Semantic Web Interest Group IRC Chat Logs for 2003-02-26

This is an automatically generated IRC chat log made by the logger bot from the Semantic Web Interest Group IRC chat at irc://irc.freenode.net/rdfig (also known as server irc.freenode.net channel #rdfig if that URI does not work for you).

NOTICE: #rdfig logs are being turned off 2004-12-03. Please switch to the new and shiny #swig channel for Semantic Web Interest Group chat. Change your client to #swig and enjoy the new experience. Or read the latest #swig logs to see what you've been missing :)

Semantic Web Interest Group Logs > 2003 > 2003-02 > 2003-02-26 (Latest) (Search)

00:25:40 <Jibbler> Jibbler is now known as Jibbler|sleep

01:04:06 * rajiv is away: gone

02:27:15 <jlinden> jlinden is now known as jlinden_sleep

03:30:44 * rajiv is back (gone 02:26:37)

03:55:14 * rajiv is away: gone

04:14:40 * burtonator is away: talking to roommate

04:28:45 * burtonator is back (gone 00:14:02)

06:26:49 <mdupont-zZZzz> mdupont-zZZzz is now known as mdupont

07:01:37 <lilo> [Global Notice] Hi all. Services restore may be necessary. Additional information on WALLOPS, /mode <yournick> +w .... thanks.

07:21:59 * rajiv is back (gone 03:26:44)

07:34:22 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

08:15:11 * rajiv is away: gone

09:06:19 <Jibbler|sleep> Jibbler|sleep is now known as Jibbler

09:17:57 <golbeck> golbeck is now known as golbeck_zzz

09:24:44 <mdupont> mdupont is now known as mdupont-work

09:24:54 <mdupont-work> mdupont-work is now known as mdupont-

09:28:28 <mdupont-> mdupont- is now known as mdupont-work

10:50:45 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

10:50:45 <arnarl> hi

10:52:02 <dajobe> hi

12:22:18 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

12:26:21 <dajobe> foo

12:57:36 <danb_offsick> hi hugo

12:59:22 * danb_offsick hmms at http://lists.w3.org/Archives/Public/xml-dist-app/2003Feb/0120.html

13:45:12 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

13:47:46 * logger_1 is already logging

14:05:03 <zool>http://opengis.net/gml/01-029/GML2.html

14:05:04 <dc_rdfig> A: http://opengis.net/gml/01-029/GML2.html from zool

14:05:22 <zool> A:|GML (Geography Markup Language) 2.0

14:05:22 <dc_rdfig> Titled item A.

14:08:05 <shellac>http://www.salon.com/tech/feature/2003/02/26/loebner_part_one/index.html

14:08:05 <dc_rdfig> B: http://www.salon.com/tech/feature/2003/02/26/loebner_part_one/index.html from shellac

14:08:21 <nmg> wow - gml uses xlinks

14:08:21 <shellac> B:|AI Stupidity

14:08:22 <dc_rdfig> Titled item B.

14:09:08 <shellac> B:A warning from history...

14:09:09 <dc_rdfig> Added comment B1.

14:09:17 <zool> nmg: i met a very interesting cartographer last night, he of webmapper.net, very interested in collaborative mapping / RDF metadata on GPS traces typr project

14:09:29 <nmg> interesting...

14:09:29 <zool> also i find out the cartography community is very very keen on SVG

14:09:51 * nmg must get round to putting his RDF'd unlocode data online

14:10:31 <zool> that would be good...

14:11:37 <zool> wow, it seems heavy the GML... i am looking for a plausible subset that could become an RSS like format that bloggers, in partic could offer autodiscovered from the web

14:12:39 * zool wonders how it connects if at all to that RDFMap

14:20:57 <nmg> not quite sure what the relationship is between GML and OpenGIS - GML seems to make a lot of assumptions that presumably OpenGIS solves

14:26:19 <zool> i thought GML was the opengis.org standard though?

14:26:38 * zool shakes head still so much to learn

14:26:43 * space waves at zool

14:26:45 * space too

14:27:20 <zool> hi space!

14:30:28 <space> zool: Ian Malpass was asking some RDF questions on (void) last night. Are you still subbed or did you give up?

14:33:14 * zool just attempted to answer them, babbling about RDFS and OWL

14:33:23 <space> okeydoke

14:33:39 <zool> two days resubscribed and already i get to be RDF cheerleading, rathergood ;)

14:34:11 <space> well, it's been quite on topic. Which is nice.

14:34:32 * zool tries to avoid getting into a topicmaps vs RDF conversation

14:36:36 <space> zool: good reply. that's helped me with some of the namespacey things I've been thinking about with SUDS.

14:37:17 <zool> (SUDS is Used to Define Soap Operas)++ :)

14:38:07 <space> we have richardc, blech, jenny and muttley to thank for that one :)

15:09:45 <zool> also curious as to any existing RDF vocab for representing GPS waypoints...

15:19:53 <danb_offsick> zool, not that i know of. that was one motivation for making the little geo: vocab, so it could be used by other things like waypoint description...

15:20:29 <danb_offsick> i've optimistically been taking lots of photos of local shops, in anticipation of the vocab coming together...

15:26:36 <zool> danb_offsick: getting the impression a lot of people waiting for the problem to be solved... movement towards declaring an interim schema at least... the time is now, &c

15:27:21 <zool> i have a waypoint-like representation of routes in my spacenamespace which i'd forgotten about writing, but i've never used it and its only tied into the uris, not to the geodata itself

15:30:53 <danb_offsick> danb_offsick is now known as danbri

15:31:14 * danbri still offsick, but no point in wallowing ;)

15:31:28 <danbri> yes, definte time is now feeling :)

15:46:24 <Morbus> hey bitsko. long time.

15:46:32 <bitsko> hey

15:47:23 <bitsko> been heads-down in a new job and remodelling a bathroom. my daughter's two year old is my taskmaster, "papa, go work on the bathroom"

15:47:35 <Morbus> heh!

15:47:44 <bitsko>http://bitsko.slc.ut.us/~ken/bathroom/

15:47:44 <dc_rdfig> C: http://bitsko.slc.ut.us/~ken/bathroom/ from bitsko

15:47:59 <bitsko> egads, /me totally forgot

15:48:40 <bitsko> C:|Ken's Taskmaster

15:48:40 <dc_rdfig> Titled item C.

15:48:45 <Morbus> cute pics.

15:49:00 <bitsko> C:oops, didn't really mean to log that, but, there ya go!

15:49:00 <dc_rdfig> Added comment C1.

15:50:58 <bijan> You can delete :)

15:51:58 <bitsko> really? how?

15:52:34 <dajobe> do C:=new URL

15:53:03 <bitsko> heh. that means I have to go research an "interesting" URL ;)

15:54:03 <dajobe> or just change it to your site, and chump what you've been doing apart from bathrooms

16:00:11 <bitsko> eh, no worry. what's a little eclecticism among friends :D

16:00:51 <bitsko> I had a question tho: anyone know what bot 'datum' is, or is it an AaronSw original?

16:01:31 <DanC>http://lists.w3.org/Archives/Public/www-rdf-calendar/2003Feb/0013.html

16:01:32 <dc_rdfig> D: http://lists.w3.org/Archives/Public/www-rdf-calendar/2003Feb/0013.html from DanC

16:01:42 <DanC> D:|Agenda for next IRC RDF calendar meeting: 26th Feb 2003, 17:00 GMT

16:01:42 <dc_rdfig> Titled item D.

16:01:57 <DanC> D:I think I'll make a BLURB for each item...

16:01:57 <dc_rdfig> Added comment D1.

16:02:23 <DanC> BLURB: datestamps (why aren't they considered as datetimes?)

16:02:23 <dc_rdfig> E: datestamps (why aren't they considered as datetimes?) from DanC

16:02:35 <DanC> E:from [agenda|http://lists.w3.org/Archives/Public/www-rdf-calendar/2003Feb/0013.html]

16:02:36 <dc_rdfig> Added comment E1.

16:02:58 <DanC> BLURB: would bugzilla help for rdf-calendar stuff?

16:02:58 <dc_rdfig> F: would bugzilla help for rdf-calendar stuff? from DanC

16:03:03 <DanC> F:from [agenda|http://lists.w3.org/Archives/Public/www-rdf-calendar/2003Feb/0013.html]

16:03:03 <dc_rdfig> Added comment F1.

16:03:30 <DanC> BLURB: is it OK to check GPL'd code in under http://www.w3.org/2002/12/cal/ ?

16:03:30 <dc_rdfig> G: is it OK to check GPL'd code in under http://www.w3.org/2002/12/cal/ ? from DanC

16:03:41 <DanC> G:from [agenda|http://lists.w3.org/Archives/Public/www-rdf-calendar/2003Feb/0013.html]

16:03:41 <dc_rdfig> Added comment G1.

16:04:34 <DanC> D:hmm... they show up in reverse order in [the weblog|http://rdfig.xmlhack.com/]. Oh well.

16:04:34 <dc_rdfig> Added comment D2.

16:05:21 <DanC> D:sure would be nice if the weblog showed the chump label (A/B/C).

16:05:22 <dc_rdfig> Added comment D3.

16:06:20 <bitsko> hmm, thought it did in the HTML source, but I don't see it now.

16:06:47 <bitsko> D:hmm, thought it did in the HTML source, but I don't see it now.

16:06:47 <dc_rdfig> Added comment D4.

16:08:54 <bitsko> F:Random kibitzing, but one person's opinion (mine) is that a issue tracking tool is an indicator of low quality in a process, not a solution. Fix the process upstream.

16:08:55 <dc_rdfig> Added comment F2.

16:12:13 <DanC> I'm not all that interested in using bugzilla for rdf-calendar, but surely there are good times/places for issue tracking tools, no? how else do you reasonably queue work?

16:13:52 <bitsko> yes, there are good times/places for issue tracking tools. operational support (work requests) and customer service are excellent examples

16:15:12 <bitsko> for queueing work, I'm particulary targeting to avoid where "work" == list of bugs/featurelets (one-liners) among the list of big chunks of work effort

16:16:31 <bitsko> for big chunks of work effort, it often depends on the size of the project and it's level of geographical distribution

16:17:46 <DanC> any projects you consider exemplary in regard to process, bitsko?

16:19:03 <bitsko> for small to midsize projects: if all team members are colocated, a spreadsheet or XP's index cards are good; if the team is distributed, a spreadsheet or wiki is good

16:20:49 <DanC> are spreadsheets/wikis issue tracking tools?

16:22:03 <bitsko> re spreadsheets, Painless Software Schedules: http://joelonsoftware.com/articles/fog0000000245.html

16:22:25 * bitsko always adds a caveat when linking to Joel's site: Joel is not a vague person, he talks in details. Sometimes he has the details wrong when the message is right.

16:22:35 <Morbus> ++ on that link. i've been using that system for a good long time, with OmniOutliner.

16:23:38 <bitsko> ah, yes, an outliner should work equally as well. notably, so does a ToDo list. there begins to show an overlap between process and tracking tool

16:24:16 <Morbus> yup. i use OmniOutliner for my TODO, my outliner, and my schedule tracker thing. it's really really handy.

16:25:48 <bitsko> re wiki, my current two projects are using a wiki to track big chunks of work. at their peaks, we have about 50-70 chunks of work

16:27:18 <Morbus> what i really want in a scheduler is for me to say "i've got 30 minutes left today. give me a list of things i can complete in that time, ranked by "things I haven't worked on in a long time""

16:28:12 <bitsko> one project, a more recent one, has integrated the Gnats issue tracking system to the wiki. I've found that to be more constraining, and it demonstrates a process breakage because one reason for adding the issue tracking system was to track "bugs". my statement that issue-tracker==indicator of process breakage fell on deaf ears.

16:29:26 <jlinden_sleep> jlinden_sleep is now known as jlinden

16:29:31 <Morbus> i'm not done much with wikis - i started to use them for a book, but then we moved over to MT.

16:29:43 <bitsko> the project using just the wiki, but with the better process, doesn't have bugs to track

16:31:21 <bitsko> F:speaking specifically to Bugzilla as a tracking tool in-and-of itself, it's very cumbersome to use and comes preconfigured with overbearing amounts of information to enter and track.

16:31:21 <dc_rdfig> Added comment F3.

16:32:51 <bitsko> F:We're using Gnats on a project, which comes out of the box a little less overbearing, and then stripped out every unnecessary step and field in the Gnatsweb web interface.

16:32:51 <dc_rdfig> Added comment F4.

16:34:17 <bitsko> F:our Gnatsweb has two main modes: list of open issues, and view/edit issue

16:34:17 <dc_rdfig> Added comment F5.

16:38:17 <bitsko> F:we consolidated all the header items into a two-column header of seven lines, followed by an "add comment" box, then the list of comments in most-recent-first order

16:38:18 <dc_rdfig> Added comment F6.

16:38:31 <bitsko> I'll see if I can dig up an example...

16:40:23 <bitsko> nope. we must have cleaned out our test project. only the live project available

16:46:17 <bitsko> DanC: an exemplary project: http://c2.com/cgi/wiki?MarketInfo

16:48:36 <bitsko> 10.5 months, six colocated developers, 14 defects reported in released code. tho it doesn't say, as an XP project they likely either used either story cards or, given the several references to excel, a spreadsheet.

16:48:38 <DanC> hmm... what is it?

16:48:59 <bitsko> software project

16:49:18 <DanC> hm... six colocated developers... that's a style very far from what I usually do.

16:49:51 <DanC> "software project". Gee thanks. I mean: what does the software do?

16:50:57 <bitsko> real-time market data to remote locations

16:51:52 <bitsko> it was 168 chunks of work

16:56:57 <bitsko> our two projects are both four remote developers. both projects are web applications, one a long-distance provider subscriber registration and customer care site, the other a corporate wireless costs management system

16:59:20 <DanC> "our"?

16:59:27 <bitsko> in one the queue of work is on one wiki page, in the other the queue is in gnats comingled with issues. in both, in our weekly meeting we create a new wiki page for the week that we copy-n-paste this weeks work items to.

16:59:43 <bitsko> the company I'm working with, as opposed to the one I just linked to

17:00:03 <bitsko> M@rketInfo isn't one of my projects

17:00:13 * DanC is pretty high on wiki these days.

17:00:17 <libby> hm, time for calendaring I reckon

17:00:25 <DanC> so it is.

17:00:33 <libby> afternnon all

17:00:40 <libby> ------- calendar meeting --------

17:00:49 <DanC> Let all those with rdf-calendar business draw near, the honorable Libby Miller, presiding.

17:00:49 <DanC> ;-)

17:00:55 <libby> cough

17:00:59 <libby> heheh

17:01:34 <DanC> DanC has changed the topic to: RDF Calendar 1700Z Wed. http://rdfig.xmlhack.com/

17:01:49 <libby> ok, all the agenda is currently my queestions at the moment. does anyone else have anything they would like to talk about?

17:02:11 <DanC> pathcross

17:02:25 <libby> pathcross? have you got a url?

17:02:55 <libby> hey hugo btw, glad you could make it

17:03:03 * hugo waves

17:03:09 <libby> ok, well let's start with 1 datestamps (why aren't they considered as datetimes?)

17:03:13 <libby> for now

17:04:55 * libby wondered why although datesstamps have the syntactic format of datetims, in ical2rdf.pl you dont get a datetime prooperty, e.g. http://www.w3.org/2002/12/cal/test/mtg.rdf

17:05:20 <libby> well, this is because dtstamp is a property and datetime is also a property so we'd nbeed somethign in between I suppose

17:05:32 * libby wonders if she's talking to herself ;)

17:05:42 <DanC> BLURB: anybody wanna play pathcross?

17:05:43 <dc_rdfig> H: anybody wanna play pathcross? from DanC

17:06:04 <DanC> H:i.e. connect your travel schedule to your foaf info, look for intersections with other travel schedules

17:06:05 <dc_rdfig> Added comment H1.

17:06:20 <libby> H: ooh excellent idea

17:06:20 <dc_rdfig> Added comment H2.

17:06:24 <DanC> ok... datestamps..

17:06:24 <bitsko> cool

17:06:36 <libby> yeah, less exciting

17:06:50 <DanC> vegetable first...

17:06:52 <DanC> vegetables first...

17:07:03 <DanC> <dtstamp>2002-12-19T20:53:57Z</dtstamp>

17:07:21 <DanC> hmm... I don't see the issue. what would you expect?

17:07:22 <libby> no, I always tholught vegetables first, but I think that's a big mistake. sometimes you never get to teh meat ;)

17:07:52 <libby> I guess I was thinking of datetimes as typed objects rather than properties

17:08:05 <libby> perhaps there isn;t an issue

17:08:56 <DanC> "typed objects"... not sure I understand.

17:09:33 <libby> well the thing at the end of a datestamp looks the same sort of thing at the end of a datetime. but perhaps we don;t need to capture that

17:09:33 <DanC> oh... perhaps I do...

17:09:44 * libby doesnt care that much about datestamps really

17:10:27 <DanC> E:we looked at [test case:http://www.w3.org/2002/12/cal/test/mtg.rdf]

17:10:27 <dc_rdfig> Added comment E2.

17:10:48 <DanC> PROPOSED(E1): all is well.

17:10:53 * DarkPenguin doesn't quite see the issue

17:11:41 <DanC> perhaps DarkPenguin would 2nd E1?

17:11:48 <DarkPenguin> sure

17:11:53 <libby> go ahead dark penguin

17:12:02 <DarkPenguin> I second E1

17:12:02 <DanC> objections?

17:12:21 <libby> nope

17:12:36 <DanC> E:RESOLVED: all is well. nothing needs changing. datestamp in this test case is ok.

17:12:36 <dc_rdfig> Added comment E3.

17:12:39 * libby vows to find more interesting agenda items for next time

17:12:50 * DanC likes the occasional easy one

17:13:13 <libby> ok

17:13:14 <DanC> on to F:...

17:13:18 <libby>

17:13:29 <libby> erk, what happpened then?

17:13:35 <DanC> ?

17:18:52 * DanC thx mattb

17:18:52 <DanC> my approach to x-props and namespaces is rather uninspired, but I haven't found inspiration for anything more elegant

17:18:57 * libby looks at some exaple prodids

17:19:09 <DanC> I noodled on making a namespace name out of the prodid, but...

17:19:37 <DanC> ... unless the guy who made up the prodid is gonna actually serve an RDF schema via http, I didn't see much use in i.

17:19:39 <DanC> in it.

17:19:52 <libby> seems unlikely on tha face of it

17:20:08 <libby> there is a standard way of contructing them maybe...?

17:20:16 <DanC> we could make a product registry under http://www.w3.org/2002/12/cal/ somewhere.

17:20:48 <libby> yeah

17:21:04 <DanC> i.e. PRODID:-//Apple Computer\, Inc//iCal 1.0//EN maps to http://www.w3.org/2002/12/cal/prod/Apple_Comp_4o984

17:21:42 <libby> I think all that was bugging me was that it sort of went outside the syntactic-only conversion because you had to pass it soemthing external.

17:21:49 <DanC> where 4o984 is a hash of "-//Apple Computer\, Inc//iCal 1.0//EN" and Apple_Comp is the first few uri-happy characters, for mnemonic purposes.

17:21:57 <libby> again, I don;t care passionately about it or anything

17:22:09 <DanC> the passing is a pain

17:22:14 <libby> i.e. a textfile or soemthing?

17:22:26 <DanC> textfile? what do you mean?

17:22:41 <DanC> I was thinking of putting an RDF schema at http://www.w3.org/2002/12/cal/prod/Apple_Comp_4o984

17:23:02 <libby> ah

17:23:05 <libby> great

17:23:28 <DanC> it's a bit of a pain for my perl code, cuz now I can't do the conversion in one pass.

17:23:34 <DanC> but not a big deal.

17:23:54 <DanC> it's better than what i'm doing now, because...

17:24:20 <DanC> xmlns:x='http://www.w3.org/2002/12/cal/prod_evo#'

17:24:24 <DanC> is 404.

17:24:42 <libby> ok, yep

17:25:31 <libby> PROPOSED (I1): that DanC write

17:25:34 <libby> damn

17:25:55 <libby> do we need to proposed this? it sounds like a good idea to me and only you and I have implementations as yet danc

17:26:33 <DanC> PROPOSED(I.reg): to institute an ical product registry ala 'http://www.w3.org/2002/12/cal/prod/' + firstFew(prodID) + hash(prodId) and to put RDF schemas there based on data we find from deployed products

17:26:45 <libby> thanks DanC

17:26:52 <lilo> [Global Notice] Hi all. Services has split and we have time to move it. Details on wallops (/mode yournick +w)...thanks.

17:27:04 <libby> seconded

17:27:20 <Jibbler> Jibbler is now known as Paul|trampsoc

17:27:22 <DanC> objections?

17:27:39 <DanC> volunteers to implement it?

17:27:52 <libby> i.e. write the schema?

17:28:15 <DanC> no, I mean add support to your java code and check in the code and update the tests

17:28:37 <libby> ok, yep, I can do that

17:28:45 <DanC> in particular, the 3 blessed tests: DATA_FOR_SCHEMA=$(CALTEST)/mtg.rdf $(CALTEST)/cal01.rdf \

17:28:45 <DanC> $(CALTEST)/gkexample.rdf

17:29:02 <libby> ok. blessed?

17:29:13 <DanC> blessed in that the ical schema is built from them.

17:29:21 <libby> ah, ok

17:29:26 <libby> shall we move on?

17:29:33 <DanC> just a sec...

17:29:57 <DanC> I: RESOLVED to institute an ical product registry ala 'http://www.w3.org/2002/12/cal/prod/' + firstFew(prodID) + hash(prodId) and to put RDF schemas there based on data we find from deployed products

17:29:58 <dc_rdfig> Added comment I2.

17:30:13 <DanC> I:action libby: implement it, add/update some tests

17:30:14 <dc_rdfig> Added comment I3.

17:30:33 <DanC> I:I hope to implement it too, in due course

17:30:33 <dc_rdfig> Added comment I4.

17:31:01 <libby> you are writing the registry danc?

17:31:10 <DanC> ?

17:31:36 <DanC> oh... you mean cvs adding the directory?

17:32:01 * libby thinks I'm not completely clear what you';re suggesting, sorry

17:32:30 <DanC> hmm... let's go over the current state...

17:32:47 <DanC> do you know how to build ical.rdf, the schema?

17:33:24 <libby> no, at least I've not tried it

17:33:34 <DanC> oops.

17:33:47 <DanC> type make in the 2002/12/cal directory

17:33:56 <libby> ok

17:34:38 <DanC> you might need to make clean first

17:35:02 <DanC> any volunteers to document building the ical.rdf schema?

17:36:08 * DanC notes these "test cases" are a little silly in that there's no mechanism to check the experimental results against the expected results

17:36:10 <libby> hm, I could have a go, depends if I can get python working

17:36:29 <libby> dont have cwm etc at the moment

17:36:39 <DanC> ah... yes... it depends on cwm and such.

17:36:47 <libby> danc, I do....kinda - I have your versions and then jena's graph match to check them

17:36:57 <DanC> ouch. I *really* don't want to be the only person who can do this.

17:37:02 <libby> or is taht not what you meant

17:37:23 <DanC> graph match: that's *exactly* what I meant!

17:37:39 <libby> ok, then action libby to document building ical.rdf schema

17:37:47 <DanC> do you have a way for folks to do regression tests with your code easily?

17:38:04 <DanC> BLURB: maintaining the ical.rdf schema

17:38:04 <dc_rdfig> J: maintaining the ical.rdf schema from DanC

17:38:13 <DanC> J:currently underdocumented. (sorry)

17:38:13 <dc_rdfig> Added comment J1.

17:38:30 <DanC> J:action libby: have a go at figuring it out, documenting it.

17:38:31 <dc_rdfig> Added comment J2.

17:38:57 <DanC> J:clues are in the [Makefile|http://www.w3.org/2002/12/cal/Makefile]

17:38:57 <dc_rdfig> Added comment J3.

17:39:00 <libby> you can type make test, though I probably dont have the conventions right

17:39:21 <DanC> BLURB: regression testing

17:39:22 <dc_rdfig> K: regression testing from DanC

17:39:36 <libby> make compare compares the graphs e.g. $(JAVA_HOME)/bin/java -cp $(CP) jena.rdfcompare \

17:39:36 <libby> test/generatedrdf/mtg.rdf test/perlrdf/mtg.rdf RDF/XML RDF/XML

17:39:44 <DanC> K:libby seems to be doing the right thing with graph matching and all that. DanC is really curious.

17:39:44 <dc_rdfig> Added comment K1.

17:40:24 <libby> K:see [makefile for java ical2rdf|http://ilrt.org/discovery/2003/02/cal/mimedir-parser/Makefile] and [readme|http://ilrt.org/discovery/2003/02/cal/mimedir-parser/readme.libby]

17:40:24 <dc_rdfig> Added comment K2.

17:40:50 <DanC> K:spiffy. you're one step ahead of me.

17:40:50 <dc_rdfig> Added comment K3.

17:40:52 <libby> it was no biggie - because jena does graph matches, I figured taht was better than a syntactic diff

17:41:03 <DanC> I think I read that readme before, but it leaked out.

17:41:14 <DanC> well, I've never installed jena.

17:41:40 <libby> you dont need all of it, just a jar I think in this case. or maybe 3

17:41:53 <DanC> well, you need a working jdk. or at least a JRE.

17:42:01 <libby> yeah, 3: http://ilrt.org/discovery/2003/02/cal/mimedir-parser/lib/

17:42:03 <libby> ah

17:42:13 <DanC> what jdk do you use?

17:42:52 <DanC> I try to rely on debian for tool support, but they don't seem to support java well enough for most of the java-based toys I try to play with.

17:42:58 <libby> heh not sure on my mac. on my debian box I think it's blackdown

17:43:09 <libby> right, yeah, it is a bit of a pain

17:43:11 <DanC> sigh.

17:43:32 * libby just having a look

17:44:16 <DanC> ok, let's pretend you can build ical.rdf ... lemme continue the x-properties story from there?

17:44:20 <libby> so, to get back to resolved for (I), doi I need to work out what the hashes are etc?

17:44:25 <libby> yep, good idea

17:44:52 <DanC> ok, consider:

17:44:52 <libby> java version "1.3.1"

17:44:52 <libby> Java(TM) 2 Runtime Environment, Standard Edition (build Blackdown-1.3.1-02b-FCS)

17:44:52 <libby> Java HotSpot(TM) Client VM (build Blackdown-1.3.1_02b-FCS, mixed mode)

17:44:54 <DanC> $(CALTEST)/cal01.rdf: $(CALTEST)/cal01.ics ical2rdf.pl

17:44:54 <DanC> $(PERL) ical2rdf.pl --xnames $(EVO_NS) $(CALTEST)/cal01.ics >$@

17:45:13 <DanC> yes, you have to work out the hash or whatever.

17:45:31 <DanC> now when you've done that, we won't need the --xnames $(EVO_NS) bit, right?

17:45:36 <libby> ok

17:45:37 <libby> yep

17:46:03 <DanC> now look in http://www.w3.org/2002/12/cal/ical.n3 (or .rdf if you prefer)

17:46:20 <DanC> do you see x_:evolutionAlarmUid a rdf:Property;

17:46:20 <DanC> ss:label "evolutionAlarmUid";

17:46:20 <DanC> ss:rangeIntersects s:Literal .

17:46:20 <DanC> ?

17:46:44 <libby> yep

17:47:12 <DanC> that belongs not in http://www.w3.org/2002/12/cal/ical but in http://www.w3.org/2002/12/cal/prod/ximian_evo_4o84 , right?

17:47:55 <libby> is the 4o84 bit made up?

17:47:58 <DanC> yes.

17:48:29 <libby> right

17:48:31 <libby> yes then

17:48:50 <DanC> so I'd like you to make all that happen. ;-)

17:48:56 <DanC> 1/2 ;-) actually.

17:49:12 <libby> ok, well I think I get it, still not sure where hashes are coming from. I'll ahve a go

17:49:24 <DanC> do as much as you're inspired to, and send mail to www-rdf-interest reporting your progress.

17:49:30 <libby> okey dokey

17:49:57 <DanC> I'm thinking base64(sha1(prodid))

17:50:06 <danbri> -interest or -calendar

17:50:14 <libby> ok

17:50:15 <DanC> -calendar

17:50:19 <libby> probably calendar

17:50:37 <libby> ok, next agenda item?

17:50:54 <libby> 3. will the mailing list do for this kind of thing, or would bugzilla

17:50:54 <libby> help?

17:51:16 <DanC> aka F

17:51:32 <libby> I thought bugzilla or something might be useful, but actually, this meeting seems to work well enough

17:51:39 <DanC> PROPOSED(F.no): no thanks, bugzilla is too much hassle

17:51:56 <DarkPenguin> I'll go with that

17:52:05 <DanC> bugzilla seems good when you have zillions of writers to the bug list and you need to constrain them

17:52:31 <danbri> I'm not finding bugzilla much of a hassle, for foaf i'm using not unlike a more structured wiki...

17:52:38 <DanC> PROPOSED(F.wiki): try the wiki instead for a while?

17:52:48 <libby> I was only thinking taht if people were wondering why we made such and such decision it might eb a bit complex to find out

17:53:10 * DanC pretty much hates anything he needs to manage Yet Another Password for

17:53:17 <libby> yeah

17:53:23 <libby> we already have the wiki right?

17:53:36 <DanC> yes. http://esw.w3.org/t/view/ESW/RdfCalendar

17:54:01 <libby> so what woudl we do with it? record decision made? changes to code?

17:54:07 <DanC> new node/page just for CategoryRDFCalendarBug?

17:54:16 * danbri is becoming dependent on mozilla remembering his passwords...

17:54:23 <libby> yeah, good idea danc

17:54:40 * DanC uses at least three different browsers that don't share password caches

17:55:01 * libby seconds F.wiki

17:55:13 <DanC> I still prefer F.no, btw. I'd like to hear more about where the bugzilla idea came from.

17:55:30 <danbri> yeah, wiki ok

17:55:47 <DanC> do we have one example bug to track at this point?

17:56:05 <libby> I've not really tried bugzilla properly. it wa sin my mind because danbri was playing with it, and we have something called resourcetracer at work that does somethign similar I think

17:56:34 * DanC is thinking YouAren'tGonnaNeedIt

17:56:49 <libby> I'm ahappy not to go with it

17:56:54 <libby> a wiki should be fine

17:58:10 <DanC> F: the [RDFCalendar wiki|http://esw.w3.org/t/view/ESW/RdfCalendar] is there

17:58:10 <dc_rdfig> Added comment F7.

17:58:13 <danbri> fwiw http://rdfweb.org/issues/ example issue: http://rdfweb.org/issues/show_bug.cgi?id=7 from the foaf site. Is being used in wiki-like fashion anyways...

17:58:41 <DanC> can i write to it without a new password/account?

17:58:42 <mdupont> has anyone tested the new redland with python and foaf?

17:58:49 <libby> let's move on. any objections to F.wiki? danc, would you rather stick with the mailing list alone?

17:58:54 <mdupont> sorry

17:58:59 <libby> I think anyone can write to the wiki

17:59:03 * mdupont did not notice the meeting

17:59:13 <DanC> no big deal, mdupont

17:59:33 <DanC> "I need a legitimate e-mail address and password to continue." -- http://rdfweb.org/issues/process_bug.cgi

17:59:38 <DanC> :_{

17:59:41 <DanC> :-{

18:00:04 <danbri> yeah, wonder if that can be swtiched off

18:00:09 <danbri> its pretty untweaked...

18:00:20 <DanC> now if it would somehow pick up my credentials by foaf-magic, I'd be impressed ;-)

18:00:54 <libby> yeah, that would be neat. it's online backing that bugs me the most. esp as you have to go and physically ask them out loud in a shop for a number furst

18:01:10 <libby> s/backing/banking

18:01:23 <libby> are we ok with F.wiki?

18:01:39 <DanC> yes

18:01:59 <DanC> F:RESOLVED: try the wiki instead for a while?

18:02:00 <dc_rdfig> Added comment F8.

18:02:06 <libby> great.

18:02:32 <libby> ok, final agenda item

18:02:34 <timbl> instead of?

18:02:34 <libby> is it OK to check GPL'd code in under

18:02:34 <libby> http://www.w3.org/2002/12/cal/ ?

18:02:34 <dc_rdfig> L: http://www.w3.org/2002/12/cal/ from libby

18:02:39 <libby> arse

18:02:47 * DanC feels very productive/efficient. how incurably geeky.

18:02:54 <libby> timbl, instaed of bugzilla

18:03:04 <libby> heheh

18:03:37 <DanC> aka G:

18:03:55 <libby> oh actually, it's not the last agenda item. we also have pathcross

18:03:56 <DanC> libby, who owns the code you write?

18:04:00 <DanC> MIT owns mine

18:04:11 <libby> universirty of bristol I think

18:04:24 <libby> pretty sure actually. that's the copyright I put on when I remember

18:04:55 <DanC> I know I can release code under the W3C software license. I don't know if I can release code that I write under GPL.

18:05:08 <libby> all my java calendar stuff depends on daniel resare's GPLed code

18:05:11 <mdupont>http://www.iata.org/sked/index

18:05:11 <dc_rdfig> M: http://www.iata.org/sked/index from mdupont

18:05:15 <danbri> (offtopic?) re credentials/magic, i scribbled an idea for this in http://rdfweb.org/mt/foaflog/archives/000005.html

18:05:28 <danbri> DanC, you or others could always relicense it under GPL...

18:05:32 <mdupont> M: The Scheduling Services department is responsible for the development and maintenance of standards and procedures for the exchange of schedule data between airlines and airports and airport Coordinators

18:05:32 <dc_rdfig> Added comment M1.

18:05:45 <DanC> BLURB: using foaf magic for authentication credentials

18:05:45 <dc_rdfig> N: using foaf magic for authentication credentials from DanC

18:05:52 <libby> I would tend to GPL it anyway even if it didnt....

18:05:58 <mdupont> libby: me to

18:06:03 <mdupont> get sued

18:06:09 <libby> get sued?

18:06:10 <mdupont> make slashdot, write a book

18:06:17 <libby> heh

18:06:37 * mdupont shuts up

18:06:51 <DanC> danbri, I don't own the code i write. I can't set licensing terms. Only MIT does/can.

18:07:08 <libby> I'm happy not to put it up on the w3c site, but someone suggested I did, and I dont want to put stuff up teher that shouldnt be

18:07:10 <DanC> does the SWAD-EU project have a license policy?

18:07:17 <libby> 'open source'

18:07:34 <mdupont> I would like to help map the RDF calendar to the IATA scheduling file format

18:07:35 <libby> we didnt specify a license - we've got a variety

18:07:37 <lilo> [Global Notice] Services have been restored from 16:10 backup. Server should be considerably more stable. Thanks for your patience.

18:07:38 <DanC> sigh... doesn't answer our questions, does it?

18:07:49 <libby> nope

18:08:16 <libby> well, you did say taht we coudl not make a decision here....I dont think a big problem.

18:08:28 <danbri> one thing w/ swad-e is that folks have pre-existing toolkits, with existing licenses, that they're doing additional work on via swad-e, eg. dave's redland, maybe jena, inkling... so it wasn't a fresh start on a swad-e specific toolkit.

18:08:30 <DanC> mdupont, I'll give you 300 points for exporting IATA schedules into HTTP space in RDF format.

18:08:41 <mdupont> I will have to deal with these IATA stuff alot more now, and I think it would have some synergies with RDF

18:08:51 <libby> right danbri

18:08:53 <mdupont> DanC: for my new job, i will have to learn this inside and out

18:09:01 <DanC> ok, well, who owns code written in swad-e?

18:09:17 <libby> um, HP, Uniiversity of bristol

18:09:21 <danbri> the organisations that employ the individuals...

18:09:25 <libby> stilo ltd?

18:09:30 <mdupont> the timezone information and airport codes from IATA is already used by unix

18:09:35 <libby> dunno what RAL comes under

18:09:37 <mdupont> bsd i think

18:10:18 <DanC> the unix tz database is public domain. I recently studied that carefully. http://www.w3.org/2002/12/cal/#tzd

18:10:38 <mdupont> ok, let me catch up, sorry to blurt out

18:10:51 <libby> s'ok mdupont

18:10:51 * mdupont looks

18:11:40 <libby> the tz stuff looks very useful danc

18:11:53 <DanC> PROPOSED(G.yes): sure. check in GPL code under http://www.w3.org/2002/12/cal/ . let's try it and see what breaks

18:12:15 <mdupont>http://rates4linux.sourceforge.net/r4l/index.php3

18:12:16 <dc_rdfig> O: http://rates4linux.sourceforge.net/r4l/index.php3 from mdupont

18:12:55 <mdupont> O:Rates4Linux the Linux rating application, a big user of the Time infomation

18:12:55 <dc_rdfig> Added comment O1.

18:12:56 * danbri 2nds PROPOSED(G.yes)

18:13:03 * timbl wonders whether sandro chumped his usgs data

18:13:07 <DanC> G:I'm note sure I can contribute to GPL code; MIT owns the code I write.

18:13:08 <dc_rdfig> Added comment G2.

18:13:24 <DanC> usgs data is in sql: space, not http: space

18:13:56 <DanC> and no, I'm not aware of any public announcement of it

18:14:00 <timbl> In fact in mysql: space, I think we decided?

18:14:14 <libby> objections to G.yes?

18:14:19 <DanC> hang on...

18:14:21 <libby> what's usgs?

18:14:34 <DanC> timbl, any thoughts on G, i.e. is it OK to check GPL'd code in under http://www.w3.org/2002/12/cal/ ?

18:14:48 <timbl> United States Geographic Survey

18:14:52 <libby> ah

18:15:18 <DanC> yes, should be mysql: space

18:15:20 <timbl> Yes, it is certianly OK for GPL coe to be hosted on 12/cal

18:15:49 <timbl> IIRC any W3C licence code eg as writteen by Dan can be taken and includeed in a GPL distribution.

18:15:57 <DanC> timbl, can I release code that I write (and hence MIT owns) under GPL?

18:16:29 <timbl> Lots of people at MIT write code under the GPL.

18:16:31 <DanC> but what about, e.g., contributing to libby's GPL'd code?

18:16:55 <DanC> ok, 2nded, G.yes.

18:16:59 <mdupont>http://www.iata.org/NR/ContentConnector/CS2000/Siteinterface/pdf/sked/sked_ssim_ig.pdf

18:16:59 <dc_rdfig> P: http://www.iata.org/NR/ContentConnector/CS2000/Siteinterface/pdf/sked/sked_ssim_ig.pdf from mdupont

18:17:00 <timbl> The problem is only if you "pollute" say a SWAD project with GPL code, which would require the downgrading of its W3C licence to GPL.

18:17:13 <timbl> This is my non-expert understanding.

18:17:43 <DanC> well, that's what we're talking about, timbl. we're talking about the license of the RDF calendar codebase (assuming it's homogeneous)

18:18:01 <mdupont> P:|IATA SSIM STANDARD SCHEDULES INFORMATIONS MANUAL - EDIFACT implementation guide for Scheduling Messages

18:18:01 <dc_rdfig> Titled item P.

18:18:05 <libby> can 'pollution' like that really happen?

18:18:05 <timbl> { ?x a :Discalimer } => { <> daml:imports ?x }. ;-)

18:18:11 <mdupont> P:PDF file

18:18:12 <dc_rdfig> Added comment P1.

18:18:36 <mdupont>http://www.iata.org/sked/publications/index

18:18:37 <dc_rdfig> Q: http://www.iata.org/sked/publications/index from mdupont

18:18:42 * DanC recalls discussion with Aaron on N3 rules for license terms... much fun...

18:18:59 <mdupont> Q:|IATA SSIM STANDARD SCHEDULES INFORMATIONS MANUAL - they dont have a copy online. Expensive

18:18:59 <dc_rdfig> Titled item Q.

18:19:09 <timbl> Libby: poluution: yes, if a company were to plan to use cwm in a product (no, seriously, just as a n example, stop laughing) and they wanted not torelease their changes to it, then inclusion of GPL code in it would mess it up.

18:19:26 <libby> right

18:19:42 * timbl has a URI for "lessPowerFullicenceThan" somewhere.

18:20:07 <mdupont> if someone wants to contribute to a W3C project, they cannot have a problem with a W3C member using it in non-free software

18:20:16 <libby> but that's different frm saying using cwm with some GPLed code used a s a library

18:20:17 <timbl> We don't make the copyleft requirement on SWAD code. But we do contribute to GPL projects.

18:20:26 * sandro did not chump his USGS thing because from his side of things it's only accessible to people on swada.w3.org.

18:20:29 <timbl> different: yes.

18:20:34 * libby gets in a mess with this sort of thing

18:20:49 <libby> people seem to get very paranoid about it

18:20:51 <DanC> no, it's not, libby. GPL spreads from libraries to main programs. LGPL does not; that's why it exists.

18:20:58 * timbl wondres whether Sandro had a mamoth file of the data in RDF format somewhere along the line

18:21:07 * timbl at least the schema.

18:21:27 <libby> danc, how about from code to the libraries used with it - taht's what I meant, sorry

18:21:38 <mdupont> www.iata.org/NR/ContentConnector/CS2000/SiteInterface/ pdf/sked/cdma/fds/sc108min.pdf

18:21:42 <mdupont> www.iata.org/NR/ContentConnector/CS2000/SiteInterface/pdf/sked/cdma/fds/sc108min.pdf

18:21:47 <mdupont>http://www.iata.org/NR/ContentConnector/CS2000/SiteInterface/pdf/sked/cdma/fds/sc108min.pdf

18:21:47 <dc_rdfig> R: http://www.iata.org/NR/ContentConnector/CS2000/SiteInterface/pdf/sked/cdma/fds/sc108min.pdf from mdupont

18:21:55 <libby> we might need to be very clear abotu separating out the graph matching part and the rest

18:22:08 <libby> (of the java code I've been pulling together)

18:22:14 <libby> hey lotr

18:22:16 <mdupont> R: The killer meeting that dashed hopes of making an XML fileformat for SSIM

18:22:16 <dc_rdfig> Added comment R1.

18:22:20 <LotR> hi

18:22:24 <mdupont> R1: The killer meeting that dashed hopes of making an XML file format for SSIM

18:22:24 <dc_rdfig> Replaced comment R1.

18:22:27 <DanC> { ?WHOLE :part ?PART. ?PART :license :GPL } => { ?WHOLE :license ?GPL}. # that's how I understand it.

18:22:30 <danbri> danbri is now known as danb_stilloffsic

18:22:43 <sandro> Nope. It was a wonderfully RDF-free experience. :-) In theory it was just re-running a script I wrote in 1998, but bugs in MySQL made it much more tedious.

18:23:00 <mdupont> R: [[IATA and SISC have received inquiries as to whether XML standards will be produced for SSIM. Some exploratory work will be undertaken, but no major work is anticipated at this time.]]

18:23:00 <dc_rdfig> Added comment R2.

18:23:21 <libby> ok, perhaps we can move on?

18:23:23 * DanC doesn't mind a little side discussion, but is now having trouble following the meeting

18:23:30 * libby too

18:23:32 <LotR> how was the calendar meeting?

18:23:40 <DanC> *is*, not was.

18:23:43 <libby> there are less tahn 10 minutes scheduled left anyway

18:23:44 <LotR> oh

18:24:06 <DanC> move on: pls, no, we just swapped in all the context. let's decide something.

18:24:09 <LotR> long one

18:24:17 * DanC thinks 90min is normal

18:24:21 <libby> sorry, thought we had

18:24:34 <mdupont> So IATAhad a working group to make an xml file format standard for schedule information, the first agenda item was to can the idea.

18:24:45 <LotR> DanC: it's been 60 every time I attended

18:24:47 * DanC checks the record, finds no decision

18:24:51 <libby> ok

18:25:37 <DanC> I'd like RDF calendar (i.e. http://www.w3.org/2002/12/cal/) to have a homogeneous license.

18:25:48 * timbl recalibrates his clock

18:25:49 <libby> so we had a proposal, G.yes and it was seconded

18:26:07 <DanC> well, I'm reconsidering.

18:26:09 <libby> I guess we're still workign through potential objections

18:26:23 <mdupont>http://www.iata.org/srs/phase1

18:26:23 <dc_rdfig> S: http://www.iata.org/srs/phase1 from mdupont

18:26:37 <DanC> homogeneious... so that we keep the option of packaging it up for release

18:26:48 <libby> I gues I could alk daniel if he feels like releaseing his code as non GPL, no idea if he would say yes

18:27:04 <mdupont> S: [[The data will be made available via real-time pull technology, such as Web Services based on XML, alternatively the data can be delivered as a SSIM file to the user of the data to compile as required by their e-commerce channel]]

18:27:04 <dc_rdfig> Added comment S1.

18:27:14 <libby> (daniel resare, who wrote the java mime-dir parser my java ical stuff depends on)

18:27:23 <libby> or I could look for an alternative parser

18:27:41 * DanC supposes our reservation of the channel has run out; can't follow the meeting with all these other things being chumped

18:27:43 <mdupont> S:|The Schedule Reference Service (SRS) is a neutral source of schedule data that provides the industry with quality service for the collection, validation, consolidation and distribution of airline schedules and related data.

18:27:43 <dc_rdfig> Titled item S.

18:27:50 <danb_stilloffsic> libby, http://www.w3.org/Consortium/Legal/2002/copyright-software-20021231 is now better suited to non-w3c-owned stuff than it used to be

18:28:14 <libby> mdupont, could you leave it for 5 mins or so while we wind up? sorry to be a pain

18:28:22 <mdupont> sure

18:28:25 <DanC> thx

18:28:30 <libby> right, but my stuff depend heavily on gpled stuff at the moment

18:28:40 <libby> cheers

18:28:53 <libby> how about we postpone a decision, see if there's a way around

18:28:55 <danb_stilloffsic> ...i meant, in case the author wanted to relicense, that license might suit them now

18:29:02 <libby> ok, right, yeah

18:29:47 <libby> in any case though, probably my code woud be owned by uinversity of bristol still while danC's code would be MIT....

18:29:59 <libby> so the copyright different even if the license the same

18:30:02 <DanC> postponing appeals to me, but I'd like to have some actions that ensure progress

18:30:08 <danb_stilloffsic> thats ok

18:30:18 <libby> well I'll action myself to talk to daniel resare

18:30:30 <libby> - ask if w3c license would suit his purposes

18:30:48 <DanC> danb_stilloffsic, any opinion on (a) whether it's worthwhile to have a homogeneious rdf-calendar license, (b) what it should be?

18:31:22 <DanC> G: Action libby: talk to daniel resare- ask if w3c license would suit his purpose

18:31:22 <dc_rdfig> Added comment G3.

18:31:54 <DanC> multiple owners doesn't conflict with a uniform license.

18:31:57 * timbl wanders off looking for something to eat

18:31:59 <libby> I personally dont think they need to be the same license

18:32:19 <DanC> you understand the cost of mixed licenses, right libby?

18:32:27 <libby> ...but perhaps that means we shoudl not mix them up on the same site

18:32:32 <libby> probbaly not

18:32:33 <danb_stilloffsic> i'd prefer to have everything available under W3C license, even though i'm a GPL sympathiser. But I don't think we _need_ a homogenously licensed codebase

18:32:42 <DanC> somebody wants to use rdf calendar stuff, and they have to look at each file to see what its license is

18:33:33 <libby> well I guess I was suggesting we dont bundle it together. after all, different languages and all

18:33:42 <DanC> another cost: if we package it up, we can't make a simple { :PKG :license ?X } claim for any ?X.

18:34:06 <libby> unfortunatelty we're out of time...

18:34:10 * danb_stilloffsic expects there to be lots of rdf calendar 'stuff', just as there is lots of RSS, Dublin Core etc stuff. The vocab and supporting test case data should be homogenous. Supporting software... also would be nice but that's an extra step.

18:34:15 * DanC has time for this

18:34:44 * libby wonders if theere are people who needs to go, need to discuss next meet?

18:34:50 <DanC> I agree that the license on the tests and schema is the critical bit.

18:35:17 <DanC> G:hmm... do we want/need a homogeneous license?

18:35:17 <dc_rdfig> Added comment G4.

18:35:33 <DanC> G:seems valuable for the data + vocab

18:35:33 <dc_rdfig> Added comment G5.

18:35:38 <danb_stilloffsic> On that front, ie the schema, can anyone comment on the situation re producing a derivative work from an IETF spec?

18:35:54 <DanC> ew. good question.

18:35:59 <DanC> pls G: it

18:36:24 * DanC is ready to postpone G: for now

18:36:38 <libby> yeah, LOTR asked a question abotu the copyright of posts on ieff calendar list recently

18:37:06 <libby> G:POSTPONED

18:37:06 <dc_rdfig> Added comment G6.

18:37:09 <DanC> I believe folks who do IETF work give their stuff (e.g. posts) to ISOC.

18:37:26 <libby> I dont remember what the answer was now

18:37:31 <libby> G:danb_stilloffsic: On that front, ie the schema, can anyone comment on the situation re producing a derivative work from an IETF spec?

18:37:32 <dc_rdfig> Added comment G7.

18:38:04 <bitsko> I read into the IETF license deeply at one point. only ISOC may make derivites. Internet Drafts are considered unpublished.

18:38:35 <libby> ok, well we still have one agenda item left, pathcross. also we could do with organising another meeting. also I'll eb at the tech plenary next week, and wondering if worth having a chat there

18:39:14 * DanC notes the irony... the tech plenary being a pathcross example

18:39:22 <libby> yeah ....

18:39:39 <libby> well maybe we could have a look at that now, desipte time being out...

18:39:48 <DanC> emerging convention is every other Weds at 1700Z?

18:39:53 * bitsko aside, RFCs in general do not require IETF licenses, only standards-track licenses do

18:39:57 <libby> sounds good to me

18:40:07 * DanC checks availability...

18:40:49 <libby> 12th march?

18:41:19 <DanC> I have Wednesday, March 12, 2003 11:00 am America/Chicago reserved in my calendar. http://calendar.sidekick.dngr.com/event?id=399

18:41:31 <DanC> 90min.

18:42:15 <DanC> so RESOLVED.

18:42:18 <DanC> move to adjourn.

18:42:20 <libby> D: next meeting [march 12th 2003, 17:00 UTC|http://www.timeanddate.com/worldclock/fixedtime.html?day=12&month=3&year=2003&hour=17&min=0&sec=0&p1=0]

18:42:20 <dc_rdfig> Added comment D5.

18:42:30 <libby> seconded

18:42:35 <DanC> ADJOURN.

18:42:52 * DanC is likely to hang out and play pathcross anyway...

18:43:06 * libby was just goign to ask you that

18:43:25 <libby> do you have a file already that describes your trip danc?

18:43:40 <DanC> does anybody keep foafy stuff connected to calendary stuff?

18:44:00 <libby> I had a link to a calendar at one point

18:44:04 <bitsko> and if not, why not? ;)

18:44:06 <libby> from my foaf file

18:44:11 <DanC> my itinerary for next week: http://www.w3.org/2003/03dc-bos/bos-me.rdf

18:44:28 <DanC> not connected to my foafy stuff:

18:44:50 <DanC> ... http://www.w3.org/People/Connolly/smart-home.rdf ... oops.

18:45:07 <libby> do you keep that uptodate?

18:45:16 <DanC> DanC has changed the topic to: RDF & SemWeb hack-n-chat http://rdfig.xmlhack.com/

18:45:23 <libby> I had foaf;subscribesto a calendar which isnt quite right

18:45:35 <libby> http://swordfish.rdfweb.org/people/libby/rdfweb/webwho.xrdf

18:46:18 <DanC> up to date... I make an .rdf file out of itinerary proposals from W3C travel staff before I accept them, as a rule...

18:46:35 <DanC> ... and I often link them from my homepage.

18:46:45 <timbl> How robust is it, now -- should I be cloning it?

18:47:08 <DanC> I have done some work on generating foafy stuff from my homepage, but I haven't closed the loop.

18:47:29 <DanC> how robust: umm... fairly. I've been using it every month for over a year.

18:47:52 <DanC> you have to teach grokItin.pl about each new airport

18:48:07 <timbl> Whats the rpocedure? (documented?) copy the directory, keep the Makefile?

18:48:45 <DanC> documented: I hope http://www.w3.org/2000/10/swap/pim/travel.html will serve as docs.

18:48:56 <DanC> don't believe it's enough for you now.

18:49:09 <DanC> yes, basically, make a new directory, copy a makefile and tweak

18:49:24 <DanC> wanna try it?

18:49:45 <DanC> i have rules to make .ics from the .rdf.

18:49:52 <timbl> But in fact, the travel message is teh only input file which is trip-specific?

18:50:00 <DanC> in fact, current version has the .ics properties in the published .rdf

18:50:23 <DanC> yes, ... only ... .

18:50:57 <DanC> umm... actually, the makefile has the lat/long of the focus of the .png view

18:51:13 <DanC> lat/long and view angle

18:52:23 <libby> ooh, hello areggiori!

18:52:30 <libby> how you doing?

18:52:47 <areggiori> hi libby

18:52:55 <danb_stilloffsic> re finding one's calendar, yeah i'd like to add something to foaf that says of a person, or organisation, 'here is its calendar'...

18:53:12 <timbl> The latlong of the veiw angle should be calculable from teh lat/long of the viewable things - but my spherical trig is nonexistent.

18:53:15 <DanC> ooh... are you the query use cases guy, areggiori? excellent work, that.

18:53:20 <danb_stilloffsic> ...just not sure about (a) granularity (personalCalendar, workCalendar etc), (b) format, whether to have different properties for ical vs ical-in-rdf

18:53:23 <libby> right, cos ical etc let you subscribe to them, but that's not enough

18:53:30 * danb_stilloffsic waves to alberto

18:53:31 <libby> danc, yes he is :)

18:53:55 <DanC> should be, yes, timbl. But hacking the makefile manually hasn't been sufficiently onerous to cause me to automate it.

18:53:59 <areggiori> just passing by - checking connection for tomorrow - talk to you tomorrow 15:00 GMT cheers

18:54:12 <libby> great, cheers alberto

18:54:29 <DanC> clue me in? what's up at 1500 tomorrow?

18:54:55 <libby> oh, sorry, we're going to chat about testcases for rdf query

18:55:19 <timbl> Oh ... I tried to add one an Mozilla ened up sending an RDF message to rdf-comments.

18:55:31 <libby> it's because I've got to write a report for swad-europe

18:55:55 <timbl> Jos and Patrick S then thought i was makin a statement on the list in RDF

18:56:04 <libby> heh

18:56:06 <timbl> ... and started criticising the form of the RDF.

18:56:12 * DanC is chairing a meeting next Thu/Fri where rdf query test cases are rather relevant... http://www.w3.org/2001/sw/meetings/tech-200303/

18:56:39 <libby> what about an rdf-only mailing list. that'd cut down that waffle.

18:56:44 * timbl ws adding them to swap/test/query/t01.n3 and so on and only got to 2, not checked in yet

18:56:55 <libby> oh yes

18:56:57 <libby>http://lists.w3.org/Archives/Public/www-rdf-rules/2003Feb/0011.html

18:56:58 <dc_rdfig> T: http://lists.w3.org/Archives/Public/www-rdf-rules/2003Feb/0011.html from libby

18:57:19 <timbl> I just was filling inthe form ASAP and didn't check the to: field.

18:57:21 <libby> T|:query testcase manifest chat agenda

18:57:30 <libby> T:|query testcase manifest chat agenda

18:57:30 <dc_rdfig> Titled item T.

18:57:37 * sandro sees a lot of rdfs:Comment triples in Libby's imagined mailing list. :-)

18:57:45 <libby> T:everyone welcome - 3pm gmt for an hour

18:57:45 <dc_rdfig> Added comment T1.

18:57:52 <libby> heh sandro

18:58:28 <libby> ok, danc, didnt see that

18:58:31 <timbl> T:"I still think that it would be very useful to come up

18:58:31 <dc_rdfig> Added comment T2.

18:58:31 <timbl> with a common way of describing a very simple conjunctive RDF query

18:58:35 <timbl> rats

18:59:12 * DanC finds the use of am/pm with GMT ... unexpected/odd for some reason. If you're gonna be stilted and international and use GMT or Z, I wouldn't expect to use the anglo/western am/pm

18:59:24 <libby> i

18:59:38 <timbl> T2:"I still think that it would be very useful to come up with a common way of describing a very simple conjunctive RDF query"

18:59:38 <dc_rdfig> Replaced comment T2.

18:59:44 <libby> I thought 17:00 etc wasn;t really in commion use in the US

19:00:02 <DanC> it's not. but nor is GMT

19:00:11 <timbl> Just do it ... they will learn ;-)

19:00:25 <libby> I guess. I would use UTC but my feeling is that hardly anyone would understand that

19:00:27 * eikeon converted clock to 24 hour UTC and feels much better now ;)

19:00:41 <DanC> so 5pm Boston time makes sense to me, or 1700Z. but not 5pmZ

19:01:00 * timbl sets his camera to Z but not his watch

19:01:32 * DanC got a digital video camera and was delighted to see it groks timezones.

19:01:50 * DanC sets his stupid digital camera to Ztime too.

19:02:01 <timbl> But anyway its 15:00Z which is 10:00ET Olson://America/NewYork etc

19:02:24 * timbl impressed very much by videocamera!

19:02:51 <deltab> Olson:// ?

19:03:06 * timbl wonders if he could get some N3 test cases done by 10:00ET tomorrow

19:03:09 <DanC> speaking of Olson, i'd like to collect some points for http://www.w3.org/2002/12/cal/#tzd , tim. I think you asked for that, yes?

19:03:32 <timbl> points?

19:03:45 <DanC> I believe you call them kudos.

19:04:19 <timbl> Oh, bounty. Yes. lots. I don't know what the current exchange rate its but you hereby get 150 rallods

19:04:35 * DanC pockets them, smiling.

19:05:24 <DanC> do you recall exactly why you wanted them? i.e. what you can do now that you couldn't before?

19:05:43 <DanC> and would you mind taking a look at http://www.w3.org/2002/12/cal/tzrules.n3 ?

19:06:05 <timbl> 100 more rallods bounty for an RDF conversion of the timezone boundary data , and 500 for an algorithm to generate zulu time from lat, long, localtime, using DanC's tz data and the tz boundary data.

19:06:36 * DanC hunts for "the timezone boundary data"

19:07:14 <timbl>http://www.nationalatlas.gov/timeznm.html

19:07:14 <dc_rdfig> U: http://www.nationalatlas.gov/timeznm.html from timbl

19:07:28 * timbl notes you hgave to look for time%20zone

19:07:55 <DanC> U:|Time Zones of the United States: Map Layer Description File

19:07:56 <dc_rdfig> Titled item U.

19:07:56 <timbl> U:|USGS Time Zones of the United States: Map Layer Description File

19:07:56 <dc_rdfig> Titled item U.

19:08:03 <timbl> snap

19:08:22 * DanC notes the pseudo-RDF in the HTML file

19:08:24 <timbl> U: 100 rallods bounty for conversion to RDF.

19:08:24 <dc_rdfig> Added comment U1.

19:08:40 <timbl> U: DanC notes the pseudo-RDF in the HTML file

19:08:40 <dc_rdfig> Added comment U2.

19:08:43 <timbl> :)

19:09:00 <timbl> There's got to be data behind that metadata

19:09:12 <DanC> er... where's the data?

19:09:22 <DanC> ah... "Download this map layer in Shapefile format."

19:10:18 <timbl> U: It is in two different forms of binary

19:10:18 <dc_rdfig> Added comment U3.

19:10:42 * rajiv is back (gone 10:55:31)

19:10:49 * timbl did that yesterday and got a pointer to an ISO spec from README.txt in the second format.

19:11:04 * DanC notes "tar zxvf" is much harder to type without the use of 4 of the fingers on my left hand

19:12:21 <justme> bounty

19:13:15 <danb_stilloffsic> re bounty etc http://www.favors.org/FF/

19:13:21 <danb_stilloffsic> re bounty etc see http://www.favors.org/FF/

19:13:25 <danb_stilloffsic> oops

19:13:26 <timbl> You need to reprogram a mouse click to mean "unwrap that, unzip it, untar it and convert it to RDF" ;-)

19:14:07 <DanC> ctrl-click for "and get the original publisher to keep the RDF up to date with changes". yes, of course.

19:15:27 * timbl deltab, Olsen? timezone identifiers in Evolution are of the form /softwarestudio.org/Olson_20011030_5/America/Chicago, which is presumably the punchline of a story about the work of a Mr Olsen....

19:15:44 <DanC> yes, fascinating work...

19:16:08 * timbl notes that if this were a video game DanC would be able to trade his rallods for fingers.

19:16:17 <DanC> LOL!

19:16:31 <deltab> I was surprised by the use of URL syntax

19:16:49 <timbl> URLish

19:17:41 <timbl> But now thanks to danc we can say http://www.w3.org/2002/12/cal/tzd/America/Chicago#tz

19:18:41 <DanC>http://www.rosinstrument.com/cgi-bin/showtext.pl/Technical/LinuxDocs/zoneinfo-96i-5/Theory-ps50-pn1

19:18:41 <dc_rdfig> V: http://www.rosinstrument.com/cgi-bin/showtext.pl/Technical/LinuxDocs/zoneinfo-96i-5/Theory-ps50-pn1 from DanC

19:19:01 <DanC> V:Theory, from the Olsen timezone stuff

19:19:02 <dc_rdfig> Added comment V1.

19:19:40 <timbl> Would it be possible to (a) make a big file which ahs all the tzdata in it and (b) make an annotation in th schema that that resource is a log:definitiveDocument for stuff like :tzid?

19:19:57 <DanC> V:unpacked by somebody from [its home at nih|ftp://elsie.nci.nih.gov/pub/]

19:19:58 <dc_rdfig> Added comment V2.

19:19:58 <mdupont> hey guys

19:20:18 <mdupont> did anyone find that Rating2LInux and IATA stuff interesting for scheduling?

19:20:29 <DanC> no, tim, cuz tzids are local.

19:20:55 * DanC hasn't looked yet, mdupont

19:21:41 <timbl> Ok, the ones given in the rdf fiels then are jsut arbitrary intersting strings.

19:22:40 <DanC> I'm thinking more about (a) a definitiveDocument for the toplevel tz stuff, e.g. America which points to all the divisions, e.g. Chicago, (b) one for each thing like Chicago that (b1) gives its convex hull/shape, (b2) relates it to a timezone.

19:22:45 <DanC> or something like that.

19:22:51 <timbl> But "standard" maybe is a property for which all possible statemements are in fact given in your files?

19:23:40 <DanC> no, actually, you and I can make up a timezone right here and now. that switches from standard to daylight every third tuesdsay.

19:23:52 <timbl> Ok.

19:24:08 <timbl> In fact, the timezone for me is a really useful one when sorting photos.

19:24:37 <DanC> you .jpg files have timezone info in them?

19:24:40 <DanC> your

19:24:54 <timbl> So one could only really use auto-lookup on the timezone URIs by looking up the subject and objects in a query on the web (as ooposed to just the predictaes).

19:25:23 * DanC isn't following.

19:25:39 <timbl> My jpeg files I set to Z. Other people's (eg jp phtopgraphers) gave theirs set to local time. jhead allows you to specify an offset when you convert to rdf.

19:26:22 <timbl> But when I actually make a photo album, it has to divide the photos by when I was asleep, not when Grenwich was asleep.

19:27:16 * timbl following: I am interesting in leaving breadcrumbs for cwm-s --mode= modes to automatically get needed stuff by dereferencing terms it comes across.

19:28:21 <DanC> well, I don't yet see enough info for cwm or anything else to come to the relevant conclusions. somebody's gotta say, for example, where the photo was taken, yes?

19:28:40 <jhendler>http://geonames.usgs.gov/pls/gnis/web_query.gnisprod?f_name=chicago&variant=N&f_state=Illinois&f_cnty=&f_ty=building&elev1=&elev2=&cell=&my_function=Send+Query&last_name=&last_state=&last_cnty=&page_cnt=&record_cnt=&tab=Y

19:28:40 <dc_rdfig> W: http://geonames.usgs.gov/pls/gnis/web_query.gnisprod?f_name=chicago&variant=N&f_state=Illinois&f_cnty=&f_ty=building&elev1=&elev2=&cell=&my_function=Send+Query&last_name=&last_state=&last_cnty=&page_cnt=&record_cnt=&tab=Y from jhendler

19:28:58 <jhendler> W:|Lat Long info for all major buildings in Chicago

19:28:58 <dc_rdfig> Titled item W.

19:29:05 <DanC> V:|Theory, from the Olsen timezone stuff

19:29:05 <dc_rdfig> Titled item V.

19:29:15 <jhendler> W: using [http://geonames.usgs.gov/pls/gnis/web_query.gnis_web_query_form]

19:29:15 <dc_rdfig> Added comment W1.

19:29:30 <jhendler> W: USGS Mapping service - wonder how available their DB is

19:29:30 <dc_rdfig> Added comment W2.

19:29:36 <DanC> W:now I'm getting nervous

19:29:36 <dc_rdfig> Added comment W3.

19:30:36 <DanC> W:but I guess yahoo gives lat/long for everything in its yellow pages, so I should stop worrying and learn to love the SemWeb, I guess.

19:30:36 <dc_rdfig> Added comment W4.

19:32:02 * jhendler asked USGS for the lat long of Dan's house, but it didn't have it - guess it's not so bad...

19:32:22 * DanC bets it's not hard to find

19:34:03 * jhendler well, it is close to 390xxxN 0943yyy which is the airport in Clay County... (xs and ys to make it harder to bomb Dan's house)

19:35:49 <_joshua> i have timezome shape files somewhere

19:36:16 * DanC just downloaded some timezome shape files, but doesn't grok them yet

19:36:29 <_joshua> are they esri .shp files?

19:36:45 <DanC> "grok" == "to understand well enough to convert, by machine, to RDF. or to do such a conversion."

19:37:02 <_joshua> I'm not sure how you'd convert a polygon to rdf

19:37:53 <DanC> { [ a :Polygon; :vertices (10 20 33 44 55 66) ] }

19:38:01 <_joshua> interesting

19:38:29 <DanC> salt to taste re x/y interleaving, pairs, whatever

19:38:30 <jhendler> does embed in SVG count?

19:38:58 <DanC> well, maybe...

19:39:19 <timbl> SVG on a sphere? might be stretching it ...

19:39:20 <DanC> ... but only if you give me a cwm built-in for testing whether a point is inside an SVG path

19:39:40 * DanC moans at bad pun

19:39:48 <jhendler> lots of GIS sites contain the outline info for various maps (just found all the counties of Kansas)

19:40:14 * DanC hopes for convex hull approximations, rather than real outlines

19:40:25 <jhendler> wonder if any make the underlying data available - would be possible to do the dump -- unfortunately, they all graph as

19:40:29 <timbl> I think on a plane its simpler than you might think -- you do sytg like take the parity of the number of lines it is to the right of.

19:40:52 <jhendler> err, I mean display, as jpgs/gifs etc -- so we cannot use our screen scraper on them - foo

19:41:06 <DanC> right/parity... exactly. easy enough that i'd be willing to brave it in N3

19:42:00 <_joshua> yeah

19:42:11 <DanC> I learned them as left-tests. v. hard to think of them as right, timbl ;-)

19:42:16 <timbl> DanC, re photo - you can assume the photo on my camera was taken where I was and you can assume that I was at the event at which I was a participant and which was running at the time.

19:42:45 * timbl was just assuming that you'd have to switch hands in your current state ;-)

19:42:48 <DanC> ok, if we have that sort of info somewhere, i can see how it all comes together, timbl

19:44:17 * DanC studied convex hull stuff in a seminar course by clien at uexas.edu; sorta fun, but I learned numerical analysis is not my thing

19:44:21 <DanC> cline

19:44:36 <DanC> as in the guy in all the [cline] citations in http://www.google.com/search?q=convex+hull+cline

19:44:56 <DanC> utexas.edu

19:45:37 <jhendler> MINDSWAP working on an OWL ontology to include Mpeg 7/mpeg 21 stuff plus various photo things - goal to link the photo info to content ontologies and the like

19:45:46 <jhendler> wil share when we have a decent version

19:45:55 <DanC> oops.. there seems to be another Cline that does convex hull stuff!

19:46:46 * rajiv is away: gone

19:47:14 <DanC> nifty, jhendler. We've tried telling the MPEG community they should use RDF in the past, but we didn't have a lot of ammo, and I think they went and did some XML Schema stuff. sigh.

19:47:18 * jhendler cannot come up with a "cline bottle" pun in real time.

19:48:08 * DanC goes back to foafy+calendary...

19:51:23 <DanC> H:ok, found my [foaf thingy|http://www.w3.org/People/Connolly/home-smart.rdf]

19:51:23 <dc_rdfig> Added comment H3.

19:51:54 <DanC> H:now... to connect that to [my itinerary|http://www.w3.org/2003/03dc-bos/bos-me.rdf]

19:51:54 <dc_rdfig> Added comment H4.

19:52:44 <timbl> That's a shame -- http://www.w3.org/People/Connolly/home-smart.rdf doesn't give an id for you

19:53:10 <timbl> <con:Person about="#me"> ?

19:53:20 * timbl can't talk, has no foafy stuff, no id

19:53:30 <DanC> sounds good... let's see what it costs...

19:53:36 <DanC> Overview.html:74: error: Entity 'eacute' not defined

19:54:05 <DanC> grumble... tidy to the rescue...

19:55:27 <DanC> cost was low; one-liner...

19:56:39 <DanC> the semweb travel itineraries don't refer to me in the most useful way either.

19:57:54 <DanC> <a rel='$WHAT' href=".../itin.rdf" />

19:58:31 <DanC> suggestions? i.e. relationship between ?WHO's homepage an (one of) ?WHO's itinerary(s)

20:00:00 <DanC> perhaps same as relationship between homepage and calendar full of events I participate in

20:02:26 * DanC thinks seeAlso will actually work

20:05:03 * DanC commits

20:07:17 <jhendler> hmm, interesting, playing with Open Cyc in Owl -- definitely Owl Full - they put a unique datatype property (GUID) on each class

20:07:42 <mdupont> jhendler: is that good or bad?

20:08:31 <DanC> opencyc and RDF get along great. but no way is Cyc anywhere near OWL DL.

20:09:00 <DanC> it fits well with OWL Full, yes.

20:09:06 <jhendler> agreed, although interestingly they use almost nothing from outside the Lite vocabulary -- i.e. another example of "Full Lite"

20:09:42 <DanC> hmm.. for namespace purposes, <http://www.w3.org/2002/12/cal/tzd/America/Chicago#tz> is a pain. <http://www.w3.org/2002/12/cal/tzd/America#Chicago> could be abbreviated ustimes:Chicago.

20:10:49 <DanC> but then, argh, http://www.w3.org/2002/12/cal/tzd/America/ is a directory, and using a .rdf file at the same address is dicey

20:11:26 * DanC reminds jimh that the WG didn't actually specify a "lite vocabulary"

20:11:51 <golbeck_zzz> golbeck_zzz is now known as golbeck

20:12:06 * rajiv is back (gone 00:25:19)

20:12:33 * jhendler note the double quotes around "Full Lite" -- it's my own name, not WG

20:14:51 <mdupont> what tools do you use for OWL?

20:15:14 <DanC> cwm

20:15:34 <jhendler> a bunch of our own homebrews

20:15:52 <DanC> I use Euler on a good day too

20:16:13 <DanC> and otter, and swi-prolog sometimes.

20:31:27 <DanC> timbl? seen this bit by Bijan on social meaning? http://lists.w3.org/Archives/Public/www-rdf-comments/2003JanMar/0366.html

20:32:00 <timbl> yes. (sigh)

20:32:06 <bijan> Heh

20:32:20 <timbl> Bijan, I was goin to reply to that.

20:32:50 <timbl> I have an email message half written someowhere, but I suppose we could do it here.

20:33:18 <timbl> I think you overestimate the amount of stuff the RDF spec has to define.

20:33:36 <timbl> No one is asking them to make a MT specification of human society.

20:33:47 <bijan> Niether was I.

20:34:14 <timbl> In other words, *underneath* any RDF document is a load of context aboutwhen teh message was senmt to whom etc

20:34:16 * timbl brb

20:35:02 * bijan oth

20:35:05 <jhendler> dan/tim - there's a later email by Bijan (today) at http://lists.w3.org/Archives/Public/www-rdf-comments/2003JanMar/0423.html

20:35:13 <bijan> (on tenterHooks)

20:35:21 * DanC wonders if EricM is around re social meaning, meeting logistics

20:35:39 <DanC> bijan, gonna be there on Thu 6Mar?

20:35:54 <bijan> You mean the day before I go to Japan?

20:35:58 <bijan> No :)

20:36:16 <bijan> I would like to have been, but I'm lowly

20:36:18 * DanC waves

20:36:20 <bijan> Hey jeremey

20:36:26 <bijan> I owe you a reply about the test case stuff.

20:36:42 <jcarroll> I found your last call comment compelling.

20:36:44 * timbl back

20:36:53 * em waves

20:37:15 * jcarroll waves back

20:37:16 <bijan> Thanks.

20:37:24 <bijan> We have a quorem!

20:37:35 * jhendler waves

20:37:35 <jcarroll> I am using it to structure the session at the plenary,

20:37:40 <timbl> It is however important for the RDF spec to tell you how to interpret an RDF document, or it isn't a spec.

20:37:47 <DanC> jjc, I'm gonna link http://lists.w3.org/Archives/Public/www-archive/2003Feb/att-0095/plenary from the meeting agenda

20:37:48 <bijan> jcarroll: in other news, I implemented a sound and complete shif reasoner

20:37:54 <bijan> Will add datatypes.

20:38:00 <bijan> And an owl lite parser

20:38:01 * timbl wondres which msg jcarroll is refering to

20:38:10 <em> DanC, agreed... http://lists.w3.org/Archives/Public/www-archive/2003Feb/att-0095/plenary looks good

20:38:27 <bijan> timbl: Er..I totally disagree.

20:38:37 <bijan> Revised syntax is a spec, shouldn't say a word about interpretation.

20:38:39 <DanC> bijan, you have perhaps the most coherent objection to RDF concepts section 4. pity you won't be there on 6Mar; let's prep a proxy now, shall we?

20:38:59 <em> jjac, can you make sure to close loop with all people speakiong on this and give them heads-up to this (if you havent allready)?

20:39:02 <bijan> Sure.

20:39:08 <em> err... s/jjacjjc

20:39:08 <bijan> Or, jim willing, I could come up

20:39:12 <em> err... s/jjac/jjc

20:39:20 * em keyboard skicking

20:39:29 <DanC> em, jjc is "all the people speaking", no?

20:39:36 <jcarroll> I want to finish off the detail at the end, and then I will post link on three mailing lists

20:39:41 <jhendler> do we have telecon bridge for Thurs/Fri?

20:39:58 <timbl> Bijan, as W3C director I can't condone a specification which purports to dfeine a langauge (or metalangauge) and doesn't.

20:40:00 <DanC> oh, so you intend to revise http://lists.w3.org/Archives/Public/www-archive/2003Feb/att-0095/plenary , jjc?

20:40:15 <jcarroll> I ran out of time on the issuettes

20:40:16 <jhendler> bijan, I'm willing, but that would be a trciky thing timing wise - I checked flights

20:40:24 <timbl> If the SVG spec didn't tell you how to interpret an SVG document, it would be bad.

20:40:29 <bijan> Well, i was thinking of flying up and back same day :(

20:40:34 <em> DanC, not sure.. thought jjc was still working on this

20:40:36 <DanC> jcarroll, put it under RDFCore or something so it'll stay still as you revise it, pls?

20:40:42 <jcarroll> The last five all need quotes from Concepts, Bijan or Peter or Peter or Aaron

20:40:52 * em takes call

20:40:57 <em> brb

20:41:10 <jhendler> timbl, I've heard you speak on social meaning several times, not clear to me section 4 of concepts really says what IO've heard you say...

20:41:13 <jcarroll> OK, I will put it ....

20:41:21 <DanC> thx

20:41:36 <bijan> timbl: I don't think I can debate you with this traffic :)

20:41:37 <DanC> sotd

20:41:41 <timbl> There was a good message of Pat Hayes ... dunno if i can find it

20:41:52 * bwm lurks and hopes nobody minds

20:41:57 <timbl> Bijan, same here

20:42:04 <DanC> hah! Speak Of The Devil and Status Of This Document abbreviate to the same thing!

20:42:13 <bijan> timbl: I responded to pat's message in detail

20:42:14 <AaronSw> heh, heh!

20:42:37 <timbl> :-)

20:42:46 * DanC laments the logistical challenges between bijan and timbl

20:42:54 <timbl> ?

20:42:59 * DanC heisenlament

20:43:28 <DanC> the 6Mar event is convenient for timbl, but not for bijan

20:43:46 <DanC> timbl tried to reply to bijan's mail, but didn't find time to finish.

20:43:47 <bijan> We're trying to figure out a way, danc

20:44:05 <timbl> Ok, my take on it i sin brief that the import of an RDF statement is defined by its preduicate. Anyone can define a new predicate if they can allocate its URI.

20:44:31 <DanC> "define a new predicate" meaning what?

20:44:33 <AaronSw> Is this a social meaning discussion? Someone should change the topic.

20:44:37 <timbl> The meaning of an RDF document is the total (conjunction) of the meanings of the sttaments within it.

20:44:44 <AaronSw> s/a/the/

20:44:59 * jhendler will fund Bijan to Boston for Thurs, but we're trying to see if logistics will work

20:45:16 * DanC cheers

20:45:36 <timbl> "define a new predicate" meaing - allocate a URI to it and give an (english or otherwise) definition of what it means, with s and o as parameters.

20:46:23 <DanC> but not any english defn will do, right? "use of this predicate causes all statements about the subject to have their sense inverted" is no good, right?

20:46:25 <timbl> Like s my:color o means that the thing s is mostly painted the color c where c is an srgb numeric triple in hexformat15.

20:46:49 * bwm can't find Pat's social meaning email in the archives, wonders if bcc's aren't archived

20:48:12 <timbl> There are some statements you can try to make which would be paradoxical, or just wrong.

20:49:23 <timbl> "use of this predicate causes all statements about the subject to have their sense inverted" is indeed not something you can say, because statements are defined by their own predicates and are oethrwise indepndent of other people's definitions.

20:49:25 <DanC> welcome to the party, pfps

20:49:35 <timbl> well, status of this document!

20:49:43 <DanC> hoot

20:49:45 <timbl> ;-)

20:50:14 <jcarroll> I can put the pleanry stuff in WWW/2001/sw/meetings/tech-200303/social-meaning.html

20:50:31 <DanC> very well, jcarroll

20:50:50 <jcarroll> (I don't much like being jcarroll - but jjc and jeremy were both taken)

20:51:11 <timbl> I am sure there are people who could talk about constrainng the interpretations

20:51:15 <jhendler> danC what time is the social meaning session (working on Bijan's logistics)

20:51:21 * DanC didn't pick the meeting pg address, btw... the /meetings/ is one too many for my tastes

20:51:29 <jcarroll> 16-18.00

20:52:01 <bijan> Hey peter

20:52:16 <DanC> " 16:00 - 18:00 Second afternoon session

20:52:16 <DanC> formal meaning meets social context" -- http://www.w3.org/2001/sw/meetings/tech-200303/

20:52:26 <timbl> It is important to note that you can't say my:car is such that all statements of the form my:car ?c ?b mean that my car is a wonderfully fast and desirable car.

20:52:27 <DanC> I expect it to go into the evening

20:53:21 <DanC> my point was that I don't see how we can be clear about which english property definitions are OK and which are not in the RDF spec.

20:53:37 <DanC> or maybe I do...

20:53:48 <timbl> So Bijan and PPS, you both are of teh same mind that RDF documents should not be construed as meaning anything?

20:53:55 <DanC> they can only constrain the possible worlds, i.e. eliminate interpretations.

20:54:05 <bijan> That's not at all what *I* said, and I can't believe that's what Peter would say

20:54:17 <timbl> You can't define which definitions are OK.

20:54:22 <timbl> You don't have to.

20:54:33 <timbl> The XML spec doesn't define the set of all valid XML applications.

20:54:42 * em returns from call

20:54:48 <timbl> The RDF spec doesn't have to defibne "OK" RDF property definitions.

20:54:50 * em catches up with log

20:55:28 <timbl> Well, you both want sction 4 to be struck, and section 4 seems to eb th only place that the meaning of an RDF document is mentioned?

20:56:01 <bijan> which meaning?

20:56:03 <timbl> Maybe it is not well written, but there needs to be something somewhere.

20:56:22 <timbl> the meaning.

20:56:28 <DanC> gee, tim, there's a whole lot about the meaning of RDF documents in the semantics doc.

20:56:51 <jcarroll> That meaning does not relate to anything outside the semantic domain

20:57:04 <jcarroll> (at least not without section 4)

20:57:31 <jcarroll> I am not convinced that section 4 is necessary though.

20:57:41 <DanC> wow... i hand't seen "0.1 Specifying a formal semantics: scope and limitations" before. http://www.w3.org/TR/2002/WD-rdf-mt-20021112/#intro

20:58:10 <timbl> Maybe the confusion is that the OWL and RDFS terms (properties and classes) can be defined quite mathematically, but in general RDF properties are defined in connection to the real world, an indeed convey real world meaning.

20:58:17 <reagle> huh, an xmlliteral needs to have a root element of rdf-wrapper?!

20:58:17 <bijan> Tim: do you mean the one interpretation of the graph that corresponds to the actual state of affairs something soemthign seomthing?

20:58:24 <reagle>http://www.w3.org/TR/2003/WD-rdf-concepts-20030123/#dfn-rdf-XMLLiteral

20:58:24 <dc_rdfig> X: http://www.w3.org/TR/2003/WD-rdf-concepts-20030123/#dfn-rdf-XMLLiteral from reagle

20:58:32 <reagle> example nine does not have that root element, so what's the deal there...?

20:58:38 <reagle>http://www.w3.org/TR/2002/WD-rdf-syntax-grammar-20021108/#section-Syntax-XML-literals

20:58:38 <dc_rdfig> Y: http://www.w3.org/TR/2002/WD-rdf-syntax-grammar-20021108/#section-Syntax-XML-literals from reagle

20:59:01 <jcarroll> It was a legacy problem :(

20:59:08 <DanC> rdf-wrapper is an artificial construct to make a piece of XML content into a well-formed wrapper before feeding it to other algorithms, reagle

20:59:21 <timbl> "The semantics given here restricts itself to a formal notion of meaning which could be characterized as the part that is common to all other accounts of meaning, and can be captured in mechanical inference rules. " - semantcs doc, quite appropriately

20:59:55 <ericP> reagle, have you started the c14n discussion?

21:00:03 * ericP just got here

21:00:03 <DanC> is that section not enough, timbl?

21:00:17 * timbl hunting for Bijan's text

21:00:31 <reagle> ericP, DanC answered, "rdf-wrapper is an artificial construct to make a piece of XML content into a well-formed wrapper before feeding it to other algorithms, reagle"

21:00:39 <DanC> ericP, reagle, a discussion of social meaning, involving jjc/timbl/bijan/pfps has sorta... spontaneously erupted

21:00:40 <reagle> which is what i understood, but isn't how the spec reads to me

21:01:04 <reagle> DanC, are you asking me to not talk on this channel?

21:01:15 <bijan> Scheduling folks: The latest it seems I can leave boston is 7:30, which includes getting to the airport

21:01:35 <DanC> no, reagle, just noting there's competition for attention

21:01:41 <reagle> ok

21:01:44 <timbl> Bijan, the problem is that if you try to talk formallyabout the non-formalized meaning of RDF statements, you ask yourself immediately for a formal systemto describe society. This doesn't exist and isn't achievable.

21:02:06 <jhendler> IMHO, I'd like to see section 4 catch what Tim is saying - that the NL (comments etc) "situates" the formal meaning, as opposed to is part of it - a rewrite rather than a deletion

21:02:19 <reagle> so ericP, I don't buy that DanC's characterization agrees with the spec.

21:02:20 <DanC> includes in which sense... flight leaves at 7:30p?

21:02:21 <ericP> reagle, would you like to wait a bit? i'd like to hvae this discussion in a public forum, but some time when it gets attention.

21:02:24 <bijan> Well, i didn't ask for that tim, and didn't try to.

21:02:34 <reagle> ok, let's wait till this wraps up

21:02:41 <ericP> roger

21:03:18 <bijan> In my reply to pat, I indicated that there was probably some claims, preferably damn weak, that I could see as reasonable, and perhaps helpful.

21:03:33 <timbl> That is, any RDF document sits in a context in which a whole bunch of assumptions common betwen sender and receiver define world about which they are talking. You can't model that/ But then within that context, an RDF message conveys some information.

21:03:35 <jcarroll>http://www.w3.org/2001/sw/meetings/tech-200303/social-meaning

21:03:35 <dc_rdfig> Z: http://www.w3.org/2001/sw/meetings/tech-200303/social-meaning from jcarroll

21:03:46 <timbl> That information is defined by a bunch of english in schemas in most cases.

21:04:13 <timbl> It is the conjunction of the meanings of the statemeents, and each statement's meaning is defined by the predictae.

21:04:20 <bijan> But *that* isn't anywhere near what's in section 4 nor what's in what you're saying.

21:04:46 <jhendler> ditto.

21:05:00 * DanC wonders if going past 26 chump items has been tested ;-)

21:05:07 <dajobe> DanC: yes, AA is next

21:05:19 <bijan> Plus, frankly, social meaning can usually take care of itself. IT doesn't need our help :)

21:05:24 <sandro> logger_1, pointer?

21:05:24 <sandro> See http://ilrt.org/discovery/chatlogs/rdfig/2003-02-26#T21-05-24

21:05:35 <timbl> Well, maybe I read it as saying what i want to hear.

21:05:46 <bijan> Yes, and that's not good enough for a specification.

21:05:54 <bijan> A normative part of a specification

21:06:03 <jcarroll> Personally I think that the weak statement on which their might be consensus would be so weak as to not worth having.

21:06:12 <bijan> jcarroll: agreed.

21:06:21 <bijan> That's my "vacuous" branch :)

21:06:42 <DanC> social meaning has not been taking care of itself in the P3P case. Some guy, a laywer, actually, is saying folks aren't accountable for what they put in the p3p files on their web sites.

21:06:48 <jcarroll> Oh no we tried that and it wasn't weak enough - much less than that

21:06:53 <timbl> Ok, so suppose we remove section 4 as badly written, then where do I find in eth RDF specs what I need?

21:07:03 <jcarroll> You don't.

21:07:08 <bijan> timbl: i really just don't know what you need, even.

21:07:19 <bijan> DanC: No, that's social meaning taking care of itself

21:07:20 <timbl> (a) That the meaning of an RDF document is the conjunction of the meaning of the statements.

21:07:26 <jcarroll> rdfms-assertion was addressed by the text in section 4

21:07:27 <pfps> I agree with bijan, Section 4 doesn't say anything near ``Any use of RDF sits within some externally-given (usally socially-given) context, which does whatever it does with no help from RDF.'''

21:07:29 <bijan> It's not the social meaning *you* want, or thought you were getting.

21:07:37 <timbl> (b) That the menaing of the statemenmt is defined by the definition ofthe predicate term.

21:07:57 <reagle> (btw: in the P3P he's a lawyer and was pointed to the specific text in the spec that negated his point but continued to make his claim anyway)

21:07:57 <timbl> "social emaning" is not a well defined term.

21:07:58 <danb_stilloffsic> rdf social meaning as i see it: we all agree RDF documents are supposed to have propositional content, ie. carry/encode/represent claims about the world, which could turn out on examination to be true, or false. For this to be deployable, we need to tie RDF's abstract structures to the world, so we can assess which arrangements of the world would satisfy the propositions encoded in some RDF.

21:08:00 <bijan> timbl: you have to disambiguate "the meaning" for me

21:08:10 <bijan> timbl: yes, adn that's a HUGE part of the problem.

21:08:17 <reagle> clear text can help dispute lawyers, but can never stop them arguing FOO if that's what they want to do...

21:08:18 <DanC> no, pfps? "Such an assertion should be understood to carry the same social import and responsibilities as an assertion in any other format."

21:08:37 <timbl> So lets try to do this without saying "social meaning" and just saying "meaning".

21:08:46 <bijan> No.

21:08:49 <danb_stilloffsic> +1 on losing the 'social'

21:08:55 <bijan> Because there's distinctions here that need to be maade

21:08:59 <reagle> DanC, that sounds like a dsig spec/law

21:09:06 <bijan> So yes, scratch social meaning, but "meaning" is JUST as ill defined

21:09:10 <jcarroll> Other words are pragmatics or effective meaning

21:09:10 * timbl becoems clearer that section4 means many things to many people

21:09:26 <bijan> pragmatics: no, I don't think

21:09:37 <bijan> "affective meaning" is worrisome to me

21:09:43 <reagle>http://www.w3.org/Submission/2003/01/Comment

21:09:43 <dc_rdfig> AA: http://www.w3.org/Submission/2003/01/Comment from reagle

21:09:46 <DanC> many/many: ah... then this has been worth the price of admission

21:09:51 <timbl> Bijan: Why do you wnat >1 meaning?

21:09:55 <reagle> In fact, Article 5.2 rules that Member States shall ensure that an electronic signature is not denied legal effectiveness and admissibility as evidence in legal proceedings solely on the grounds that it is: ... in electric form

21:10:04 <bijan> Because there *is* more than one

21:10:12 <timbl> Give ma an example.

21:10:17 <bijan> At the very least, you have to indicate what you are excluding.

21:10:19 <timbl> or even me

21:10:25 <bijan> Denotation vs. connotation

21:10:33 <bijan> RDF specs don't give you even denotation

21:10:53 <bijan> I don't actualy believe the URI specs give you denotation

21:10:55 <pfps> No. The external (social) context doesn't have to abide by the treatment that RDF gives to anything. For example, some social context might disavow all ``insults'', while some other social context might treat any use of prohibited words as deadly insults.

21:11:08 <timbl> The RDF specs refr tp the URI specs. The RDF specs do not have to and should not take over the job of the URI specs.

21:11:22 * bijan agrees with pfps

21:11:57 * DanC doesn't think the URI spec (+related specs) nails down a single denotation for each URI

21:12:15 <dmiles> i guess processing (inference) on RDF is a little simplier then english.. but yet has most of the same problems

21:12:20 * danb_stilloffsic neither, unfortunately (though maybe we can pretend...)

21:12:26 <danb_stilloffsic> danb_stilloffsic is now known as danbri

21:12:26 <bijan> timbl: There is sufficient disagreement about what uris denote, even what they *can* denote, that one can fairly say its not nailed down.

21:12:50 * sandro wonders if DanC thinks the URI specs nail down the fact that there IS a single denotatation for each one....

21:12:54 <timbl> There are social contexts which, if you like, quote an RDF document and express some syntactic or other permutation on it. They don't make reference to THE meaing ofthe document. But there is still THE meaning ofthe document, even if the social context uses the syntax ofthe document for wallpaper.

21:13:17 <bijan> timbl: general "the meaning" is a nonstarter.

21:13:18 <DanC> no, it's not a matter of lack of consensus, bijan. it's a matter of the "single intended model" fallacy.

21:13:43 <timbl> Bijan, The problems with URis may be rife, but those are problems with URIs. And you overlook the fact that we are using them fairly effectively.

21:13:47 <DanC> only in a newtonian approximation, sandro

21:13:48 <bijan> DanC: sorry, I meant that as orthoganal to your comment

21:13:54 <bijan> No tim, I'm not

21:13:57 <dmiles> at least RDF to Logical Infernce is easier to code then English to Logical Inference language :)

21:14:08 <bijan> I'm sayin gwe can use them effectively without them getting denoation nailed down (for exmaple)

21:14:24 <bijan> And with wide disagreement on their semantic and pragmatic properties

21:14:26 <bijan> So too for RDF

21:14:35 <bijan> Hence, no need for a section 4 for RDF to be a useful specification

21:14:38 <bijan> QED :)

21:15:44 <timbl> But Bijan, I think you are ignoreing a lot of conventions which allow stuff to work. Though you use them implicitly all the time. It is unfair not to explainthem for others.

21:16:00 <sandro> Bijan, how would you approach solving the P3P problem? We want people to be able to publish things on the web and have their statements taken seriously. Maybe not too seriously without a XADES digital signature, but still fairly seriously.

21:16:09 <timbl> The fact that to understand s p o you look up p is something you just assume.

21:17:05 * sandro still doesn't see any reason to single-out P in SPO, but that's a different subject.

21:17:07 <timbl> The fact that the OWL spec defines the meaning of statments written with predicates in the OWL namespace, for example.

21:17:22 * em chuckles at sandro's pun

21:17:39 <timbl> sandro, I define my:car to be such that my:car ?p ?o always means "My car is wonderful".

21:17:39 <bijan> timbl: i'm not ignoring the conventions

21:17:44 <dmiles> mainly P holds the meanings and SO are just denotations that have built in identity crisis

21:17:46 <timbl> Good

21:18:06 <timbl> (good, bijan, not ignoring conventions. can we write them down?)

21:18:08 <bijan> I'm pointing out that those convention eithe 1) evolve in practice (not in a priori specification) 2) are outside your juristiction

21:18:22 <bijan> yes, we do all the time. Rules, contracts, specs, laws.

21:18:26 * eikeon finally gets sandro's pun :)

21:18:34 <pfps> I think that I would modify Bijan's statement to something along the lines that RDF is suitable only in contexts that are ``assertional'', i.e., ones that don't have some of the weirder (but very common) aspects of social conventions

21:19:02 <timbl> There are many bits which are outside the RDF spec jurisdiction. There are VERY FEW which are inside. But RDF has a duty to define those.

21:19:43 <bijan> sandro: "solve the p3p problem"? You want people to take stuff seriously, you can't jsut SAY "you must take stuff seriously"

21:19:48 <bijan> You can, but it prolly won't work.

21:20:09 <bijan> timbl: well, I've not seen a narrow enough range that was 1) interesting, 2) useful to make a normative section about

21:20:47 <timbl> RDF for example is NOT responsible for defining to what extent the sender of email is responsibl for the content of email, or a web page editor or publisher or domain owner can be held accountable for the information on it, and so on.

21:21:15 <bijan> And I'm *very* leary about the talk of assertional vs. non-assertional contexts where that isn't handled by the rest of the docs at all.

21:21:16 <danbri> pfps, i've found uses for rdf in non-assertional contexts (eg. queries, representing things that aren't true (fiction, artwork etc)).

21:21:17 <timbl> It is NOT responsible either for arguing whether "metalic grey" is a color.

21:21:23 <sandro> What else can you do, Bijan? Saying it is 80% of the magic spell which makes it happen. The 20% is in the the consensual illusion we call society.

21:21:38 <bijan> Uhm. Well, you may have it backwards.

21:21:56 <sandro> True - my spells sometimes blow up in my face

21:22:09 <timbl> I think I have said I want to eb able to find two thinsg in the spec soemwhere.

21:22:14 <sandro> And passing laws would be somewhat stronger, but is less practical.

21:22:23 <bijan> furthermore, I thought the "Semantic web" thing was *not* about legistlating best practices

21:22:27 <sandro> And perhaps much more dangerous.

21:22:32 <pfps> hmm, I guess I should have said context that have a form of representational transparency, i.e., it is the meaning that matters, not the form of the expression

21:22:38 <timbl> We are *not* talking aboyt legislating vbest parctices.

21:22:42 <bijan> I think it's *futile* to declare things that well, you have no authority to declare

21:22:55 <timbl> Like?

21:22:57 * danbri scribbles an example -- http://lists.w3.org/Archives/Public/www-archive/2003Feb/0100.html -- wondering what it would take to fix the meaning of 'http://xmlns.com/foaf/0.1/uncle' to approximate the rules we use for 'uncle' in english.

21:23:07 <dmiles> sematic web is about constructing new course materials for DBAs

21:23:10 <bijan> Like the legal significance of a piece of rdf

21:23:20 <sandro> This is a "recommendation from the W3C (and its 400+ members)" That's a very nice way to declare something.

21:23:20 <dmiles> (errm i thought)

21:23:33 <bijan> Either that's already governed by existing law and practice (as I believe) or it isn't

21:23:34 <timbl> I am *NOT* asking anyone to specify the legal signicance of a bit of RDF!

21:23:41 <bijan> if it isn't, nothing you say will make i tso

21:23:58 <bijan> If it is, then there's nothing to specify

21:24:07 <timbl> Exactly.

21:24:17 <bijan> Doesn't have to be a bit, it can be the general relationship.

21:24:39 <timbl> Ok, her's an example of a simpler case. Let's define eng13, a new langauge whcih is defined as english rot13-d.

21:24:53 <timbl> Suppose you write a spec for eng13.

21:25:25 <timbl> You say that the meaning of an eng13 document is the meaning of the document yuo get by rot-13ing each character of the eng13 document. You are done.

21:25:58 <timbl> You don't say that, and eng13 is usefless.

21:26:25 <bijan> Hmm. Rarely do I see the explicit metalinguistic version

21:26:31 <bijan> There's a reason for that

21:27:54 <jhendler> timbl, I don't think the eng13 analogy is right -- the DOCUMENT with the eng13 doesn't say that the meaning is via the rot -- the spec does

21:27:57 <bijan> In programming spec, "same meaning" is usually referring to the operational or denoational meaning of a programming langauge construct (for example)

21:28:01 <jhendler> not the same thing at all.

21:28:09 <eikeon> danbri: Wouldn't your foaf test cases be another example of non-assertional contexts?

21:28:21 <timbl> But think of specs of OFX, the online financial exchange format. That defines what an OFX resource means as a bank statement

21:28:33 <timbl> The spec is essential to the use fo OFX for communicating.

21:28:47 <bijan> Plus, I'm a lot *LESS* leary of *specific* meaning mappings

21:28:48 <jhendler> i.e. the definition of RDF says that the meaning of the RDF document depends on what is in the strings, which makesit possible for the document itself to "change" its meaning

21:28:52 <jhendler> OFX a better example

21:28:54 <timbl> People using RDF forthat will not define a langauge - they will define a vocabulary and RDF as the langauge.

21:29:01 <bijan> I.e., it's easier to get those right, or right enough.

21:29:13 <bijan> Though there are a number of people who disagree even with that

21:29:21 <bijan> Pointers in my response to Pat

21:29:33 <timbl> Well, the world works by peopel agreeing on specs like OFX.

21:29:35 <danbri> Yes, good point eikeon. the rdf at http://rdfweb.org/2003/02/a-z/ isn't asserted by the publisher, its just online. I'm prepared to think of this as those documents lying (but not the webmaster lying).

21:29:55 <bijan> timbl: no, it works by people doing lots and lots and lots of things :)

21:30:18 <timbl> The web architecture is that it works by people writing and reading specs.

21:30:30 <timbl> The role of the RDF spec is to be part of that architecture,

21:30:38 <bijan> Tendentious opinion on your part. What's the proof?

21:30:45 <timbl> The fact that other things happen does nbot need to be addressed by the sepc.

21:31:12 <timbl> Proof? Assertion.

21:31:49 <timbl> Definition ;-)

21:31:52 <bijan> Tim, sorry, should we really be making a technical normative recommendation based on your mere assertion?

21:32:13 <bijan> Granted, you have the social power to make that so.

21:32:18 <bijan> As director and all

21:32:40 <bijan> But I'm at a loss to see how you'll enforce conformance.

21:32:41 <timbl> Without my assertion that the web works by normative recommendations, you don't have to make any at all, as the sepcifciation would have no role anyway.

21:32:42 <DanC> web architecture works cuz it's easier and rewarding to do what everybody else is doing than to not. The specs help maybe a little, but they're hardly the real backbone. e.g. specs that put too much burden on the wrong parties just get ignored.

21:33:23 <bijan> timbl, not at all, as danc just enuciated, there's other views of the *very useful* role of normative specs

21:33:26 <dmiles> i think getting normative specs going is a good thing.. whether the acomplish the intended goals.. is that the question?

21:33:31 <timbl> That is true, but each spec is written as a peice of the puzzle in attem[pt to make as cleran a machine part as possible.

21:34:02 <bijan> Well, that's maybe your intention

21:34:11 <bijan> It's not clear that that's the real role

21:34:30 <bijan> Which is a nice illustration of the gap between intended meaning/role and "sentence" meaning/role

21:34:35 <timbl> Give me an example of another useful role of normative specs.

21:34:43 <bijan> Regulative ideal

21:34:50 <bijan> Inspiration

21:34:54 <bijan> Branding

21:35:15 <bijan> Description of practice that tries to guide practice

21:35:17 <timbl> What makes such specs "normative"?

21:35:31 <bijan> Their normativity? :)

21:35:57 <bijan> let me roll things back a moment.

21:35:59 <bijan> <timbl> Without my assertion that the web works by normative recommendations, you don't have to make any at all, as the sepcifciation would have no role anyway.

21:36:09 <dmiles> using chr(32) between words is like a normative spec right?

21:36:11 <DanC> the rule of law is what makes specs normative, no?

21:36:28 <DanC> i.e. consent of the governed, due process, and all that.

21:36:34 <bijan> This is clearly not quite right. The web "works" by its structural and operational architecture

21:36:35 <danbri> re social/legal, all I want to know is: if I have in my hands an RDF document, what procedure(s) can I follow to examine it and the world, to discover whether it truly describes the world vs fails to do so.

21:36:53 <bijan> that may *or may not be* correctly described by some specification

21:37:23 <eikeon> danbri: How about adding some metadata to the test documents that indicates the intent of the assertions. And documents in general... then it is up to the consumers of the documents to key off the metadata appropriately.

21:37:26 <bijan> I.e., the normative spec can, often does, and perhaps should, lag actual practice

21:37:29 <timbl> But the interesting thing is that there is a single Normative meaning.

21:37:39 <bijan> Er...?

21:37:42 <bijan> of what?

21:38:25 <danbri> eikeon, I might add a 'util:Bogus' type to each doc...

21:38:28 <timbl> <danbri> re social/legal, all I want to know is: if I have in my hands an RDF document, what procedure(s) can I follow to examine it and the world, to discover whether it truly describes the world vs fails to do so.

21:38:28 <timbl> [16:36] <bijan> that may *or may not be* correctly described by some specification

21:38:36 <timbl> Correctly?

21:38:36 <bijan> (I find it hard to keep going in the face of appeals to "the single meaning". I don't know what you mean. Plenty of non po mo philosophy, for exmaple, would count you as off the page)

21:38:47 <bijan> That is, the prctice doesn't conform to the spec.

21:39:01 <timbl> I am quite aware of the philosophy of meaning through use.

21:39:08 <DanC> "po mo"

21:39:12 <bijan> post modern

21:39:23 <bijan> That's not the only one tim.

21:39:38 <timbl> However, I would respectfully submit that th eeffectiveness of teh Internet as a medium has relies on a social convention

21:39:45 <bijan> In fact, "no single meaning" is perhaps pretty close to a consensus view :0

21:39:50 <bijan> timbl: *lots* of them!

21:39:55 <bijan> And specs play a role in them!

21:40:01 <bijan> So let them play their role :)

21:40:15 <bijan> They will anyway

21:40:20 <timbl> ... that users of the Internet (or their soofwtare agetnts) sare a single meaning for many terms and bitfield valeus which is defined in a series of normative specifications.

21:40:27 <danbri> everyone _pretends_ there's a single meaning, even if we never agree on it. Same as natural language imho...

21:40:32 <bijan> Er...

21:40:37 <bijan> I don't pretend such.

21:40:47 <eikeon> danbri: That would be good... I and others can then key off it so that the test foafers do no show up in our foaf apps ;)

21:40:53 <timbl> However, I would respectfully submit that th eeffectiveness of teh Internet as a medium has relies on a social convention that users of the Internet (or their soofwtare agetnts) sare a single meaning for many terms and bitfield valeus which is defined in a series of normative specifications

21:41:03 <DanC> no, not a single meaning, timbl. I think they share certain constraints on the possible meanings, but the constraints don't narrow it down to one.

21:41:04 <bijan> Disagree

21:41:11 <bijan> Agree with DanC

21:41:21 <bijan> And how tight an agreement depends on context and interest

21:41:33 <bijan> How tight an agreement is necessary

21:41:36 <danbri> yup re danc's comment

21:42:12 <timbl> Yes, where I say "meaning" you can say "set of meanings which are for all intends and purposes equivalent for the matterat hand but area really different".

21:42:25 <bijan> but tim, frankly, this very discussion makes my general point against section 4

21:42:27 * jhendler is puzzled because he agrees with most of what timbl says and most of what Bijan's says, but they seem to be disagreeing :->

21:42:31 <bijan> Oh, so you have >1 meaning?

21:42:56 <bijan> Or one version of it.

21:43:02 <timbl> No, I think your insietnce on having >1 bijan:meaning for the one tim:meaning is fiddley and not useful.

21:43:26 * danbri revisits http://plato.stanford.edu/entries/content-externalism/

21:43:30 <bijan> Uhm.

21:43:45 * rajiv is away: gone

21:43:48 <timbl> Example. The 5th bit in the 6th byte of an X10 packet may have the meaning that if 1, the light is turned on, and if "0" it is turned off.

21:43:49 <eikeon> jhendler: lol :)

21:44:08 <bijan> "the light"?

21:44:13 <bijan> "on"?

21:44:16 <bijan> "off"?

21:44:22 <timbl> the lights addressed bythe bits in the rest of the packet.

21:44:23 * bijan can play this game, doesn't really want to.

21:44:44 <bijan> tim, i think we need a methodology

21:45:06 <danbri> does working from test cases make sense for this debate?

21:45:15 <bijan> If there were test cases, I'd welcome that

21:45:40 <bijan> Beacuse, to get back to the spec, I don't see anything resembling tim's position in section 4 (I say again)

21:45:41 <timbl> In practice , the lights turn on and off, and people agree on what the bit means, and it is realy useful that it is written down somewhere, and Bijan, is you want to sit thee saying "light"? "meaning"? "on"? you can, but I'm moving on.

21:46:18 <bijan> well, i *don't* want to play that game.

21:46:24 <bijan> but it feels like the next move

21:46:35 <timbl> You played it at 16:43 ET

21:46:48 * bijan can play this game, doesn't really want to.

21:46:50 * eikeon thinks that attempting to constrain things to a single meaning is a good thing.

21:46:53 <bijan> See right after.

21:46:55 <timbl> Ok, so section4 uses too may philosophical worsa nd should not say social.

21:47:20 <bijan> tim, take the first sentence if 4.4:

21:47:26 <bijan> """Human publishers of RDF content commit themselves to the mechanically-inferred social obligations."""

21:47:48 <bijan> What does that attempt to impose on me, joe RDF user.

21:48:21 <timbl> What I take that to mean, as Pat explained, is that if you tell me that the car you are selling me is a LotusCar, and LotuCar is a subClass of IBMCar, then I can expect to get an IBMCar.

21:49:19 <timbl> It imposes on you as a RDF user that when you use a term, you should use it according to how it is defined.

21:49:21 <bijan> If that's all that's expected, I don't see how the *absence* of an explicit statement about it would prevent that from being the reasonable expectation

21:49:30 <bijan> What defines it?

21:49:38 <bijan> What's the canonical definition?

21:50:14 <bijan> What happens if the definition changes? am I still responsible?

21:50:35 <timbl> This is not a formal system, remmeber. However the web is wonderful in that it provdies a way fo getting information which is authoritative about terms.

21:51:01 <danbri> OK, test case. An RDF document: http://rdfweb.org/2003/02/a-z/c-foaf.rdf -- how would the world need to be arranged for that document to be a true description of the world?

21:51:04 <bijan> This isn't about formal systems. It's about being even as tight as laws, regulations and the like

21:51:44 <timbl> Bijan, when you ask about changes, and lying, and misunderstanding, then you go into the areas where the law can step in with lot sof social systems and precedent for resolving such issues - quite effectively.

21:51:49 <DanC> c-foaf looks pretty easy.

21:52:11 <jhendler> Tim - what you are saying makes more sense to me than what section 4 says -- it is talking about the langauge in the comments, you're talking about the mapping of the meaning of the document to real world expectations

21:52:18 <jhendler> these don't look the same to me.

21:52:19 <timbl> The only think which can *** it up is if we drop the ball and give the lawyers the ability to say "this is just and RDF statement, it has no meaning".

21:52:22 <bijan> timbl: but the spec says I'm committed to the mechanically-inferred social obligations

21:52:27 <DanC> I'd have to find somebody who'd say "yes, I call myself 'Christine Christianson'. and watch me update my weblog" and such.

21:52:51 <bijan> Er..first, I don't think a section 4 wouuld take away their ability if they have it

21:52:56 <danbri> it should be easy. subtleties are when we get into things like 'uncle', and the fact that the rdfs:label and rdfs:comment can't capture the rules for uncle-ness (though N3 could).

21:53:24 <bijan> I think I need to study this p3p case. Pointer?

21:53:33 <danbri> ...so we get to pick amongst the different aspects of meaning re 'foaf:uncle', and test different arrangements of the world against our intuitions.

21:54:04 <DanC> I think the semantics of foaf:unlce are sufficiently grounded. I could get, e.g. that newspaper writer to testify in court as to what it means.

21:54:06 <bijan> tim, plus, I see ambiguity in your "has no meaning", again: 'uninterpreted', 'lacking legal/social import'

21:54:24 <timbl> I have spoke on the phone with a very arrogant spammer who told me taht yes, he didput my email address in the "From:" field, but he could put whatever he pleased in the from field, becauze it was just an internet message, and internet specs aren't law.

21:54:30 <bijan> the former is typically false (i.e., rdf will have interpretations in a context)

21:54:49 <bijan> The latter is exactly what lawyers, judges and legistlatures should hve the power to determine

21:54:53 <bijan> Well, and *do* have that power

21:55:00 <bijan> Regardless of what a spec says.

21:55:07 <bijan> People entering into contracts, too.

21:55:08 * jhendler marvels that Tim phones spammers :-> bet that takes a lot of his time

21:55:50 * timbl points out that if veryone did it they would give up. And it makes him feel better and them feel worse.

21:55:52 * DanC notes there's nothing special about lawyers/judges etc. except that they're sorta more convenient and scalable than angry mobs and such. power rests with the people

21:56:09 <bijan> DanC: yes

21:56:21 <bijan> I was taking the power of layers, etc. to have been delegatted power

21:56:36 <jhendler> we have a mailer which builds the mail metadata from the message, so if it was "from: hendler@foo" the rdf would include me as the dc:creator - are you arguing the RDF would somehow have more import than the email itself tim?

21:57:14 <timbl> When you have a model of delegated power, etc, you can have a model of architecture like http://www.w3.org/DesignISsues/Stack in which specs nest, and life is orderly.

21:57:48 <timbl> if your philosophy of life is anarchic, then I'm not sure what you would be writing a w3C spec for.

21:58:05 <timbl> jhendler, no.

21:58:07 <bijan> Er..

21:58:10 <DanC> I think timbl will eventually win against the spammer, because, as I said earlier, it's valuable to The People if we can pretty much count on the connection between mailbox names, From: fields, messages, and people.

21:58:36 <bijan> I thought a major point I mades is that I take section 4 as written to arrogate illegitmately (or attempt to) power that belonged elsewhere

21:58:54 <bijan> So, in that sense, I'm resisting your power grab, tim :)

21:59:01 <AaronSw> value to The People has to be balanced with enforcement costs

21:59:30 <DanC> I'm a little bit afraid we may have to make it more costly, computationally, to commit From: fraud, becuase there seems to be considerable economic benefit to forging From: headers, yes, AaronSw

21:59:34 <AaronSw> and judges aren't just scalable angry mobs, they're specifically designed to be undemocratic

21:59:37 <jhendler> yes, I agree with what DanC says re spammer, but it is because the "social meaning" part of is happening off line - in either law courts or courts "of the people"

21:59:45 <timbl> But if the RFC822 spec writers had just sat there saying "From"? "light"? "message"? "deliver"? and not written what the field meant because they didn't fiedl; that there was really any one true meaning to any RCC822 header from the philosophical statendpoint, we would nothave email.

22:00:09 <DanC> I think we would.

22:00:20 <bijan> <sigh/> /me just wants to say that I wasn't intending to be all that restrictive or substational with his list of empowered entities

22:00:32 <bijan> :)

22:00:39 <DanC> if we could get software deployed without specs, I think the value of the software and the protocols would perpetuate itself.

22:01:14 <AaronSw> Historically, I think you'll find that software deployed without specs has been more successful than specs deployed without software

22:01:23 <jhendler> tim, I'm confused again about meaning of trhe spec vs. meaning of the document -- i.e.the RFC for the email says "from: should be where the sender's email goes" - the implementors of mailers understand the spec, and build it that way -- absolutely social meaning.

22:01:23 <DanC> specs work iff they lower the cost of the "ooh! how do I join the fun!" process.

22:01:45 <bijan> And I believe more strongly in after the development and deployment specs

22:02:04 * DanC likes concurrent code/doc/test

22:02:09 <jhendler> but the mail itself useing the word "from:" doesn't define the term - if the RDF called it "ZIppy" the implementors might still go along, and it would hold the same content

22:02:22 <jhendler> and you'd yell att he spammers for putting your name in the "zippy" field

22:02:35 <bijan> tim, btw, are you wearing your director hat in this conversation?

22:03:06 <bijan> At one point you explicitly said, "as director..."

22:03:21 <DanC> he's listening as director, but not speaking as director. 1/2 ;-)

22:03:35 <bijan> heh.

22:03:52 <bijan> Well, there is one place where he did speak as director, which is a key one

22:04:10 <DanC> The Director speaks only by email to w3c-ac-members, I believe.

22:04:28 <kendallc> heh, sorta like Moses and Y**H...? :>

22:05:09 <bijan> 20:39:58 <timbl> Bijan, as W3C director I can't condone a specification which purports to dfeine a langauge (or metalangauge) and doesn't.

22:05:14 <bijan> How do I interpret that, danc?

22:05:29 <kendallc> (as a threat or as a promise?! :>)

22:05:42 <bijan> It sounds like if I got my way with this issue, RDF wouldn't become a recommendation.

22:06:24 <timbl> You can never tell what will or what won't become a recommendation.

22:06:29 <timbl> :-)

22:06:45 <timbl> The weirdest things have....

22:06:50 <bijan> Indeed.

22:06:52 <timbl> DanC was right ... speaking with hat off.

22:06:59 <bijan> Including that?

22:07:10 <jhendler> I took that statement as a contrapositive - i.e. "As Director I can't condone..." but since he is not currently wearing his director hat he could condone it... :->

22:07:26 <bijan> because, even if the "real" meaning is that it's up in the air, it certainly alters my perceived risk

22:07:42 <bijan> Which could lead me to yeild where I don't think it's wise in order to reach compromise :)

22:07:56 <DanC> I took it as "eventually, I have to make a decision on this. And right now, I can't see how I could agree with an RDF spec that doesn't have (something like) concepts secion 4"

22:08:41 <timbl> the factthat you got some characters from my IRC cleint which contained "as W3C " etc doesn't necessarily have any meaning.

22:08:50 <timbl> That is, it of course has many menaings.

22:08:56 <timbl> "director"?

22:09:17 <bijan> Hmm. I guess you didn't say *you* didn't want to play that game!

22:09:19 <bijan> "light"?

22:09:25 <bijan> ;)

22:09:42 <timbl> As light, I can't condone ...

22:09:46 * jhendler looks for section 4.4 of the IRC spec to see what a director is...

22:09:57 <DanC> my money has (1) the WG taking out section 4 (2) timbl filing a formal objection (3) the WG requesting PR anyway, (4) The Director agreeing to go to Proposed Rec over timbl's objection, in the interest of getting RDF to REC before the sun burns out.

22:10:07 <bijan> Hehe

22:11:10 <AaronSw> Hm. I wonder if the section 4 stuff will be removed from the mime type draft as well

22:11:11 <timbl> I note that specs have been stopped before when people have noticed there were a lot of words but nothing actually defined.

22:11:22 <bijan> So timbl, out of curiousity, what do the W3C layers say about all this?

22:11:28 <bijan> lawyers

22:11:32 <bijan> samn "w" key

22:11:34 <bijan> damn, even

22:12:02 <DanC> I got djw, the T&S domain lead and a lawyer, to review it, and he thought it was ok.

22:12:10 <bijan> Ah.

22:12:16 <DanC> circa the time the WG made the decision

22:12:16 <timbl> Why do you keep wanting to bring legal issues into a technical spec?

22:12:20 * bijan looks for *another* lawyer, one that agrees with him.

22:12:23 * jhendler lawyers nothing, what does Janet say :->

22:12:37 <bijan> Because the spec as stands is trying explicitly to bridge that gap

22:13:06 <timbl> So Bijan, would you agree to my two points as being useful ino for the aspiring RDF user?

22:13:30 <bijan> Uhm...maybe, if I remembered what they were or had decent search in my irc transcript...

22:13:30 <AaronSw> Hm, I suspect Lawrence Lessig would also file an objection if section 4 was removed.

22:13:44 <AaronSw> He's been talking a lot about making sure the Creative Commons RDF is enforcable lately.

22:13:56 <timbl> (a) That the meaning of an RDF document is the conjunction of the meaning of the statements. (b) That the menaing of the statemenmt is defined by the definition ofthe predicate term.

22:14:03 <DanC> I don't think legal opinions are most relevant. I think community consensus is. And as much as I agree with timbl on this, I wonder if the W3C Director should really be saying there's consensus around section4.

22:14:03 <kendallc> (that's interesting, aaron)

22:14:04 * bijan would be happier if someon elike lessig in the loop

22:14:24 <bijan> a, maybe

22:14:29 <bijan> b no

22:14:36 <timbl> hat of, i clearly there is no consesnsu around section 4.

22:14:37 <DanC> do you know if lessig has actually looked at the spec, AaronSw? that is interesting to heat.

22:14:39 <bijan> B is just false

22:14:41 <DanC> hear

22:14:45 <dajobe> (but their - CC's - RDF is in HTML comments, a whole different thing/story)

22:14:50 <AaronSw> kendallc et al, don't quote me on that, though. I don't want to put words in his mouth.

22:14:57 <timbl> I think a lot of setion 4 is saying what the spec is independnet of, and tha tcould be omitted.

22:14:57 <bijan> um...

22:14:58 <bijan> Too late?

22:15:01 <kendallc> dajobe: exactly

22:15:18 <timbl> If you don't believe in (b) then we have a problem.

22:15:33 <kendallc> i have trouble seeing how to make it enforceable insofar as i'm still unclear how to associate cc rdf in a comment with any particular resource

22:15:41 <bijan> "meanign of the statemetn is defined by the definiton of the predicate term"?

22:15:51 <bijan> I mean...first off, what about subject and object?

22:15:55 <DanC> (b) is pretty clear in the model theory...

22:16:15 <bijan> second, which definition?

22:16:45 <timbl> Because we have a nunch of specs for vocabularies which "define Properties" by defining the meaning of statements involving them in the P position.

22:16:45 <bijan> third, which meaning? Is the legal significant of an rdf statement (or social import) defined by the definition of the predicate?

22:17:03 <DanC> "if E is a triple s p o . then I(E) = true if <I(s),I(o)> is in IEXT(I(p)), otherwise I(E)= false." -- http://www.w3.org/TR/2002/WD-rdf-mt-20021112/#gddenot

22:17:12 <sandro> Shouldn't B be more like: the truth of an RDF statement/graph on the interpretations of the S,P,O URIRefs, and their interpretations can be constrained by URI owners.

22:17:29 <sandro> Shouldn't B be more like: the truth of an RDF statement/graph depends on the interpretations of the S,P,O URIRefs, and their interpretations can be constrained by URI owners.

22:17:50 <timbl> Bijan, it is socially clear which dfeinition for things lke OWL and dub,in core and http://www.w3.org/2000/10/swap/pim/contact. Do you wnat to dispute the general idea of an authoritiave definition, or quibble about details?

22:17:51 <DanC> I parameterize (b) by interpretation...

22:17:57 <AaronSw> that seems to lose the essential point of b, sandro

22:18:16 <timbl> sandro, absolutely no.

22:18:22 <timbl> See the my:car example above.

22:18:26 <bijan> Well, I think it *isn't* socially clear, or we wouln't find it all so contentious

22:18:28 <DanC> That the menaing of the statemenmt (i.e. its truth in I) is defined by the definition (extension in I) of the predicate term.

22:18:30 <bijan> putting that asside

22:18:36 <sandro> If the essential point of B is the "predicate" focus, than yes. That focus is not true.

22:18:45 <bijan> I see lots of accounts of the formal meaning

22:18:47 <bijan> That's clear

22:18:57 <bijan> And I know how to work with that

22:19:02 <bijan> Even in social contexts

22:19:04 <sandro> That is -- the truth depends on all three parts of the triple, not just the predicate.

22:19:32 <bijan> Plus, I don't know what happens if I add restrictions to someone "elses" property

22:19:53 <bijan> And I'm not just quibbling here...I've tried to write some of this up

22:20:03 * DanC goes back to pathcross coding... and lunch...

22:20:11 <timbl> Sandro, the truth depenmds on all three parts, BUT the way we build the sem web is by making a set of Property definitions where p is ocnstant and s and o are parameters.

22:20:34 <bijan> And I beleive your "it is social clear" is assertion only, not something we've determined empirically.

22:21:28 <timbl> You don't accept that the OWL spec defines eth owl: terms?

22:21:30 <AaronSw> the key thing in my mind is that P(S,O) can be defined as meaning the sky is blue for all S, O, but S and O cannot.

22:21:37 <bijan> Oh.

22:21:38 <bijan> hmm.

22:22:33 <bijan> No, they do.

22:22:39 <timbl> Thank you.

22:22:42 <bijan> But only formal meaning.

22:23:02 <timbl> What about the dublic core spec defining the dc: terms?

22:23:53 <bijan> More like "indications of use" than definition

22:24:01 <bijan> DC specs are superweak.

22:24:15 <bijan> To the degree that they are specs

22:24:44 <timbl> Well, the dublin core is alas rather vaguly put, but I thinkwe agree that there is one authoritative spec, however much we would prefer t to be cleaned up in some ways.

22:24:46 <bijan> And, I'd guess that their actual or even intended meaning is more in lines of operational behavior of search engines

22:24:51 <bijan> no.

22:24:58 <bijan> The dc community doens' tthink that afaict

22:25:06 <bijan> See Application Profiles

22:25:11 <jhendler> would I understand the dublin core documents without reading the spec (i.e. the non-document part)? I don't think so, or if I did it would because I guessed what "creator, date" etc.were implying

22:25:45 <jhendler> just from the RDF, I could not intuit what was expected of me (social meaning wise) from the documents -- so the power of the spec (the social meaning) is different from what is in the document

22:25:58 <timbl> I agree that DC is probably one of the worst examples. The OFX spec would be better (were it RDF).

22:26:02 * DanC updates itinerary to use new well-known timezone names... fun...

22:26:03 <jhendler> this is where it seems to me that somehow bijan and timbl are speaking past each other.

22:26:28 * mortenf nudges DanC towards http://xml.mfd-consult.dk/foaf/morten-0301-ams.rdf

22:26:33 <timbl> The OFX spec has numbers which have to add up, and testable implictaions like accruing interest if you don't clear your balance, and going to jail etc.

22:27:59 <bijan> Tim: I'm just not familar enough with that to dicuss it sensibly

22:28:13 <DanC> wow, morten, you learned how to do that from http://www.w3.org/2000/10/swap/pim/travel.html ?

22:28:21 <mortenf> yeah, I tried...

22:28:52 <bijan> God, is it 5:30? sheesh

22:29:16 <danbri> god, is it 10.30? sheesh...

22:29:26 <bijan> heh.

22:29:29 <timbl> time flies when you are enjoying yourself.

22:29:34 * bijan loses the sympathy battle.

22:29:46 * mortenf wishes to thank the participants of the discussion for some quite enlightening comments.

22:30:09 <bijan> hmm. i would say you're welcome, but I'm not sure any of my comments were enlightening :)

22:30:10 * timbl would like to thank morten for a quite ambiguous remark ;-)

22:30:25 <DanC> hmm... :ams-cph a i:Flight; dc:date "2003-02-03";

22:30:48 * timbl roots around for a flight defn

22:30:53 <bijan> timbl, can I get a bit of conversation closure, even if we continue rtalking right now?

22:31:12 <bijan> E.g., 1) do you agree that section 4 as stands needs serious work?

22:31:12 * jhendler goes for the sympathy vote by pointing out to Tim that I have to have these arguments with Bijan more often :->

22:31:37 <mortenf> DanC: I couldn't make sense of your use of date/time vs. departure/arrival

22:31:57 <bijan> 2) Do you still believe that not having something "like" section 4 will make RDF technically unimplemntalbe or undeployable

22:32:03 <timbl> Bijan, ok...

22:32:16 <danbri> mortenf, re <i:HotelReservation rdf:ID="hotel">

22:32:17 <danbri> <i:hotelName>The Quentin England</i:hotelName>

22:32:19 <timbl> Bijan, I agree that section 4 is vague, has no consesnus, doesn't do the job.

22:32:31 <timbl> I diagree that deleting it leaves tye rDF sepoc with everything it needs.

22:32:33 <bijan> 3) If you think it's technically implemenatble/deployable, do you think it will be pragmatically ineffective, or just unused?

22:32:33 <DanC> morten, ok, gotta work on the docs. in short...

22:32:36 <danbri> ...would be nice to annotate the hotel entry to say that it had wide open 802.11 access :)

22:32:52 <mortenf> danbri: :) (how?)

22:32:58 <DanC> hmm... :ams-cph a i:Flight; k:startingDate [ xsdt:date "2003-02-03"].

22:33:16 <bijan> 4) Given that there's lots "missing" from the current RDF spec (in terms of the semweb layer cake) do you think deferring this general issue to another spec or wg is a workable idea?

22:33:37 <danbri> (dunno:)

22:33:38 <timbl> Loosely, it is the job of teh RDF spec to tell you (indirectly) what an RDF document means.

22:33:39 <DanC> i.e. the startingDate of the flight is some day... not a string, but related to a string by the xsdt:date property

22:34:04 <timbl> (Why not make the startingDate a string?)

22:34:08 <mortenf> yeah, that could work, but I'd actually rather have a complete timestamp (as well)...

22:34:09 * DanC wonders how much N3 mortenf groks

22:34:23 * mortenf would estimate it as "some" N3...

22:34:48 <timbl> bijan, that is nore or less the only job of the rdf spec.

22:35:05 <timbl> The owl spec should tell you what owl statements mean.

22:35:13 <DanC> I didn't make up startingDate; I borrowed it from cyc. cuz I like their design, and it fits neatly with XML Schema datatypes & interpretation properties, timbl

22:35:15 <timbl> by defining owl predicates and classes.

22:35:25 <danbri> timbl, now that's leading to trouble.

22:35:25 * timbl ok, danc

22:35:30 <danbri> what's an 'owl' statement?

22:35:31 <bijan> But it does tell you, just not to the degree or in the way that you want

22:35:55 <danbri> Is _:p1 <foaf:mbox> <mailto:danbri@w3.org> an OWL statement?

22:35:55 <timbl> Maybe it does indirectly.

22:36:05 <timbl> not as I used it above.

22:36:08 <bijan> Which is all you said was needed :)

22:36:15 <danbri> ...or an RDF statement? an RDFS statement? A DAML+OIL statement?

22:36:17 <danbri> why not?

22:36:22 <bijan> But, it does tell you at least part of the meaning directly

22:36:23 <DanC> hm... :hotel contact:home :quentin .

22:36:38 <timbl> I meant, statements whose predicate is in teh owl namespace, or whose predicate is rdf:type and whose object is in the owl namespace.

22:37:12 <dajobe> OWL triples are defined right here: http://www.w3.org/TR/2003/WD-owl-ref-20030221/#Syntax

22:37:23 * mortenf found the cyc address terms to be somewhat incomplete...

22:37:32 <danbri> timbl,

22:37:33 <danbri> [[

22:37:34 <danbri> <rdf:Property rdf:about="http://xmlns.com/foaf/0.1/mbox"

22:37:34 <danbri> rdfs:label="personal mailbox"

22:37:34 <danbri> rdfs:comment="A personal mailbox, ie. an Internet mailbox associated with exactly one owner, the first owner of this mailbox. This is a 'static inverse functional property', in that there is (across time and change) at most one individual that ever has any particular value for foaf:mbox.">

22:37:34 <danbri> <rdf:type rdf:resource="http://www.daml.org/2001/03/daml+oil#UnambiguousProperty"/>

22:37:36 <danbri> <rdf:type rdf:resource="http://www.w3.org/2002/07/owl#InverseFunctionalProperty"/>

22:37:36 <jhendler> dajobe - let's not go there right now...

22:37:38 * DanC thinks foaf:mbox_sha1sum is one of the coolest ideas around... wants to exploit it somehow.

22:37:38 <danbri> <rdfs:isDefinedBy rdf:resource="http://xmlns.com/foaf/0.1/"/>

22:37:40 <danbri> </rdf:Property>

22:37:41 <timbl> I owuld like to se it explicitly pass the authority for defining the menaing of a statemnt to the authority for the predicate Property.

22:37:42 <danbri> ]] -- http://xmlns.com/foaf/0.1/

22:37:45 <DanC> ouch!

22:38:06 <DanC> ^that's why N3 exists ;-)

22:38:07 <danbri> So _:p1 <foaf:mbox> <mailto:danbri@w3.org> is an OWL statement.

22:38:14 <bijan> "Like" as in "think nothing will work otherwise" or "think it would work worse if not"...?

22:38:39 <timbl> 0.001 bounty for someone to paste that in in n3

22:39:01 <timbl> danbri, no

22:39:03 * jhendler danbri - or,more technically, a DAML+OIL statement until you run it through the converter

22:39:12 <danbri> <rdf:type rdf:resource="http://www.daml.org/2001/03/daml+oil#UnambiguousProperty"/>

22:39:13 <danbri> <danbri> <rdf:type rdf:resource="http://www.w3.org/2002/07/owl#InverseFunctionalProperty"/>

22:39:21 <danbri> I put in both assertions

22:39:32 <danbri> hedgin' those bets ;)

22:39:44 <DanC> hmm... tried $ python ~/w3ccvs/WWW/2000/10/swap/cwm.py --rdf http://xmlns.com/foaf/0.1/ --n3 | less

22:39:45 <DanC> lost.

22:39:46 <mortenf> DanC: oh, and HotelReservation != Hotel (or k:HumanResidence), right?

22:40:03 <jhendler> apologies Danbri, you did, I missed the second.

22:40:29 <timbl> danbri, [17:35] <timbl> I meant, statements whose predicate is in teh owl namespace, or whose predicate is rdf:type and whose object is in the owl namespace.

22:40:31 <danbri> I updated at the weekend btw., FOAF is now an OWL:Ontology (whatever that means)

22:41:02 * timbl agrees mbox_sha1sum teracool.

22:41:05 <DanC> I think you're right, mortenf, but I haven't studied k:HumanResidence closely

22:42:07 <DanC> hmm... danbri, I'm getting HTML when I wget http://xmlns.com/foaf/0.1/ . Do you offer RDF?

22:42:32 <danbri> It should content-negotiate to get the pure rdf. Also the rdf is embedded, and also available as index.rdf

22:42:34 * burtonator is away: coffee shop time

22:42:40 <danbri> (more bet hedging)

22:42:44 * jhendler thanks danbri

22:42:54 <DanC> index.rdf should work for now...

22:43:29 <DanC> File "/home/connolly/w3ccvs/WWW/2000/10/swap/uripath.py", line 217, in refTo

22:43:29 <DanC> if n == 0 and uri[i] == '#':

22:43:29 <DanC> IndexError: string index out of range

22:43:33 <DanC> hm...

22:43:54 <danbri> we don't use # in the namespace uri, is that a problem for cwm?

22:44:11 <timbl> I got a wird error chasing schemas ... I think on rdfs:

22:44:14 <DanC> seems to be ;)

22:44:23 <danbri> (I remember its a problem for timbl, cos he believes things named http:// can't be classes and properties...)

22:44:26 <timbl> namespace dcittionary lookup in sax parser I think

22:44:37 <danbri> I think I used the owl:includes doobry as well

22:44:38 <DanC> I think this is a cwm bug.

22:45:25 <DanC> wierd... works ok on the wgotten local file

22:45:51 <reagle> can I resume question about wrapping of xml literal? just wanna know whether i should send a last call comment

22:46:01 <DanC> foaf:mbox a daml:UnambiguousProperty,

22:46:01 <DanC> rdf:Property,

22:46:01 <DanC> owl:InverseFunctionalProperty;

22:46:01 <DanC> :comment "A personal mailbox, ie. an Internet mailbox associated with exactly one owner, the first owner of this mailbox. This is a 'static inverse functional property', in that there is (across time and change) at most one individual that ever has any particular value for foaf:mbox.";

22:46:02 <DanC> :isDefinedBy foaf:;

22:46:04 <reagle> hrm.. ericP no longer here

22:46:04 <DanC> :label "personal mailbox" .

22:46:46 * mortenf thinks timbl should fork over the .001 bounty now... :)

22:46:47 * DanC presents bill to timbl for 0.001 bounty

22:47:13 * timbl notes DanC gets 1 millirallod

22:47:28 <dajobe> reagle: well ericP is still logged in here; however ask away, jcarroll co-edits the rdf concepts

22:47:38 <DanC> hmm... rallod... do you get googlemark on that? i.e. did you make it up?

22:47:40 * jhendler notes that this is once again "OWL Full" -- another win for the good guys (not to be quoted on WOWG mailing list)

22:47:46 * timbl Danc, I only have rallods - do you have change?

22:48:13 <DanC> nope... "the currency in this country is Rallod" -- 'Problem Set 4' http://www.google.com/search?q=rallod

22:48:21 <bijan> But preserved in rdfig archives and pfps's scrollback :)

22:48:27 <AaronSw> Utopia, apparently

22:48:54 * eikeon takes note of reagle patience... that was a long wait.

22:48:55 * dajobe feels this is much more useful than trying to get everyone together in a room in Boston

22:49:04 <reagle> [1] seems to indicate that a value of type XMLLiteral will always be wrapped in <rdf-wrapper>, but this is not the case in example 9 [2].

22:49:04 <reagle> [1] http://www.w3.org/TR/2003/WD-rdf-concepts-20030123/#dfn-rdf-XMLLiteral

22:49:04 <reagle> [2] http://www.w3.org/TR/2002/WD-rdf-syntax-grammar-20021108/#section-Syntax-XML-literals

22:49:08 <bijan> dajobe: er...

22:49:25 <bijan> But I can't deploy the awesome intellectual might of my whirligig!

22:49:26 <danbri> yup, seems odd as I'm using on the most modest features of OWL... am also thinking to augment the vocab description with N3 or Lbase definitions / annotations...

22:49:47 <DanC> $traveller = genSym("traveller$given");

22:50:03 <dajobe> reagle: nope, [1] doesn't say use rdf-wrapper

22:50:07 <jhendler> Rallods are the currency of Utopia - used in financial documents to denote a fictitious currency in examples so as not to be confused with a real statement

22:50:12 <DanC> timbl, how should grokTravItin.pl come with a URI-name for the traveller?

22:50:17 <DanC> command-line arg?

22:50:24 <DanC> or just #traveller?

22:50:35 <dajobe> reagle: it tells you how to make something you can give to an XC14N algorithm which requires an XML document, and hence a doc element.

22:50:47 <bijan> So, what's the deal with the tech plenary?

22:50:57 <DanC> which deal is that?

22:51:04 <bijan> Do I still need to come? er...rephrase, is that still a good idea? :)

22:51:09 * bijan very tired...

22:51:30 <dajobe> reagle: and it defines "XML document corresponding to a string str" inside the wrapper, which is what you need to know in order to put it into the example [2]

22:51:51 <timbl> name fo traveller? $BASE.$USER ?

22:51:56 <DanC> I still expect the discussion to be interesting. The need for you to come seems considerably less; i.e. I think your viewpoint is somewhat understood. I'd rather you came than didn't.

22:52:04 <danbri> hmm does N3 or swap give a name to the relationship between a resource and its URI name(s)?

22:52:04 <bijan> Travel arrangements mean that I'd have to leave whatever between 5 and 6

22:52:15 <timbl> I think I put base() into uripath.py

22:52:17 <DanC> log:uri , danbri

22:52:24 <bijan> What time is the official session?

22:52:43 <reagle> dajobe: i have a heard time grokking that from the text.. i'll stare at it hard some more...

22:52:47 <bijan> I could phone participate or something, i suppose...

22:52:51 * DanC chuckles at official, but answers re 'scheduled'...

22:53:11 * bijan notes that his lc comment about striking RDFS got considerably less attention :)

22:53:25 <DanC> " 16:00 - 18:00 Second afternoon session

22:53:25 <DanC> formal meaning meets social context" -- http://www.w3.org/2001/sw/meetings/tech-200303/

22:53:40 <dajobe> reagle: the C14N don't tell you how to canonicalize element content, only XML docs. We need the former so have to "fake" an XML doc in order to define the XML C14N of the element content

22:53:45 <bijan> So...4-6, ugh.

22:53:45 <DanC> I'm not inclined to offer phone participation.

22:53:50 <danbri> cool. so f:mbox_sha1sum(X,Y) can be defined as f:mbox(X,M), log:uri(M,S), x:sha1sum(S,Y) ?

22:53:58 <DanC> yes, danbri

22:54:01 <danbri> phone would probably be painful...

22:54:04 <bijan> So can't be earlier?

22:54:17 * bijan just had a debate on a busy irc channel...how much more painful can it get?

22:54:23 <timbl> sha1sum is in crypro

22:54:39 <danbri> excellent. I'll add sha1sum then. Will be useful for file sharing too... [reads timbl's comment] oh ok, no need. excellent.

22:54:40 <sandro> If I were remote, I'd want to listen on the phone or web, but wouldn't care about being unable to contribute via voice.

22:54:45 <jhendler> do any of you Bostonians know a way to get a plane out of Logan after 8PM? We can't find anything later

22:55:04 * bijan notes that his lc comment about striking RDFS got considerably less attention :)

22:55:07 * DanC has been noodling on log:composition, as in { (:parent :brother) log:composition :uncle }

22:55:12 <timbl> ?me.foaf:mailbox.log:uri.crypto:sha1 OWTE

22:55:17 <reagle> dajobe, I appreciate what you are trying to do, and I'm familiar with the technique (used in XENC quite a lot), I just have a hard time coming to that understanding from that text...

22:55:21 <danbri> bijan, I'm behind on my LC comments... will get to it!

22:55:25 <bijan> Heheh

22:55:26 <danbri> s/comments/responses/

22:55:33 <dajobe> reagle: ok, take it up with the editors then - they wrote it :)

22:55:35 <bijan> DanC there's some good DL stuff on role composition

22:55:56 <bijan> General role constructors, actually

22:56:05 <danbri> is uncle-like stuff within the reach of OWL?

22:56:18 <DanC> no

22:56:29 <bijan> I *think* not, but I'm not sure given role heirarchy

22:56:42 * timbl has head swimming from trying to untangle threads

22:56:55 <jhendler> DanC, I'll try to dig up some of my old examples from when I taught Prolog, we had some interesting examples of relations that were harder than uncle

22:56:56 <DanC> I'm quite sure uncle functionality was asked for and declined in a WebOnt WG meeting.

22:57:17 <danbri> but one can nevertheless deploy an 'uncle' property using RDF/RDFS/OWL, and anticipate making its meaning more machine visible using future SW language extensions...

22:57:19 <jhendler> the Indians had one for male relative on the fathers side anywhere up the family tree - must more interesting.

22:57:20 <bijan> Can one inherite from multiple superproperties?

22:57:26 <danbri> yes

22:57:32 <jhendler> OWL cannot do uncle, but can do mother-in-law

22:57:58 <jhendler> (that is Owl DL -- in Full one could sort of kludge something using some class as instance tricks I think)

22:58:00 <DanC> early on... in requirements discussion. might be documented as an objective. (i.e. the "yeah right; right after h** freezes over" pile)

22:58:11 <bijan> so, conjunction vs chainging

22:58:16 <bijan> er..

22:58:18 <bijan> chaining

22:58:32 <jhendler> mother-in-law is functional functional (one wife w/one mother), uncle isn't

22:58:42 <bijan> I think it's a decidable extention to DL, peter would know :)

22:58:51 * timbl rats, crypto schema not complete.

22:58:56 <danbri> for some restricted definition of wife.

22:59:06 <jhendler> (multiple brothers) - Ian said it was not a decidable extension to SL

22:59:10 <jhendler> s/SL/DL

22:59:23 <jhendler> danbri - that's social meaning at it's best (or worst) :->

22:59:51 <timbl> cr:SHA a rdf:Property; s:label "SHA hash";

23:00:20 * danbri doesn't think OWL could define 'Divorce' interestingly (though maybe N3 could?)

23:00:54 <jhendler> uncle should be easy in cwm if it is just meant to be direct uncle - harder if you include "great ...." uncle

23:01:18 * jhendler OWL may be responsible for my divorce if i don't stop spending so much time on it :->

23:02:05 <jhendler> divorce requires a time - Owl can encode the primitives for time (see thetemporal daml stuff) and then an instance data could define the events of the particular marriage starts and ends

23:02:41 <danbri> anyways I'll add a bunch of this family tree stuff to FOAF so we might get some real world data to argue around...

23:02:52 <reagle> When the value of an ntriple is a url such as <http://example.com/foo> is its rdf:type == rdf:resource ?

23:02:53 <jhendler> so it depends on the meanign of "define divorce' (i.e. owl can handle uncle just fine -- :uncle a owl:class. But it's the brother of the father stuff that is harder

23:03:19 <reagle> (I mean what is the type of that of that URI)

23:03:25 <mortenf> danbri, pixel's schema might be a start.

23:03:38 * jhendler gonna go work on preventing divorce - waves good bye to bijan and timbl - been fun to observe

23:03:39 <DanC> value... you mean object, reagle? everything has type rdfs:Resourve, reagle

23:03:44 <danbri> yes reagle; from RDF's pov anything is Resource

23:04:10 <danbri> rdf:resource lower case, is just RDF syntax. rdfs:Resource, by confusing contrast, is a class of thing.

23:04:25 <danbri> yup, will definitely ack and plunder Pixel's schema :)

23:04:54 <DanC> danbri, what's the norm in foafy-land w.r.t. URIs for people. Do any tools expect it? or do they all know to key off mbox or homepage or whatever?

23:05:34 <reagle> DanC, yes: if subject/predicate/object == resource/property/value

23:05:56 <danbri> I generally discourage folk from expending too much effort worrying about URIs for people. FOAF apps generally use identifying properties instead (mbox, sha1sum_mbox - barring astronomically unlikely clashes, homepage, dna_checksum...)

23:06:23 <danbri> ...but a few people 'give themselves' URIs in their FOAF files, and that's fine by me. I just ignore them...

23:06:26 <DanC> is there dramatically more support for mbox than the others?

23:07:01 <mortenf> methinks mbox_sha1sum is actually prefered

23:07:03 <danbri> yes, though sha1sum stuff is catching up. I was a fool not to promote foaf:homepage as uniquely identifying until recently... so turns out nobody used it that way.

23:07:19 <DanC> bummer.

23:07:25 <danbri> yeah a lot of people seem really wary of exposing their mailboxes

23:07:40 <mortenf> that too.

23:08:08 <mortenf> rather, exposing other peoples?

23:08:11 <danbri> I need to come up with a spec for foaf:dna_checksum, maybe get a darpa grant...?

23:09:26 <mortenf> perhaps foaf:iris_sha1sum is even possible?

23:10:02 * danbri wonders what other uniquely identifying properties are around, that can be shared (eg. countrycode+nat insurance number is no good, as in america at least, people try to keep that private)

23:10:25 <mortenf> nat?

23:10:35 <danbri> national insurance number

23:11:02 <mortenf> nat=nin?!?

23:11:23 <danbri> nat=nat(ional)

23:11:34 * danbri fails yet again at saving keystrokes through abbreviation

23:11:41 <mortenf> :)

23:18:02 <DanC> btw... mortenf... cyc:containsInformationAbout-Focally is pretty much the same as foaf:topic

23:18:48 <mortenf> yeah, I gathered as much, but wanted to follow your lead as much as I could stomach :)

23:19:03 <DanC> :)

23:19:05 <danbri> sando at some point suggested having something like 'primaryTopic', to pick out a single entity that a document is about. Wonder if Cyc has that too?

23:19:26 <DanC> ah.. then foaf:topic is like cyc:containsInformationAbout

23:19:35 <DanC> cyc:containsInformationAbout-Focally is functional.

23:19:37 <danbri> ah, focally being f

23:19:38 <danbri> yes

23:20:20 * DanC still puzzling over cheapest way to connect me to my itinerary...

23:20:25 <danbri> It's pretty tempting, i just had concerns about deployability. I guess its an assertion that woudl be much more plausible coming from the author of a document, since nobody else could reasonably be expected to guess the exact topic.

23:21:25 <mortenf> DanC, I (or you) tried i:passenger.

23:22:09 <DanC> I'm happy using cyc:passengers as the property... it's the object of that property that I'm struggling over

23:22:27 <timbl> ?

23:22:39 <DanC> what name do I use for me?

23:22:44 <timbl> People/Connolly/myswebphomepage#me?

23:22:53 <DanC> in such a way that foafy tools will grok

23:23:17 <mortenf> do you need a name, isn't an anon resource (with foaf:x) enough?

23:23:19 <timbl> foafy tools do what?

23:23:23 <DanC> ok, so pass People/Connolly/myswebphomepage#me on the grokTravItin.pl command line?

23:23:48 <timbl> Yes, or expect an environment variable maybe.

23:24:15 <mortenf> foafy tools key (smush) on unique properties like mbox_sha1sum.

23:24:19 * DanC has nightmares about environment variables; no thanks

23:24:33 <timbl> Ummm. I decided to have ~/.peronal.n3 chmod 600

23:24:40 <timbl> sorry personal

23:24:49 <DanC> foafy tools follow seeAlso too, no?

23:24:53 <timbl> You can store all sorts of stuff about yorself in there.

23:25:05 <timbl> Like your favorite ID

23:25:05 <mortenf> yes (should?)

23:25:21 <mdupont> hey guys

23:25:34 <timbl> Your personal data can point to your public data but not ideally the opposite way around.

23:25:34 <danbri> hi

23:25:47 <mdupont> i was wondering if you could tell me what tools you use to do all this rdf stuff in

23:25:50 <DanC> ok, I'm going with People/Connolly/myswebphomepage and 'me' as args...

23:26:09 <mdupont> is there a "package" of tools that you can recommend

23:26:10 <DanC> emacs/make/python

23:26:19 <mdupont> ok got all that

23:26:22 <mdupont> and cwm

23:26:38 <mdupont> so all your owl work is cwm/python based?

23:26:43 <danbri> mdupont: whatever rdf parser I have to hand (often redland's), usually an SQL store, and some rdf query to sql tools (mine or libbys); or sometimes Cwm

23:26:49 <DanC> if you have cwm, you're bleeding-edge. I don't have any packages to nominate

23:26:54 * timbl updates http://www.w3.org/2000/10/swap/crypto.n3 to include the actual builtins

23:27:15 <danbri> i've heard good things about eikeon's package too

23:27:19 <mdupont> danbri: what sql tools you use?

23:27:25 <mdupont> i think i need to do that next,

23:27:31 <DanC> eikeon's rdflib is good stuff, yes

23:27:34 <mdupont> the n3 files i have are too big

23:27:41 <mdupont> 6mb

23:27:43 <mdupont> and up

23:27:50 <timbl> ouch

23:27:54 <DanC> are they regular?

23:28:01 <DanC> cwm talks to mysql these days

23:28:14 <danbri> mdupont, see http://www.w3.org/2001/12/rubyrdf/dataquery.html for crap docs on my sql stuff.

23:28:19 <mdupont> ok

23:28:20 <mdupont> thanks

23:28:24 <danbri> it was based on libby's rewrite of mattb's perl code

23:28:30 <danbri> .google matt triplequerying

23:28:31 <datum> matt triplequerying: http://lists.w3.org/Archives/Public/www-rdf-interest/2002Aug/0129.html

23:28:32 <timbl> Did I understand right that ericp is paying with a mixed mode store, sql+remaining triples?

23:28:41 <danbri> I believe so

23:28:41 <DanC> yup

23:29:00 <timbl> Cool.

23:29:00 <reagle> if I have an instance from an anonymous Class in the namespace with the prefix 'foo' is a decent representation for nt 'foo:_blah45555' ?

23:29:15 <DanC> ericp hasn't started porting it from perl to python yet, though

23:29:20 <reagle> (actually, nt would expand foo to the namespace...)

23:29:22 <mdupont> danbri++

23:29:26 <mdupont> looks cool

23:29:29 <reagle> so

23:29:48 <timbl> foo:_blah is not anonymous. _:fooblah is

23:30:26 <reagle> timbl, that's what i suspect, but then i figured since _:fooblah can be of a class with a known namespace... (and confused myself :) )

23:41:12 <timbl123> Wow .. for some reason cwm won't parse http://www.w3.org/2000/01/rdf-schema because it uses the xml: namespace which is not declared expclitly. I wonder what changed.

23:42:30 <timbl123> Guess i could have a weird sax impl.

23:43:33 * DanC winning...

23:43:35 <DanC> <> <http://www.w3.org/2000/01/rdf-schema#seeAlso> <http://www.w3.org/People/Connolly/home-smart>.

23:43:35 <DanC> <#thisTrip> <http://opencyc.sourceforge.net/daml/cyc.daml#passengers> <http://www.w3.org/People/Connolly/home-smart#me>.

23:44:11 <mortenf> nice!

23:44:26 <DanC> timbl, does seeAlso fit into this cwm mode stuff somehow?

23:45:02 <timbl123> I was just about to ask who did anything wiht seeAlso!

23:45:15 <DanC> the foaf crawlers use it

23:45:21 <timbl123> So no, but I suppose it *could* ... what would it do and when?

23:46:02 <timbl123> Do the foaf crawlers have any termination/cut algorithm? No ... they are after all crawlers.

23:46:31 <timbl123> What would one do about the whole world wide web of seeAlso's?

23:46:54 <DanC> hm... cwm foo1.rdf foo2.rdf --seeAlso --think

23:47:04 <timbl123> I think I'd prefr to just add a mode where it will look up any subject or object.

23:47:22 <timbl123> wher --seeAlso means?

23:47:27 <DanC> the --seeAlso arg could cause it to load all the objects of rdfs:seeAlso

23:47:34 <timbl123> recursively

23:47:39 <timbl123> --seeAlso=6

23:48:27 <DanC> or { [] s:seeAlso ?DOC } => { ?DOC a log:Truth }. maybe

23:48:49 <timbl123> Or maybe a combination. When trying to resolve a query involving s p o, and when the racine of s or o is the object of a seeAlso statement, then dereference it.

23:49:02 <DanC> or { [] s:seeAlso [ log:semantics ?F ] } => { ?F a log:Truth }. rather

23:49:38 <timbl123> Sounds like daml:Import

23:49:39 <DanC> that sounds interesting. (... racine ...)

23:50:06 <DanC> yes, I argued that we didn't need daml:imports cuz we have, among other things, rdfs:seeAlso.

23:50:15 <DanC> but the WG wanted something with more teeth

23:50:38 <timbl123> seeAlsoOrElse

23:50:53 <DanC> ;)

23:50:59 <timbl123> (BTW, can you read rdfs: with cwm?)

23:51:05 <DanC> more like seeAlsoAndBelieve

23:51:14 <timbl123> I am still camping on a cygwin box

23:51:46 <DanC> yes, i win when I do $ python ~/w3ccvs/WWW/2000/10/swap/cwm.py --rdf http://www.w3.org/2000/01/rdf-schema --none

23:53:55 <timbl123> what's the version on your /usr/lib/python2.2/xml/sax/expatreader.py ?

23:53:58 <DanC> btw... timbl... on # vs /... the best names for airports don't have #s in them: <http://www.daml.org/cgi-bin/airport?BOS>

23:54:01 <timbl123> I have 0.20

23:54:18 <timbl123> airports ... i know :(

23:54:30 <DanC> version = "0.20"

23:54:43 <timbl123> ?! bizarre.

23:55:51 <timbl123> I get File "/usr/lib/python2.2/xml/sax/expatreader.py", line 246, in start_element_ns

23:55:51 <timbl123> prefix = self._ns_stack[-1][apair[0]][-1]

23:55:51 <timbl123> KeyError: http://www.w3.org/XML/1998/namespace

23:56:08 <timbl123> Maybe I have some incompatibility with the C bits.

23:56:22 <DanC> ok, foafy folks, check out http://www.w3.org/People/Connolly/home-smart and http://www.w3.org/2003/03dc-bos/bos-me.rdf

23:56:33 <timbl123> Rats. I wanted to do the fourth query test case using auto schema lookup

23:56:44 <danbri> <timbl123> Do the foaf crawlers have any termination/cut algorithm? No ... they are after all crawlers.

23:56:52 <danbri> at the moment, we're grateful for all the RDF we can find ;-)

23:57:04 <DanC> just predeclare xml: to http://www.w3.org/XML/1998/namespace as a work-around, timbl

23:57:09 <timbl123> cwm has to watch its intake.

23:57:15 <danbri> but there are a few rdfs:seeAlsos that point to huge RDF docs, or IRC logs or whatever, so folk are starting to use more careful traversal rules

23:57:28 <timbl123> Its in the sax code, not in mine

23:57:29 <danbri> eg using typing of the thing pointed to, or from...

23:57:42 <DanC> oh. ouch, tim

23:57:50 * danbri wonders how hard it would be to write a crawler in Cwm/N3

23:58:07 <timbl123> too easy.

23:58:25 <DanC> this would pretty much do it, danbri: { [] s:seeAlso [ log:semantics ?F ] } => { ?F a log:Truth }

23:58:46 <timbl123> Danc, cwm doesn't import facts from Truths.

23:58:52 <timbl123> kinda silly, I know...

23:58:56 <danbri> but I guess Cwm wouldn't stop and do anything until after it finished traversing the entire linked web...?

23:59:18 <DanC> { [] s:seeAlso [ log:semantics [ log:includes { ?S ?P ?O } ]] } => { ?S ?P ?O }. # actually, since log:Truth only works for rules

23:59:27 <danbri> My crawlerette calls a bunch of registered handlers after each document is parsed, so i can update HTML, databases, logs etc...

23:59:41 <DanC> let's try it...

The IRC chat here was automatically logged without editing and contains content written by the chat participants identified by their IRC nick. No other identity is recorded.

Alternate versions: RDF Resource Description Framework Metadata and Text

Provided by Dave Beckett. Hosted by Useful Information Company.