Semantic Web Interest Group IRC Chat Logs for 2003-10-05

02:14:22 <Davey> Davey is now known as D[a]vey

02:26:16 <D[a]vey> D[a]vey is now known as Davey

04:09:25 <Davey> Davey is now known as D[a]vey

06:39:33 <mdupont> hey mea_culpa

06:39:39 <mea_culpa> hey mdupont

06:39:47 <mdupont> hows things going?

06:39:51 <mea_culpa> ok

06:40:12 <mea_culpa> I have this essay I have to write by monday and I haven't even started yet...

06:40:18 <mea_culpa> well I guess I kinda did

06:40:29 <mea_culpa> I have a ruff outline?

06:40:34 <mea_culpa> something like that

06:40:37 <mdupont> what essay?

06:40:39 <mea_culpa> how are you

06:40:44 <mea_culpa> for english

06:40:54 <mea_culpa> 101

06:41:00 <mdupont> good, i am finally getting something exciting done!

06:41:05 <mea_culpa> :)

06:41:12 <mdupont> converting rdf data into vectors

06:41:19 <mdupont> girl, it is *fast*!

06:41:36 <mea_culpa> :)

06:41:59 <mdupont> i finally appreciate my linear algebra

06:42:10 <mdupont> and the matrix stuff

06:42:56 <mea_culpa> Oh I got the most awsome calculator in the world! (speaking of math)

06:43:21 <mea_culpa> I look at is as an investment for my college career!

06:43:24 <mea_culpa> hehehhe

06:43:52 <mea_culpa> I got a TI-83

06:44:24 <mdupont> yea? i was looking into those, got a casio graphic fx

06:44:39 <mea_culpa> I like it

06:44:46 <mdupont> cfx9850GB plus

06:45:14 <mea_culpa> I'm still learning about... I haven't had much time yet... but it's pretty damn cool

06:45:33 <mea_culpa> it's got all kinds of cool features

06:46:07 <mea_culpa> It have the periodic chart on it! I thought that was way cool

06:46:13 <mea_culpa> have=has

06:46:16 <mdupont> thats good stuff

06:46:26 <mdupont> thos things are so expensive

06:46:35 <mea_culpa> it was 129 bux

06:46:40 <mdupont> not bad

06:47:39 <mea_culpa> no and I bought it at the place my mom works and got a 10% discount :)

12:53:10 <botlars> botlars is now known as larsbot

14:08:07 <libby> hello

14:08:19 * libby wonders if anyone minds us logging a meeting here?

14:08:39 <libby> we are talking about SW stuff in Bristol, doing demos

14:10:29 <libby>http://rdfweb.org/topic/PubsDiscussionLaptops

14:10:29 <dc_rdfig> A: http://rdfweb.org/topic/PubsDiscussionLaptops from libby

14:10:49 <libby> A:|An informal semantic web meeting in Bristol, UK

14:10:49 <dc_rdfig> Titled item A.

14:14:03 <libby>http://techquila.com/tmsinia.html

14:14:03 <dc_rdfig> B: http://techquila.com/tmsinia.html from libby

14:14:14 <dajobe-lap> kal describes b:

14:14:25 <dajobe-lap> looking at wasy of organising things

14:14:30 <libby> B:|Kal Ahmed's demo: topicmaps patterns

14:14:30 <dc_rdfig> Titled item B.

14:14:42 <dajobe-lap> thesaurus work with RIchard Light (sp?)

14:14:55 <dajobe-lap> model for descr thesauri in XTM

14:15:08 <dajobe-lap> came up with two ways to do it

14:15:16 <dajobe-lap> ISO came up with third way, BSI fourth

14:16:24 <dajobe-lap> vague UML diagram shown (in visio)

14:16:33 <dajobe-lap> term-centered, concept-centered

14:16:42 <dajobe-lap> assoc between terms are TM assocs

14:17:22 <dajobe-lap> plus other bits scope notes, wqarrants ec.

14:17:50 <dajobe-lap> 2nd form concept bnased, BT, NT are related between concepts

14:17:55 <dajobe-lap> terms become labels for concepts

14:18:33 <libby> B:relatedly: [http://www.w3.org/2001/sw/Europe/plan/workpackages/live/esw-wp-8.html SWAd-Europe WP8, theasauri]

14:18:34 <dc_rdfig> Added comment B1.

14:18:34 <dajobe-lap> prefered terms use TM scopes; nonprefered have more local scope

14:18:43 <libby> B1:relatedly: [http://www.w3.org/2001/sw/Europe/plan/workpackages/live/esw-wp-8.html|SWAD-Europe WP8, theasauri]

14:18:43 <dc_rdfig> Replaced comment B1.

14:19:09 <dajobe-lap> lose term by term use of warrants in 2nd form

14:19:38 <dajobe-lap> 3rd form is the RAL thes work (after Desire) in OWL

14:19:54 <libby> owl to topicmap conversion intersting:

14:20:06 <libby> directed to non-directed graph

14:20:56 <libby> three distinct uris in tm not one; had to add new concepts in

14:22:38 <libby> B:seealso [http://www.w3c.rl.ac.uk/SWAD/thesaurus/tif/tif.html|A Thesaurus Interchange Format for the Semantic Web] by Alistair Miles

14:22:38 <dc_rdfig> Added comment B2.

14:23:46 <libby> ...ideally would like an automated way of converting between OWL and TM

14:24:02 <dajobe-lap> note that it wound be friend to encourage recordign of symmetric ec. properties to aid auto convesion

14:24:16 <libby> nmg: were tms expressive enough to capture all the OWL information?

14:24:48 <libby> kal:transitive, symmetric difficult

14:24:55 <libby> nmg: local range contraints?

14:25:13 <dajobe-lap> (reminder for later, swade thinking of tm workshop at xml europe)

14:25:24 <libby> kal: no standard constraint language for TM. maybe shoudl be a requirement to map from OWL and RDF

14:25:39 <dajobe-lap> kal: tmcl still being worked on

14:26:21 <libby> ...can you make the RDF/OWL to TM conversion without losing anything

14:27:37 <dajobe-lap> an xslt to convert a psvi output froms xsv into tm

14:27:42 <dajobe-lap> ^- kal

14:27:46 <dajobe-lap> running into real problems

14:27:55 <dajobe-lap> sine psvi is not easily understoond by me

14:28:04 <dajobe-lap> (we wait for jenni t to help us)

14:28:58 <dajobe-lap> had to split into local and globally edeclared, some with local scope, some global

14:29:48 <libby> ...currently processing just PSVI, hopefully WSDL and XSLT

14:30:11 <libby> ...make this chenge to teh schema and then do a query to see what will be affected by the change

14:30:43 <libby> ...PSVI doesn't give you this, but converting to TM or RDF can enable this

14:31:23 <libby> ...not sure if XSLT2 wuld give you the right information

14:33:20 <libby> --sean palmer

14:35:44 <libby> -demos his python RDF API

14:36:49 <libby> ...ntriples parser - easy to parse, useful to store data in

14:37:03 <libby> ..wanted a really good ntriples parser - as perfect as possible

14:37:20 <libby> ...trickiest paert was handling unicode. has a lot of tests for unicode

14:37:42 <libby> ...maybe incorporate into RDF tests

14:37:51 <libby> ...woudl like a standard test suite for APIs

14:38:39 <libby> B:[http://ilrt.org/discovery/chatlogs/rdfig/2003-10-05.html#T14-14-30-1|discussion on #rdfig]

14:38:40 <dc_rdfig> Added comment B3.

14:39:06 <libby> BLURB:Sean palmer shows his python RDF api

14:39:06 <dc_rdfig> C: Sean palmer shows his python RDF api from libby

14:39:18 <libby> C:[http://ilrt.org/discovery/chatlogs/rdfig/2003-10-05.html#T14-33-20|discussion on #rdfig]

14:39:18 <dc_rdfig> Added comment C1.

14:39:40 <libby> ...main class for resources and literals is Node

14:40:32 <libby> ...wanted sublcasses; also disambiguation of input from the ntriples versions - feed to it a constrcutor and give you the right subclass

14:40:48 <libby> ...really trciky in python; not possible until 2.3

14:42:06 <libby> ...main class is Graph, bit like a context. Can initialize as ntriples or with a uri - xml regex else pass to ntreipls. maybe later n3

14:42:42 <libby> dajobe-lap: how do you do graph comparison with graph hashes?

14:43:19 <libby> sean: isomophism code using internal python hashes. passes NTC tests for iphomorphism

14:43:36 <libby> ...canonoicalizing and then using a python hash on canonicalized form

14:43:56 <libby> ...really useful for graph merging where you have to rename bnodes

14:44:10 <libby> ...really annoying

14:44:33 <libby> ...you want to be able to forbid a certain prefix space.

14:44:48 <libby> ...sbp renames all the bnodes using the hash of the graph

14:45:13 <libby> nmg: do something similar in 3store - hash of the uri

14:45:47 <libby> ...currently working towards path stuff, and going further on the query stuff.

14:46:15 <libby> ...want to add multiple metadta contexts. cwm has had them for years now. useful for rules taht create rules, which happens once in a blue moon.

14:46:33 <libby> ...prob is that no contexts in RDF/XML as it stands

14:47:08 <libby> ...(esp nested contexts)

14:48:15 <libby> ...tension in the rdf world between hhacking and academia - useful stuff and provability

14:48:32 <libby> nmg: when the SW is involved in cash machines, I want it to work!

14:48:56 <libby> sean: depends on the app - if you dont care about provability....

14:49:05 <libby> nmg: yep if working in a walled garden

14:50:06 <libby> libby: contexts *so* important for foaf stuff

14:50:15 <libby> ...in an application sense

14:50:49 <libby> sean, nmg, problem with trusted sources for names

14:51:07 <libby> jim: already sorts by trusted sources

14:51:28 <libby> sean: chaals is the worsdt for that - 6 different spellings of his name

14:52:05 <libby> dajobe-lap: can;t qualify properties in RDF but can in topicmaps, e.g. preferred uri, prefeered name

15:13:31 <libby> ---jim ley

15:14:26 <libby> bluetooth presence

15:14:35 <libby>http://jibbering.com/bt-presence/example.1

15:14:36 <dc_rdfig> D: http://jibbering.com/bt-presence/example.1 from libby

15:14:49 <libby> D:|bluetooth presence SVG

15:14:50 <dc_rdfig> Titled item D.

15:16:22 <libby> D:see [http://rdfweb.org/topic/FoafMobile|foafmobile]

15:16:22 <dc_rdfig> Added comment D1.

15:17:18 <libby> - uses foaf gender for the icons

15:18:02 <libby> d2m: and [http://jibbering.com/discussion/Bluetooth-presence.1|a writeup]

15:18:11 <libby> D:and [http://jibbering.com/discussion/Bluetooth-presence.1|a writeup]

15:18:11 <dc_rdfig> Added comment D2.

15:19:05 <libby> ...uses a hashed bluetooth address as an identifier

15:20:32 <libby> (in the example we can see two blue people near together (jim and morten) and a white person (who has left), and their names)

15:21:30 <libby> ...anything that can talk bluetooth and http as a locator

15:21:47 <libby> ...meeting your blind date, the two devices start talkign to each other

15:22:10 <libby> ...even if there are lots of bt dvices, you have the ids of the one you are interested in

15:23:11 <libby> D:[discussion on #rdfig|http://ilrt.org/discovery/chatlogs/rdfig/2003-10-05.html#T15-13-31]

15:23:11 <dc_rdfig> Added comment D3.

15:24:21 <libby> nmg: ibuttons vs rfids - ibuttons are cheaper (the rfid readers are £1500-ish)

15:25:31 <libby> ---nick gibbins

15:25:39 <libby> BLURB:CS AKTive space tour

15:25:39 <dc_rdfig> E: CS AKTive space tour from libby

15:26:04 <libby> ...taken lots of public information from various adtasources in the UK research community

15:26:18 <libby> ...e.g. finding the experts in a subject in the uk and where they are based

15:27:27 <libby> ...map of uk, see CS depatrtments, hover and select them, all the different groups within 50 miles

15:28:02 <libby> ...heirarchy of subjects; who works in each one

15:28:17 <libby> ..information on each person, research interes and publications

15:28:57 <libby> ...implicit communities of practice, e.g. the conferences oyu go to, places you work, publications you ahve published in

15:29:08 <libby> ...weighted path search from a person, ujsing the ontology

15:30:14 <libby> ...loads of depiction data :)

15:30:52 <libby> ...uses smushing; these 5 nigel shadbolts are the same person.

15:31:06 <libby> ...automatic techniques are all error-prone

15:31:18 <libby> ...data is all very messy

15:31:41 <libby> ...did it automatically, and then improved by hand

15:32:02 <libby> ...thinks bnodes are actively harmful for this problem

15:33:12 <libby> kal: in TM it is similar

15:34:18 <libby> nmg: inversfunctionalproperties is the only real way. but there are very few of these

15:34:31 <libby> ...sameas is no good if people are bnodes

15:34:55 <libby> ...can;t say in a 3rd file that a in one file and b in another are the same using bnodes

15:35:16 <libby> ...shows sources of data

15:35:30 <libby> ...lots of infromationj about who gathered it, when last gathered

15:35:40 <libby> ...15-16 million triples

15:37:20 <libby> ....simple schema, s,p,o using hashes; index building takes the time

15:37:51 <libby> ...different sized circles in the uk, get more people

15:38:04 <libby> ...order by RAE result or by grant total :)

15:38:28 <libby> --nick demos 'signage

15:39:23 <libby> E:also discussed [signage|http://signage.ecs.soton.ac.uk/]

15:39:23 <dc_rdfig> Added comment E1.

15:39:40 <libby> ...demoed at several recent conferences

15:40:47 <libby> ---damian demos treehugger

15:41:26 <libby>http://swordfish.rdfweb.org/discovery/2003/09/treehugger/

15:41:26 <dc_rdfig> F: http://swordfish.rdfweb.org/discovery/2003/09/treehugger/ from libby

15:41:45 <libby> F:|TreeHugger demo - RDFPath and XSLT

15:41:45 <dc_rdfig> Titled item F.

15:42:19 <libby> ...maxf and libby discussing RDFPath languages over the summer, and norm walsh and others at extreme markup

15:42:58 <libby> ...quite simple to extend Saxon

15:43:16 <libby> ...peopel try to use XSLT on RDF but often it won;t work

15:43:43 <libby> F:see also [http://rdftwig.sourceforge.net/|norm walsh's RDFtwig]

15:43:43 <dc_rdfig> Added comment F1.

15:44:30 <libby> ...idea is that you use RDFPath language to query the RDF, combined with xslt, to ease peopel into RDF from XML

15:45:13 <libby> ...in the demo, the document root is the RDF model; the children of the document node are all the subject nodes in the graph

15:45:48 <libby> F:[#rdfig discussion|http://ilrt.org/discovery/chatlogs/rdfig/2003-10-05.html#T15-40-47]

15:45:48 <dc_rdfig> Added comment F2.

15:46:04 <libby> ...it goes node-property-node using the striping

15:46:24 <libby> ...text() is a literal

15:47:43 <libby> ...rdf:li is a sort of cheat; also rdf:about, rdf:resource are syntactic...but it works!

15:47:58 <libby> ...using the striped syntax.

15:48:19 <libby> ...because jena 2 comes with inferencing support, easy to add in some inferencing

15:49:24 <libby> ...to grab resources rather than print them you can grab them using rdf:about (as objects)

15:50:21 <libby> E:[#rdfig discussion|http://ilrt.org/discovery/chatlogs/rdfig/2003-10-05.html#T15-25-31]

15:50:21 <dc_rdfig> Added comment E2.

15:50:47 <libby> ...you can represnt bnodes using * or rdfs:resource

15:51:20 <libby> sbp: why do you distinguish subjects from objects at the root level?

15:51:40 <libby> damian: not sure. with objects you have to back up to get anything

15:52:02 <libby> sbp:doesn't think it matters if a subject or object

15:52:35 <libby> morten: because you need somewhere to start - tree?

15:53:15 <libby> damian: definitely needed tobe able to do inverse, i.e. going backwards up the tree (an anonymous inverse property)

15:53:36 <libby> ...e.g for dc:creator

15:53:43 <libby> ...probably shoudl be a function

15:54:35 <libby> ...one thing you can;t do is talk about properties - e.g. give me the labels for properties, can;t do that

15:56:16 <libby> ...ported to saxon 7, so XSLT2, Xquery (xslt2 is not strictly backwards compatible; xquery looks interesting - as a programming labuage)

15:56:44 <libby> libby: partly came from discussions at Budapest and meeting XML peopel at XML summer school with kal and jeni

15:56:52 <libby> heya verbosus :)

15:56:59 <verbosus> Hello there, Libby.

15:57:32 <verbosus> Just read about the SW meeting in Bristol: did I told you about the pub nearby my former house with open wifi?

15:58:12 <libby> A:attending - Libby Miller, Damian Steer, Jim Ley, Kal Ahmed, Morten Frederiksen, Nick Gibbin, Sean Palmer, Dave Beckett

15:58:12 <dc_rdfig> Added comment A1.

15:58:20 <libby> yeah, I remember about that :)

15:58:28 <libby> I didn;t think it was still there...

15:58:40 <verbosus> The pub or the open wifi?

15:58:44 <libby> the wifi

15:59:04 <verbosus> Because I’m sure they haven’t switched it off. Maybe it doesn’t go through the house/pub wall.

15:59:16 <libby> ah, I didnt check!

15:59:19 <libby> I will do now

15:59:21 <verbosus> But the hotspot is on the pub’s side.

15:59:22 <libby> :)

15:59:26 <verbosus> Heh ;-)

15:59:57 * verbosus is sorry not to be able to attend :_(

16:00:01 <libby> aw

16:00:26 <libby> it's been fun. we're going down the pub shortly...

16:01:10 <libby> jeni and norm are coming along.

16:01:20 <verbosus> Is Sean there?

16:01:27 <libby> yes :)

16:01:34 <verbosus> Say hi, then :-)

16:02:53 <dajobe-lap> present: shellac, libby, dajobe, jim ley, kal ahmed, mortenfn, nickg, sbp

16:11:24 <mdupont> dajobe-lap: i have something neat... a redland to c converter...

16:32:46 <arnarl> hi

16:33:25 <mdupont> hey arnarl

16:33:47 <arnarl> hi mdupont, how goes the job-hunt?

16:34:03 <mdupont> i am going to make my own company

16:34:11 <arnarl> not bad

16:34:29 <mdupont> yes, in this econonmy, it is very hard to find a long term contract

16:34:40 <mdupont> so why not go for lots of short term work

16:34:51 <mdupont> companies want low risk and low cost work

16:35:04 <arnarl> indeed

16:35:40 <mdupont> how goes your move-over?

16:35:49 <arnarl> many people make a good living from self-employment

16:36:03 <arnarl> I'm one week into a five week hiatus

16:36:10 <mdupont> cool!

16:36:51 <arnarl> mostly watching soccer so far, but next week I'll be checking out apartments closer to the new job

16:37:24 <arnarl> then I expect to spend some time preparing for a project at the new job

16:38:25 <mdupont> good

16:38:38 <mdupont> it is good to have some time off

16:38:47 <mdupont> i missed that, jumped from one to the other

16:38:50 <mdupont> that was a big mistake

16:39:00 <mdupont> i ended up partying too much at the new job

16:40:29 <arnarl> he he

16:40:58 <arnarl> yeah, one has time to read and have a life

17:00:17 <mdupont> gotta go

17:00:20 <mdupont> see you

22:44:48 <eaon|zZz> eaon|zZz is now known as eaon

22:48:43 <libby> A:[http://swordfish.rdfweb.org/photos/2003/10/05/index.html|libby's photos]

22:48:44 <dc_rdfig> Added comment A2.

23:13:55 <mof-alice> dc_rdfig, list

23:13:56 <dc_rdfig> Not understood: list

23:14:05 <mof-alice> dc_rdfig, help

23:14:06 <dc_rdfig> Post a URL by saying it on a line on its own

23:14:07 <dc_rdfig> To post an item without a URL, say BLURB:This is the title

23:14:08 <dc_rdfig> I will reply with a label, for example A

23:14:09 <dc_rdfig> You can then append comments by saying A:This is a comment

23:14:10 <dc_rdfig> To title a link, use a pipe as the first character of the comment

23:14:11 <dc_rdfig> Eg. A:|This is the title

23:14:12 <dc_rdfig> To see the last 5 links posted, say dc_rdfig:view

23:14:13 <dc_rdfig> For more features, say dc_rdfig:morehelp

23:14:35 <mof-alice> dc_rdfig, view

23:14:35 <dc_rdfig> B: Kal Ahmed's demo: topicmaps patterns (http://techquila.com/tmsinia.html)

23:14:36 <dc_rdfig> C: Sean palmer shows his python RDF api (blurb)

23:14:37 <dc_rdfig> D: bluetooth presence SVG (http://jibbering.com/bt-presence/example.1)

23:14:38 <dc_rdfig> E: CS AKTive space tour (blurb)

23:14:39 <dc_rdfig> F: TreeHugger demo - RDFPath and XSLT (http://swordfish.rdfweb.org/discovery/2003/09/treehugger/)

23:15:37 <mof-alice> A:[http://www.wasab.dk/morten/2003/10/photos/england/3/|morten's photos]

23:15:37 <dc_rdfig> Added comment A3.

23:16:13 <JibberJim> A:[http://jibbering.com/piccies/2003/10/5/|Jim's photos] (purely to joing the chumping extravaganza!)

23:16:13 <dc_rdfig> Added comment A4.

23:20:06 <mof-alice> A:seeAlso [http://www.wasab.dk/morten/2003/10/photos/england/2/|the day before] and [http://www.wasab.dk/morten/2003/10/photos/england/1/|the day before that]

23:20:06 <dc_rdfig> Added comment A5.