libical_0.23.1.20020423-1_i386.deb generated

23:08:11 <danbri> [root:pts/4] debian:~> dpkg -i libical_0.23.1.20020423-1_i386.deb

23:09:07 <Seth> .. danbri, only calendar schema i can find in my records is http://ilrt.org/discovery/2001/06/schemas/ical-full/hybrid.rdf ... is that the one your using?

23:09:40 <danbri> see above, using several. that's the latest re the icalendar family tree... there is also RSS events module and DanC's palm-based vocab.

23:12:55 <danbri> Bummer; installed everything, but calender not working w/ latest Moz nightly.

23:14:53 <AaronSw> I:I'm hacking on the converted

23:14:53 <dc_rdfig> Added comment I7.

23:18:59 <danbri> I:Go AaronSw!

23:18:59 <dc_rdfig> Added comment I8.

23:19:10 * danbri heads offline; cu

23:20:36 <danbri>http://www.daml.org/listarchive/joint-committee/1179.html

23:20:37 <dc_rdfig> J: http://www.daml.org/listarchive/joint-committee/1179.html from danbri

23:21:22 <danbri> J:|Updated DAML Query Language (DQL) draft (Fikes, Hayes, Horrocks, eds.), from DAML Joint Agent Committee mailing list

23:21:22 <dc_rdfig> Titled item J.

23:22:17 <danbri> J:A formal language _and_ protocol, coo.

23:22:17 <dc_rdfig> Added comment J1.

23:23:37 <danbri> J:Seems to be same basic idea as that advocated in [http://www.w3.org/TandS/QL/QL98/pp/enabling.html|Enabling Inference] QL'98 paper, but with more details (bNodes, protocol) worked out.

23:23:37 <dc_rdfig> Added comment J2.

23:24:29 <AaronSw> hm, can someone prefix the resources in http://sw1.ilrt.org/discovery/2002/05/rsscal/events.xml with rdf: ?

23:24:57 * danbri *could* but won't; bug Libby!

23:25:18 <danbri> I don't want to wade in and change her files unless urgent... doesn't your parser have a tolerant mode?

23:25:25 <AaronSw> I don't think so.

23:25:29 <AaronSw> Maybe http://swordfish.rdfweb.org/calendar/events/swevents.rss will work.

23:29:38 <Seth> redfoot does, reads fine in sailor :)

23:29:52 <AaronSw> I'm using redfoot.

23:30:19 <Seth> im using the redfoot lax parser

23:30:45 <Seth> otherwise 60% of the rdf on the web wont read

23:52:02 <AaronSw> anyone have an ical tool / validator?