Semantic Web Interest Group IRC Chat Logs for 2003-04-16

00:21:19 <sandro> what RSS client do you use, Danbri?

00:22:16 <danbri> netnewswirelite at moment, just got it on mac

00:22:32 <danbri> i had another one for windows laptop whose name currently escapes me, gui thing

00:22:39 <sandro> linux?

00:22:48 <danbri> nothing currently

01:26:18 <danbri> G:OK, I made a [http://tux.w3.org/~danbri/rss/ few more feeds], and begun moving this to more stable w3c hosting...

01:26:19 <dc_rdfig> Added comment G9.

01:26:55 * danbri -> zzz

04:53:11 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

06:06:25 <Schuyler_> hiya folks

08:19:39 <aharth> anybody knows where to find the wine.owl file mentioned in the spec?

08:28:26 <libby> thanks for your msg nick

08:35:28 <JibberJim> 14:00 UTC geo meeting today?

08:38:59 <libby> sked, all geo

08:39:02 <skedulike> time now is 2003-04-16 08:38 UTC. Upcoming meetings concerning geo:

08:39:02 <skedulike> 2003-04-16 14:00UTC Geo meeting #rdfig

08:39:15 <libby> yes :)

08:39:34 <libby> you able to come along jim?

08:41:21 <JibberJim> I think so.

08:41:51 <libby> cool

09:06:43 <grault> grault is now known as earle

09:06:46 <earle> morning

09:07:47 <libby> hey earle

09:07:55 <JibberJim> hey earle

09:08:01 <libby> you about later for the geo meet?

09:08:05 <earle> absolutely

09:08:29 <earle> thought I'd make an appearance early to prevent myself from forgetting. :)

09:14:40 <nmg> ack. not going to be able to make the geo meeting this afternoon (off on hols).

09:14:58 <JibberJim> that's no excuse nmg, I'm making it, and I'm on hols too...

09:15:40 * nmg doesn't think that provide net connections on wales and west trains yet :p

09:17:01 <libby> earle:yay, nmg boo... ;)

09:17:20 <libby> where are you off to nmg?

09:17:30 <libby> wales and west?

09:17:39 * JibberJim notes danbri wrote some foaf meet notes on a BRS-LON train...

09:17:40 <nmg> ulster for the weekend, to see the inlaws

09:17:47 <nmg> bath is just the first stop

09:17:53 <libby> nice

09:18:11 * libby can be online always and everywhere now....

09:18:18 <nmg> on a different note, have you had any luck finding other ontologies that talk about groups?

09:18:29 <nmg> erm, how?

09:18:39 <libby> nothing yet, although someone suggested I have a look at ldap ones

09:18:50 <libby> p800 with an irc client :)

09:18:58 <libby> thanks for your messgae nmg

09:19:01 <JibberJim> Your p800 do sat phones for when you're in midatlantic?

09:19:10 <libby> ok, maybe not everywhere

09:19:34 <libby> though I heard BA are trialling internet access on some flights to the US

09:20:02 <libby> I thought I'd have a good look at eikeon's directory of ontoologies today - didnl;t realize so many were there

09:20:27 <JibberJim> I was thinking more of a rubber dinghy post crash...

09:20:51 <libby> heh

09:21:10 <libby> don't know how good the G2.5 coverage is mid-atlantic

09:24:21 <JibberJim> hmm, anyone want to scrape http://www.sosick.org/archive.html into RDF - so I can plot SARS on my geo map?

09:27:22 <libby> wow

09:27:27 <libby> places to avoid...

09:28:12 <earle> JibberJim: if you can work out where places like "Lee Kee Building, Ngau Tau Kok" are... :)

09:28:34 <earle> unicode scraping would be hard, though, I think

09:28:57 <JibberJim> I was thinking the flight information would be the interesting one EARLE - one of the sections is all flights.

09:30:07 <earle> oh, right - I'll scroll some more...

09:31:12 <earle> ah, hmm.

09:31:19 <earle> morning danbri_

09:32:35 <danbri_> hi

09:32:41 <JibberJim> does the cyc travel ontology have something for that thing which is being moved - ie the foaf:Person foaf:Group etc?

09:32:43 <earle> mmm, it's fugly Microsoft html as well.

09:36:08 <JibberJim> - http://jibbering.com/travels/places.svg and http://jibbering.com/travels/places.rdf now updated!

09:36:31 <JibberJim> oops 4 minutes of power remaining...

09:37:51 <libby> ooh, it works for me! (the svg file) :))

09:38:08 <libby> that's so cool jim!

09:38:17 <mea_culpa> mea_culpa is now known as aem_apluc

09:38:47 <JibberJim> Some of the places in the RDF though don't have lat/long so aren't shown :-(

09:38:53 <danbri_> dc_rdfig:view

09:38:53 <dc_rdfig> C: RDF version of the JPEG image description (http://people.freebsd.org/~nik/imageinfo.rdf)

09:38:54 <dc_rdfig> D: http://search.cpan.org/author/NIKC/AxKit-App-Gallery-0.4/

09:38:55 <dc_rdfig> E: http://people.freebsd.org/~nik/AxKit-App-Gallery.mov

09:38:56 <dc_rdfig> F: AnnounceOMatic notes in ESW Wiki (http://esw.w3.org/topic/AnnounceOMatic)

09:38:57 <dc_rdfig> G: Rough cut at an RSS feed generated from W3C mailing lists (http://tux.w3.org/~danbri/rss/www-rdf-interest.rdf)

09:39:03 <danbri_>http://www.mapbureau.com/rdfmap

09:39:04 <dc_rdfig> A: http://www.mapbureau.com/rdfmap from danbri_

09:39:20 <danbri_> A:|"A new version of an RDF spec for mapping/geography" (rdfmap)

09:39:20 <dc_rdfig> Titled item A.

09:39:22 <JibberJim> did you click on the blob for the name, and maybe a photo?

09:40:00 <libby> buccaneers at cintsa

09:40:03 <danbri_> chris++ # saying 'an' rather than 'the'

09:40:08 <libby> no photos tho

09:41:13 <libby> yep, I get the names

09:41:44 <JibberJim> no photos for that one, try LHR, or Zurich they have photos arlready added.

09:42:03 <libby> ok

09:42:20 <libby> ooh

09:42:42 <libby> yep heathrow works

09:42:55 <libby> heh, pics of airports

09:43:15 <libby> what a brilliant record to have of your journey

09:43:18 <danbri_> ooh, just got mail from Simon Cox (old Dublin Core friend) who is now involved w/ GML.

09:43:37 <JibberJim> I'm also wondering how to say "this is a photo from place X" - so you can show the views etc. any thoughts on that?

09:44:00 <libby> does depicts work?

09:44:17 <libby> the 'omoion' vocab for withinNmetres of might work for that

09:44:21 <JibberJim> no, because the photo doesn't depict the location.

09:44:30 <libby> s/omoion/onion/

09:45:03 <libby> hm, deicts a point within n metres of lat/long? maybe not

09:45:58 <JibberJim> battery dead... bye bye - I'll be back for the meeting.

09:46:04 <libby> cool re Simon Cox, chris danbri

09:46:08 <libby> bye jim, see you later

09:46:21 <libby> danbri_, can we gte these guys to come along later?

09:46:30 <danbri_> jim, this is great :)

09:46:59 <danbri_> simon can't make this meeting (is Australian btw -- timeslot may be the issue)

09:47:15 <libby> bum

09:48:28 <danbri_> he mentions that GML is rdf-like in its syntax (striped). Also that we'll eventually (for 'practical and legal' reasons) want to capture the coordinate ref system in use, rather than assume just one.

09:48:56 <libby> interesting

09:49:14 * libby hasn;t looked at GML fully

09:51:06 <aem_apluc> aem_apluc is now known as eDtivrsky

09:53:24 <earle> where's the onion vocab located?

09:56:49 <libby> it's not complete yet

09:57:03 * earle nods...

09:57:56 * earle gives libby an animated "Under Construction" .gif of a man digging a hole next to a barrier with a flashing light on it

10:00:09 <libby> hehe

10:00:32 <libby> danbri made a start but tells me that the scaling's wrong currently: http://esw.w3.org/topic/GeoOnion

10:00:40 * earle has a browse

10:04:08 <libby> there's a bit here from the last calendar meet: http://ilrt.org/discovery/chatlogs/rdfig/2003-04-09.html#T15-30-00

11:03:45 <danbri_> earle, yes, scaling of the 'onion' thing is currently wrong. There is too much space between 1km and 10km (roughly, 'walkable' vs 'unwalkable', for some stereotypical trudger)

11:03:56 * earle nods.

11:04:09 <earle> 1, 2, 5, 10?

11:04:39 <earle> or perhaps just 1, 5, 10

11:05:09 <danbri_> maybe we just need to make some up. i tried to get it on some mathematically plausible scale but i'm no mathmo...

11:45:08 * earle wishes there was some better word than 'thing' to describe, well, a thing.

11:46:19 <danbri> spatialthing

11:48:34 <earle> I know, but I mean in the English language in general.

11:50:19 * danbri_ tries to remember the french... 'truc'?

11:50:35 <earle> Ah, un truc, oui

11:51:45 <earle> c'est a petit bit obscure pour mon audience generale je reckon though :)

11:52:11 <earle> hmm, my Franglais is rusty

11:54:24 <earle> hullo mattb

11:57:14 <mattb> hi earle

12:28:06 <sbp`-> sbp`- is now known as sbp`

13:24:23 <zool> helo

13:24:29 <Bdragon> morning... are you a god?

13:24:35 <Bdragon> (sorry, watched ghost busters last night)

13:24:38 <Bdragon> :)

13:24:53 * zool harrumphs quietly

13:25:57 <earle> re zool

13:28:10 <zool> re earle...

13:30:09 <zool> i made a route-drawing sort of function for my bot yesterday... so something to show -

13:30:19 <zool> . http://space.frot.org/show_route/test

13:30:30 <zool> was it something i said?

13:30:54 <mortenf> heh.

13:30:57 <zool> you can make your own iyl by talking to mudlondon and going somewhere and saying 'en route' and following the instructions

13:31:23 * sandro reads RFC3404 and the Knoxville Framework (the resolution system must be separate from the way names are assigned)

13:32:19 <zool> are you worrying about the imminent collapse of the dns system sandro

13:32:50 <sandro> not at all. should I be?

13:33:45 <zool> perhaps

13:34:11 <sandro> I have plenty of other things to worry about, thanks. :)

13:34:29 <sandro> should you be XUL perhaps, instead?

13:39:58 <zool> no!

13:42:42 * DanC wonders why the ""s in A:

13:42:51 <DanC>http://lists.w3.org/Archives/Public/www-rdf-calendar/2003Apr/0011.html

13:42:51 <dc_rdfig> B: http://lists.w3.org/Archives/Public/www-rdf-calendar/2003Apr/0011.html from DanC

13:43:03 <DanC> B:|RDF calendar IRC meeting: 2003-04-23, 1500 UTC

13:43:03 <dc_rdfig> Titled item B.

13:43:16 * DanC wonders why 1500Z rather than 1700Z

13:43:41 <zool> JibberJim couldn't amke it later cos of cybercafe-closing problems

13:43:54 <DanC> wasn't that last time?

13:44:16 <libby> can;t you make it danc?

13:44:23 <DanC> I dunno.

13:44:36 <DanC> I'm just trying to figure out how 1500Z was decided; I don't see it in the 9Apr summary

13:44:47 * libby checks

13:46:39 <libby> hm

13:46:40 <libby> 16:31:51 <libby> so are peopel ha[py with 1400utc for next wed as goe meet; 2 weeks cal meet same time as today, i.e. 1500UTC?

13:46:59 <libby> I forgot that I'd said that we would revert to 1700utc after that meet

13:47:27 <libby> (this all arose because of an initial typo; we initially said 1500 utc until end may

13:50:58 <DanC> the standing item I put in my calendar was for 11am America/Chicago, i.e. 1700Z. I guess I could change it.

13:51:44 <Schuyler_> 1700 UTC works way better for us Californians

13:51:49 <DanC> but 9am means there's basically no time in my schedule to prepare. so I'm not sure it's worth coming.

13:53:01 <libby> how about we compromise?

13:53:19 <libby> 10am America/Chicago?

13:53:30 * libby wonders if that's any good for jibberJim

13:53:47 <danbri> danbri is now known as geodanbri

13:54:16 <DanC> it's awkward to change the time of the 23Apr meeting, now that you've announced it and people agreed to it.

13:54:33 <libby> true

13:54:42 <DanC> oh... I guess if we announced the change today, that would be 1 week notice, which is pretty reliable

13:54:45 <libby> it's difficult because 5UTC is 7pm europe

13:55:01 <libby> s/5UTC/1700UTC/

13:55:34 <zool> hey early morning Schuyler

13:55:40 <libby> jibberjim, how is 1600UTC for you in SA? for next calendar meet (maybe you weren;t thinking of comig anyway

13:55:46 <libby> )

13:56:09 * Schuyler_ groans

13:56:12 <Schuyler_> hi zool

13:56:33 <geodanbri> geodanbri has changed the topic to: RDF/SemWeb Geo/mapping meeting 1400UTC for 1hr http://esw.w3.org/topic/GeoInfo weblog: http://rdfig.xmlhack.com/

13:56:47 <DanC>http://lists.w3.org/Archives/Public/www-rdf-calendar/2003Apr/0013.html

13:56:47 <dc_rdfig> C: http://lists.w3.org/Archives/Public/www-rdf-calendar/2003Apr/0013.html from DanC

13:56:47 <geodanbri> libby, was it for 1hr? this meeting?

13:56:59 <DanC> C:|RDF geo meeting, 2003-04-16, 1500 UTC

13:56:59 <dc_rdfig> Titled item C.

13:57:01 <geodanbri> yup ~1hr

13:57:01 <libby> yeah

13:57:07 <mortenf> hi all

13:57:17 <JibberJim> it depends where I am lib, to whether I can get access then, otherwise I don't really mind.

13:57:30 <DanC> C:folks who want to be recorded as attending will please say so as a comment on this item

13:57:30 <dc_rdfig> Added comment C1.

13:57:54 <geodanbri> shall we start a separate item for the agenda, or use C:?

13:58:02 <DanC> C:duration: 1hr, thru 1600Z

13:58:02 <dc_rdfig> Added comment C2.

13:58:13 <DanC> pls use a separate item for each agendum

13:58:15 <mortenf> .time

13:58:15 <datum> Wed, 16 Apr 2003 13:58:15 GMT

13:58:19 <DanC> 1400UTC?

13:58:34 <libby> ok, I'll send a note around moving cal meet to 1600UTC so easier for US

13:58:34 * mortenf is confused?

13:58:35 <Schuyler_> you know, I *like* living in a world where finding a meeting time is difficult because relevant parties are spread far enough across the globe that someone inevitably is going to have to tune in either at the crack of dawn or the middle of the night.

13:58:42 * geodanbri copied it from wiki... I believe it means not-corrected-for-summertime

13:58:53 <geodanbri> Meeting is now anyways! or in 1minute...

13:58:56 <libby> yep danbri, 1400 UTC, 1500 BST

13:59:11 <mortenf> so 1hr, thru 1500Z?

13:59:24 <geodanbri> BLURB:Agenda review / Actions / AOB

13:59:24 <dc_rdfig> D: Agenda review / Actions / AOB from geodanbri

13:59:25 <Kake> Hello.

13:59:45 <geodanbri> BLURB:Agenda item 1: cal/geo usecase 1: opening hours of a thing that has a place

13:59:46 <dc_rdfig> E: Agenda item 1: cal/geo usecase 1: opening hours of a thing that has a place from geodanbri

13:59:57 <geodanbri> BLURB: Agenda item 2: cal/geo usecase 2: describing a journey

13:59:57 <dc_rdfig> F: Agenda item 2: cal/geo usecase 2: describing a journey from geodanbri

14:00:06 <DanC> pls call it agenda item E rather than agenda item 1

14:00:18 <earle> re Kake

14:00:27 <geodanbri> F1: Agenda item 'F': cal/geo usecase 2: describing a journey

14:00:28 <dc_rdfig> Comment F1 not found.

14:00:46 <geodanbri> E1: Agenda item 'E': cal/geo usecase 2: describing a journey

14:00:46 <dc_rdfig> Comment E1 not found.

14:00:54 <geodanbri> E1:Agenda item 'E': cal/geo usecase 2: describing a journey

14:00:54 <dc_rdfig> Comment E1 not found.

14:01:00 <geodanbri> hmm

14:01:08 <mortenf> no comments, only a title?

14:01:20 <DanC> C2:duration: 1hr, thru 1500Z

14:01:20 <dc_rdfig> Replaced comment C2.

14:01:24 <geodanbri> F:|Agenda item 'F': cal/geo usecase 2: describing a journey

14:01:25 <dc_rdfig> Titled item F.

14:01:38 <geodanbri> E|:Agenda item E: cal/geo usecase 1: opening hours of a thing that has a place

14:02:05 <DanC> C:|RDF geo meeting, 2003-04-16, 1400 UTC (not 1500, despite announcement)

14:02:05 <dc_rdfig> Titled item C.

14:02:06 <geodanbri> E:|Agenda item E: cal/geo usecase 1: opening hours of a thing that has a place

14:02:07 <dc_rdfig> Titled item E.

14:02:49 <DanC> C:has anybody offered to chair?

14:02:49 <dc_rdfig> Added comment C3.

14:02:52 <geodanbri> hi everyone... Can you sign up as follows (by my example)...

14:02:57 * geodanbri offered to chair

14:03:00 * geodanbri chairs

14:03:07 <DanC> C:chair: geodanbri

14:03:08 <dc_rdfig> Added comment C4.

14:03:20 <DanC> C3:""

14:03:20 <geodanbri> C:[http://www.w3.org/People/DanBri/|Dan Brickley], W3C.

14:03:20 <dc_rdfig> Deleted comment C3.

14:03:21 <dc_rdfig> Added comment C4.

14:03:30 <JibberJim> C: Jim Ley

14:03:30 <dc_rdfig> Added comment C5.

14:03:50 <mortenf> C:Morten Frederiksen

14:03:50 <dc_rdfig> Added comment C6.

14:03:55 <ericP> libby, a while ago, i documented a program for looking for functional properties in triples stores

14:03:56 <zool> C:jo walsh

14:03:56 <dc_rdfig> Added comment C7.

14:04:05 <ericP> do you remember this? or where it went?

14:04:21 <geodanbri> D:See [http://esw.w3.org/topic/GeoInfo|ESW:GeoInfo] for background / agenda etc.

14:04:22 <dc_rdfig> Added comment D1.

14:04:33 <earle> C:Earle Martin

14:04:33 <dc_rdfig> Added comment C8.

14:04:34 <Kake> C:[http://0xff.org/accretion.pl?Kake|Kake]

14:04:34 <dc_rdfig> Added comment C9.

14:04:56 * DanC isn't sure he can afford to attend

14:05:01 <libby> C:[libby miller|http://ilrt.org/discovery/libby/]

14:05:01 <dc_rdfig> Added comment C10.

14:05:05 * geodanbri sees 7 signups, let's start...

14:05:09 <mattb> C:[Matt Biddulph|http://www.hackdiary.com]

14:05:09 <dc_rdfig> Added comment C11.

14:05:14 <DanC> I should probably report from the DAML Spatial break-out

14:05:17 <geodanbri> =========== RDFIG geo chat ============

14:05:22 <libby> ericp, am looking

14:05:33 <libby> that sounded intersting danc

14:05:34 <alberto> C:[Alberto Reggiori|http://www.asemantics.com]

14:05:34 <dc_rdfig> Added comment C12.

14:05:36 <geodanbri> DanC, that would be great. If you're time pressured, we could do that earlier rather than as AOB.

14:06:03 <DanC> I could also go over my travel tool stuff, which is now written up for a wide audience.

14:06:11 <geodanbri> We had some action items last week. let's do those first...

14:06:12 <DanC> if I can do that, it's worth my time.

14:06:13 <zool> DanC: daml spatial? no OWL spatial? :)

14:06:21 <geodanbri> From http://esw.w3.org/topic/GeoInfo :

14:06:22 <geodanbri> [[

14:06:25 <DanC> indeed, daml spatial, not OWL spatial.

14:06:35 <geodanbri> -> ACTIOn libby write a converter from chefmoz opening housr format

14:06:35 <geodanbri> to rdf calendar format (or icalendar format maybe?)

14:06:35 <geodanbri>

14:06:42 <geodanbri> Libby, any progress on that?

14:06:46 <libby> sorry, no :(

14:06:49 <libby> apologies

14:07:04 <geodanbri> OK, continued (though we didn't arrange a next meeting yet...)

14:07:12 <geodanbri> -> ACTION danbri_fr to have a draft of the 'concentric circles'

14:07:12 <geodanbri> within100m schema/vocab for next week

14:07:17 <geodanbri> danbri, progress?

14:07:29 * zool splutters

14:07:33 <geodanbri> Some progress, though minimal. See...

14:07:49 <DanC> D:ACTIOn libby write a converter from chefmoz opening housr format to rdf calendar format (or icalendar format maybe?) [continues]

14:07:49 <dc_rdfig> Added comment D2.

14:07:59 <geodanbri> D:DanBri drafted [http://esw.w3.org/topic/GeoOnion|'geonion'] page.

14:07:59 <dc_rdfig> Added comment D3.

14:08:16 <geodanbri> I'd like to continue this, and hopefully discuss it a bit here today.

14:08:36 <DanC> under one of the existing agenda?

14:08:43 <geodanbri> D:Continued, ACTION danbri to have a draft of the 'concentric circles'

14:08:43 <dc_rdfig> Added comment D4.

14:08:57 <mortenf> what happened to the fibonacci numbers?

14:09:01 * geodanbri looks, let's try it under 'F'

14:09:11 <geodanbri> mortenf, my lack of math knowledge.

14:09:21 <geodanbri> ]]

14:09:24 * zool nods, fibonacci squared seemed a nice scale - and not too complex! :)

14:09:30 * DanC wonders where this list of actions is coming from

14:09:33 * geodanbri sees no more actions from prev meeting.

14:09:52 <geodanbri> They're in http://esw.w3.org/topic/GeoInfo and IRC log, not sure where circulated (Libby?)

14:10:23 <geodanbri> DanC, shall we AOB 'update from DanC on DAML spatial & travel tool writeup'?

14:10:38 <libby> linked from here: http://lists.w3.org/Archives/Public/www-rdf-calendar/2003Apr/0013.html

14:10:42 <DanC> separate items, pls

14:10:58 <libby> .time

14:10:58 <datum> Wed, 16 Apr 2003 14:10:57 GMT

14:11:19 <geodanbri> BLURB: AOB G: DanC update on DAML spatial work

14:11:20 <dc_rdfig> G: AOB G: DanC update on DAML spatial work from geodanbri

14:11:25 * DanC doesn't see actions in 0013

14:11:39 <geodanbri> BLURB: AOB H: DanC update on travel tool writeup

14:11:39 <dc_rdfig> H: AOB H: DanC update on travel tool writeup from geodanbri

14:11:48 <DanC> D:reviewed actions from [http://ilrt.org/discovery/chatlogs/rdfig/2003-04-09.html#T15-02-25|log of 9Apr meeting]

14:11:48 <dc_rdfig> Added comment D5.

14:12:00 <libby> yep http://esw.w3.org/topic/GeoInfo was the ref, danbri right

14:12:15 <geodanbri> Any other things folk want to discuss?

14:12:30 * geodanbri wonders if the opening hours item might be a bit light...

14:12:53 <alberto> waves to geodanbri (and libby)

14:12:57 <geodanbri> ==== Agenda item E: cal/geo usecase 1: opening hours of a thing that has a place

14:12:59 <DanC> but GeoInfo isn't a (permanent) record of the previous meeting; it's editable.

14:13:17 * libby notices that although in wiki says 1400 UTC, the email said 1500 UTC, so wwe may find people arriving then. apologies - this may have been my fault :( 1400 was the correct time.

14:13:39 <geodanbri> So Libby mentioned already that she didn't manage to get the chefmoz to ical conversion done. Did anyone else make progress on this scenario?

14:13:45 * DanC notes that "correct" is highly subjective, finds just as much evidence for 1500Z as 1400Z ;-)

14:14:01 <geodanbri> E:[http://rdfig.xmlhack.com/2003/04/09/2003-04-09.html#1049903200.904995|previous discussion].

14:14:01 <dc_rdfig> Added comment E1.

14:14:09 * geodanbri thinks we've enough folk here to continue

14:14:15 * geodanbri wishes for flat planet

14:14:17 <libby> heh, true.

14:14:39 <geodanbri> OK, recap. We are looking at merging geo and calendar and related RDF data.

14:14:44 <JibberJim> the far north and south would still then want daylight savings time...

14:14:48 <DanC> in email, we discussed contexts vs subabstractions a bit; does that fit here under E?

14:15:15 <zool> where was that, on www-rdf-calendar?

14:15:29 <zool> (contexts vs subabstrac)

14:15:34 <DanC> contexts vs subAbstract on www-rdf-calendar, yes: http://lists.w3.org/Archives/Public/www-rdf-calendar/2003Apr/0007.html

14:15:35 <geodanbri> I have been gathering fodder for representing restaurant reviews, and one issue that arose (hence ESW:GeoOnion) was that sometimes the lat/long/alt we have access to isn't quite for a point in the middle of a restaurant.

14:15:35 * mattb looks forward to w3c rss meta-list feeds

14:15:44 <geodanbri> yes DanC, that sounds relevant.

14:15:48 <zool> thx

14:16:15 <geodanbri> For example with ChefMoz, we have postcode locations in their dataset, and could use postcode2latlong lookups to get some relevant geo info.

14:16:28 <DanC> E:hmm... [discussion of contexts vs. subAbstrac | http://lists.w3.org/Archives/Public/www-rdf-calendar/2003Apr/0007.html]

14:16:28 <dc_rdfig> Added comment E2.

14:17:00 <DanC> have you seen my city/state/zip->lat/long rules, danbri? (in GeoInfo)

14:17:08 <geodanbri> The Onion thing tried (didn't yet manage fibbonacci, though libby tried a bit yesterday) to invent relations like 'within100M' to connect spatial things with some indicator of nearness.

14:17:09 <zool> where cyc:holdsIn indicates a context - of any kind not just geo/cal ?

14:17:26 <geodanbri> DanC, yup, I saw that. Hmm did I get it working with Postcode? I remember trying...

14:17:37 <DanC> could somebody elaborate this use case? which parties do what?

14:17:40 <geodanbri> MattB, you were also looking at this (lookups) weren't you?

14:18:13 <DanC> my address->lat/long is only for u.s. postal addresses.

14:18:15 <zool> the backend to that is not yet publically loggable geodanbri

14:18:40 <DanC> oh... I haven't connected my u.s. postal schema to wg84. i meant to look at that...

14:18:41 <geodanbri> Its a mess of related use cases. One might be: find me a vegetarian restaurant in central bristol near to the hostel I'm staying at (whose postcode is BS6 7AD / whose latlong is within 100M of ...)

14:18:46 <zool> the UK postal data is not free-for-use sadly

14:18:52 * geodanbri nods to zool

14:19:04 <mattb> geodanbri: yeah, sample model at http://www.picdiary.com/~mattb/rdf/e97qp.rdf

14:19:06 <mortenf> the dk data is free x 2, and available.

14:19:14 <femke> what about varying projections?

14:19:16 <mattb> but as zool notes the data is not openly available, hence static snapshot url there

14:19:17 <DanC> free x 2?

14:19:26 <mortenf> beer etc...

14:19:46 <geodanbri> mattb, can you talk us through that RDF file a bit?

14:19:51 <mortenf> as in, both kinds of free.

14:19:55 <JibberJim> I'm also finding that I want to describe the cafe just down the road from where I'm staying - I don't always carry my GPS everywhere, but I will know it's 5 miles from some place I have got positions for, and that's still useful.

14:20:00 <zool> Kake did a nice simple locator irc bot using my rdf backend; 'show me pubs near here'... with plans to tie into a category/ontology system

14:20:14 <wkearney99> mattb, what scheme is it using? you'll want to allow for other schemes...

14:20:26 * DanC reviews http://www.w3.org/2000/10/swap/pim/usPlace2LatLong.n3

14:20:26 <mattb> it's a model representing the data held in the html page http://www.upmystreet.com/latlong.php3?l1=E9+7QP

14:20:30 <JibberJim> femke - I think we try and escape projections mess by defining the geo:lat/lon to be wgs84

14:20:46 <mattb> really just trying to tie together two different representations of the same geographic point

14:20:53 <mattb> "this geo:Point has both this postcode and this lat/long"

14:20:57 * zool agress with jim a lot of places in my db don't have lat/longs but are 'connected' e.g. def within Nmetres of places that have wgs84

14:20:59 * earle nods to zool... we are close to rdf'izing Grubstreet output.

14:21:01 <mattb> and some rdfs:seeAlso type links

14:21:04 <geodanbri> BLURB: AOB I: DanBri re relationship to GML (chat w/ Simon Cox)

14:21:04 <dc_rdfig> I: AOB I: DanBri re relationship to GML (chat w/ Simon Cox) from geodanbri

14:21:06 <wkearney99> forcing use of wgs84? why? encouraging it as the default is certainyl worth using. But excluding the others?

14:21:12 * geodanbri remembers

14:21:18 <libby> hey danja

14:21:23 <earle> er, to continue... output, to feed into Kake's bot.

14:21:43 <danja> hi Libby

14:21:45 * DanC doubts we're going to get thru this agenda in 1hr

14:21:53 <mattb> wkearney99: i'm not excluding anything

14:21:55 <JibberJim> you're free to use other co-ordinate systems, I've yet to see a way that is clean though, the map thing danbri posted this morning had problems.

14:22:02 <mattb> just using existing vocab to describe the data at hand

14:22:03 <zool> opening hours usefu; does grubstreet include them?

14:22:13 <geodanbri> bill, the reason at the moment is just pragmatic... there wasn't (GML being XML but not RDF) any way of doing it in RDF. So we made a start w/ the wgs_84 vocab to get some experience in mixing some simple geo vocab with other RDF vocabulary.

14:22:16 <mattb> that same geo:Point could have any arbitrary representation system info attahced to it

14:22:22 * geodanbri nods re agenda

14:22:25 <wkearney99> dc's methods support indicating an optional scheme, as goes GML

14:22:27 <Kake> zool: Yes, but currently only as a freeform text string.

14:22:29 <zool> also is it possible / desirable to specify 'likely' opening hours for somewhere of type Pub or GreasyCafe

14:22:30 <earle> zool: we will be including them as metadata very soon, yes.

14:22:38 <mattb> the idea being just to same that "point X is represented as Y in system 1 and Z in system 2; they are the same conceptual location"

14:22:53 <mattb> s/same/state/

14:22:57 <alberto> interesting mattb

14:22:59 <geodanbri> So re the AOB I just suggested, I had mail from Simon Cox, who did much of the DC geo work before moving to GML. He noted two things:

14:23:06 <JibberJim> The optional schemes suffer from a problem in everything I've seen, I can discuss the problems, but I think that's for another day.

14:23:11 * mattb confuses his words a bit; apologises, in midst of job-work

14:23:15 <geodanbri> (i) that GML has a very RDF like model for its instance data, striped syntax etc.

14:23:30 * zool very interested to see that RDFMap now proper RDf rather than RSS hack; are there any applications which use it?

14:23:40 <geodanbri> (ii) that we will for various reasons eventually want to represent other schemes than wgs_84

14:23:54 <DanC> I used to worry a lot about identity, ala "same conceptual location". I don't any more. it rarely matters. And where it does matter, there's usually an existing controlled vocabulary, e.g. u.s. state codes, zip codes, ISO country codes, internatoinal airport codes

14:23:59 <geodanbri> both of which I agree with. Think of the wgs_84 as a toe in the water for RDF geo stuff.

14:24:18 * geodanbri nods

14:24:20 <mortenf> rdfmap still has the conversion problem ...

14:24:25 <zool> yeah... i also use the Ordanance Survey grid reference where i can get them

14:24:27 * DanC is kinda lost are we on GML now?

14:24:41 * geodanbri was replying to wkearney's point (as didn't expect to get to the AOB).

14:25:07 * earle wondering what the current status of http://www.daml.org/cgi-bin/hyperdaml?http://www.w3.org/2000/10/swap/pim/contact# is

14:25:12 <JibberJim> the problem with rdfmap is that you can't have multiple x,y's with different srs in a single point - the srs applies to the point, not to the x value - I think this is a huge problem for tools trying to decide if two places are the same.

14:25:15 <geodanbri> OK, I'm not used to chairing IRC meetings! Shall we move onto describing a journey...

14:25:20 <earle> oops, strip out the hyperdaml from that, sorry.

14:25:30 * zool thinks hard to disagree re need for mulitple grid reference systems

14:25:34 <geodanbri> srs?

14:25:43 <DanC> E doesn't feel done to me, but I wasn't quite sure what it was anyway.

14:25:45 <zool> JibberJim: oh, that sounds pretty flat

14:25:55 * danja thinks he must be GTM+2...

14:25:57 <JibberJim> in rdfmap srs is the pred which defines the co-ordinate system.

14:26:02 * mattb wonders about geo-smushing disparate rdf

14:26:09 <geodanbri> In E, I had hoped to be in a position to go "TaDa! Here is ChefMoz in Geo + onion markup..."

14:26:14 <mattb> requires a bit more data and inference than just sharing same foaf:mbox...

14:26:22 * zool thinks that will be hard to do mattb

14:26:28 <libby> danja, there was confusion over the time - appologies

14:26:44 <danja> ok, thank libby

14:26:46 <JibberJim> some tools would want to do nearSmushes I think.

14:26:48 <mortenf> but the smushing should be easier with the onion vocab...

14:26:57 <geodanbri> mattb, quite... I was worried about identity conditions for g:Point, tried to make points smushable with lat/long/alt as identity conditions

14:27:19 <geodanbri> but i think 'within100M' stuff should help with data merging, though maybe in a less generic way.

14:27:25 <DanC> does anybody here have the ability to change how ChefMoz works? or were you thinking of a layered service, geodanbri?

14:27:26 <alberto> has anybody tried to describe polygons instead of just points with some RDF voc ala GML ?

14:27:42 <geodanbri> DanC, the last I heard from chefmoz contact was ~3 years ago...

14:27:46 <zool> i dont have a good enough handle on GML

14:27:50 <DanC> stop worrying about merging. just worry about notions you can exploit, like "within 100M"

14:27:53 <geodanbri> I was thinking layered / copied and fixed, initially

14:28:00 <zool> the pdf specs are so huge and overcomplex seeming

14:28:12 <geodanbri> So sticking with E, does anyone have a good idea for elaborating the 'onion' vocab I sketched...

14:28:16 <JibberJim> I've thought about alberto - trying to reuse my image region markup, but no more than that. There I just used SVG's path methods for describing a path.

14:28:16 <danja> alberto - sounds painful in straight RDF, but the co-depiction does it using SVG

14:28:24 <geodanbri> ie http://esw.w3.org/topic/GeoOnion

14:28:43 <geodanbri> ...the gaps were wrong in that scale: too much space between 1Km and 10Km

14:28:50 <geodanbri> ...which is probably the most useful bit of the space

14:29:04 <DanC> re GeoOnion, NaryRelations rears its ugly head. it's most naturally expressed as (within (miles 100) myHome theAirport)

14:29:25 <geodanbri> yes, this design does smell a bit like RDF's rdf:_1 _2 predicates

14:29:31 <zool> hmm

14:29:49 * mattb agrees re within100M thing above

14:29:50 <alberto> danja, [some satellite image RDF description|http://uranus.esrin.esa.it/showcase/env/Canada/Edmonton_ASAR_APP_Orbit5055_20030217_20030317.rdf]

14:29:57 <geodanbri> Is there anyone here who could take a look at the use of polygons in an RDF context, perhaps report back to www-rdf-interest?

14:30:01 <mortenf> hm.

14:30:03 <DanC> but I suppose GeoOnion is an 80% solution: (within100miles myHome theAirport)

14:30:10 <geodanbri> (noting that the SVG spec has a section on this... @@url)

14:30:15 <mattb> rather than actual smushing, one might produce an automatic model that says who's near who, rather than asserting node identity

14:30:19 <zool> mortenf, werent you planning to extrapolate shapes for regions from your danish data

14:30:39 <libby> :within10YardsOf rdfs:subPropertyOf cyc:near.....

14:30:40 <geodanbri> Yes, I was hpoing for 80%, but it seems to depend on getting a useful set of relations. My attempt yesterday to do that failed...

14:30:47 * DanC spent a semester studying convex hulls... has been thinking about writing left tests in N3...

14:30:51 <mortenf> yep, planning being the operative word...

14:30:59 <DanC> failed in what way, geodanbri?

14:31:03 <geodanbri> yep, I kept finding 'near' crop up as something I want, but I know its undeployably vague.

14:31:18 <geodanbri> DanC, I couldn't find some maths that generated intuitively useful sequence of distances

14:31:35 <zool> i have just used a near bnode, with distance and units, in the past, but that caused some objections here before?

14:31:51 <geodanbri> zool, do you have an example we could be pointed at?

14:32:01 <DanC> E:this seems to be a collection of related use cases. one [offered by danbri] is: find me a vegetarian restaurant in central bristol near to the hostel I'm staying at (whose postcode is BS6 7AD / whose latlong is within 100M of ...)

14:32:02 <dc_rdfig> Added comment E3.

14:32:35 <geodanbri> it sounds ok-but-verbose to me... and (as above) may be hard to query directly (without fancy datatyping support in query system)

14:32:41 <DanC> ok, so in the E3 case, 'near' means what? 100 meters?

14:33:03 <alberto> geodanbri, in the context of RDF-ying some satellite pics HTML descriptions we end up describing something using polygons, but the vocabulary at the moment is an hybrid between GML XML and a comprosime [see RDF|http://uranus.esrin.esa.it/showcase/env/Canada/Edmonton_ASAR_APP_Orbit5055_20030217_20030317.rdf] and [original HTML|http://uranus.esrin.esa.it/showcase/env/Canada/Edmonton_ASAR_APP_Orbit5055_20030217_20030317.htm]

14:33:04 <geodanbri> For me (relatively able bodied, non-driver), 100m-1000m

14:33:14 <DanC> or is near really measured in time-and-money?

14:33:41 <zool> geodanbri, e.g. http://space.frot.org/a_space/London/Holborn_Station?nearest_type=Pub&nearest_dist=2 at the bottom

14:33:42 <DanC> i.e. a <=10minute walk, or a <=$10 cab ride

14:33:51 <geodanbri> Looking to wind up discussion on E... anybody feel volunteery to work on refinements / ACTION for continued work?

14:33:58 * DanC notes there's a whole world of AI planning here...

14:34:02 <zool> i didnt even separate out the units and number,s that is a bit of a hack sorry

14:34:12 * zool grins at DanC

14:34:19 <JibberJim> near is really measured depending on the context, if I normally live in South Africa, and want to meet danbri, knowing I'm in London and he's near in BRS may be handy.

14:34:20 <dirkx> Looking at the use of SGV - I fear it is not exactly what one wants; as compared to GML or some other method.

14:34:22 <geodanbri> I will take a crack at generating some/any/a 'onion' based model of some ChefMoz data.

14:34:48 <femke> near depends on accessibility

14:34:58 <danja> sorry Jibber

14:35:00 * libby thinks some sample data would be very useful using the onion vocab

14:35:02 <zool> G:[http://carto.net|carto.net] SVG for cartographers

14:35:03 <dc_rdfig> Added comment G1.

14:35:03 <dirkx> Reason is that it does not take height and geoids into account with enough descriptive ness to allow for decoding the data, say, 20 years in the future.

14:35:06 * DanC wonders if I'm the only one who doesn't know what GML is

14:35:11 <alberto> well, the RDF query as such can not "solve" the spatial-query problem x1<x2 && y2<y1 at the moment efficently, but only geo:Point is a bit restrictive...

14:35:14 <JibberJim> probably not dirkx the reason I used in image description, is because the property is easy to generate, and easy then to reuse.

14:35:15 <geodanbri> E:ACTION: danbri produce some sample chef-moz based data using onion + wgs_84 vocab.

14:35:15 <dc_rdfig> Added comment E4.

14:35:18 <zool> the OpenGIS consortium geography ML

14:35:40 <zool> G:[http://www.opengis.org|openGIS]

14:35:40 <dc_rdfig> Added comment G2.

14:35:53 <dirkx> Aye - and the nice thing about GML is that for points and polygons it is absolutely trivial; yet able to deal with things over time.

14:35:54 <femke> GML = geographic ML

14:35:57 * geodanbri waves to dirkx, invites him/her to sign up by typing C:Firstname Surname

14:36:03 <zool> i should refocus soon can i mutter about F for a bit

14:36:06 <danja> -Jim, you mentioned 'srs' earlier - what they?

14:36:09 <dirkx> C: Dirk-Willem van Gulik

14:36:09 <dc_rdfig> Added comment C13.

14:36:20 <chrisg05> C:Chris Goad

14:36:21 <dc_rdfig> Added comment C14.

14:36:24 <geodanbri> OK, let's move on agenda....

14:36:31 <geodanbri> ==== Agenda item 'F': cal/geo usecase 2: describing a journey

14:36:34 * DanC would prefer to skip right to G

14:36:35 <JibberJim> danja - in http://www.mapbureau.com/rdfmap1.0/index.html it calls the predicate for the projection used map:srs - I know not why.

14:36:40 <geodanbri> Zool, can you introduce this?

14:36:41 <alberto> waves to dirkx

14:36:48 <danja> JJ - thanks

14:37:04 <wkearney99> C:Bill Kearney

14:37:04 <dc_rdfig> Added comment C15.

14:37:09 <chrisg05> srs means "spatial coordinate system"; terminology used,eg, by gml

14:37:10 <geodanbri> hmm zool, can we take G first?

14:37:18 <geodanbri> G being DAML spatial update...

14:37:21 <dirkx> Also - distance and nearness would not be hard - in fact it is not uncommon to find polygons or big circles around an object labeled/fetured as "this is near" and "this is very near" -- so that finding nearness is just intersect.

14:37:24 <JibberJim> srs=scs so still confusing :-)

14:37:34 <geodanbri> F:... postponed until after G

14:37:35 <dc_rdfig> Added comment F1.

14:37:38 <geodanbri> AOB G: DanC update on DAML spatial work

14:37:41 <geodanbri> ===

14:37:41 <dirkx> Aye SRS is needed to decode

14:37:46 <geodanbri> DanC? Tell us about DAML...

14:37:52 <chrisg05> whoops "spatial reference system" is what I meant to say

14:38:03 <DanC> F:my [travel schedule|http://www.w3.org/People/Connolly/travel-sched] has lat/long info. has dates too, though not in icalendar vocab

14:38:03 <dc_rdfig> Added comment F2.

14:38:24 * DanC collects thoughts on DAML spatial...

14:38:25 <JibberJim> dirkx the problem with that is rivers, and conceptual distances over real ones - I may be only 20 miles from france standing in Dover, I still think it might be easier to go to london for a decent meal.

14:38:35 <DanC> so I went to the DAML PI meeting last week in Miami

14:38:48 <DanC> <geo:lat>25.82897949</geo:lat>

14:38:48 <DanC> <geo:long>-80.28563690</geo:long>

14:38:51 <DanC> 1/2 ;-)

14:39:07 <geodanbri> DanC, can you do this in 5-10 mins?

14:39:13 <DanC> yes...

14:39:18 <JibberJim> thanks chrisg05, always tough to understand where the abbreviations come from when you're outside.

14:39:48 <dirkx> JibberJim: aye - this is something commonly a problem in GIS - GML offers some nice fixes; as do layers which define nearness for a given object as anything which is inside it. And they are often (say for a train table) NOT geographic but 'warped' as to accomidate road quality.

14:39:59 <DanC> from http://esw.w3.org/topic/GeoInfo , under "Is anybody building an RDF vocabulary for stuff like latitude/longitude?" there's a link to the daml-spatial archive

14:40:32 <DanC> G:[introductory message to daml-spatial from Jerry Hobbs|http://www.daml.org/listarchive/daml-spatial/0005.html]

14:40:33 <dc_rdfig> Added comment G3.

14:40:46 * danja wonders if nearness can be done 'confidemnce factor' style

14:40:59 <DanC> DAML is a DARPA research program...

14:41:00 <danja> confidence

14:41:22 <DanC> theire doing RDF stuff, starting with DAML+OIL, which fed into OWL; i.e. InverseFunctionalProperty and such...

14:41:41 <DanC> ... lots of people have built vocabularies (LinkMe: DAML ontology library) ...

14:41:52 <DanC> ... and, of course, time and space emerged as recurring themes...

14:41:57 <geodanbri> Are they collaborating on vocabularies? Or is it 1-per-funded-project tpyically?

14:42:06 <danja> i.e. [danc xxx:near Miami] cf:confidence 0.8

14:42:07 <DanC> starts as 1-per-funded-project

14:42:20 <dirkx> DanC: <geo:long>-80.28563690</geo:long> -> kind of is meaningless without it's lat. which is kind of hard to interpret without its projection - which kind of needs the Hight for most projections to make sense...

14:42:38 <DanC> some folks collaborated on daml-time. It's seen as a success, though when I asked if anybody was *using* it, I didn't get much of an answer.

14:43:01 * zool did some work on collaborative daml spatial ontology based on the GeoNET classifications from nima.mil/gns ...

14:43:02 <geodanbri> are you active on that list? chasing up some real uses would be cool...

14:43:16 <zool> damltime struck me as wildly overengineered

14:43:21 <DanC> dirkx, I was able to use geo:lat/long without height to plot my itinerary on a map: http://www.w3.org/2003/04dc-mia/itin-mia.png

14:43:23 <geodanbri> zool, was that the irc/aim-interfaced thing?

14:43:42 <dirkx> <position proj="xx" date="YY"><geo:long>-80.28563690</geo:long> <geo:long>-80.28563690</geo:long><geo:height ref="NAP">0.30</></>... date added in case of things like WGS84 which need a date to be able to map resonably to US or EU local grids.

14:43:44 <DanC> active: I forwarded the announcement of this meeting (with the wrong time) to daml-spatial

14:43:49 <DanC> and I browsed the archive.

14:43:57 <earle> G:[DAML ontology library http://www.daml.org/ontologies/]

14:43:58 <dc_rdfig> Added comment G4.

14:43:59 <DanC> Pease is another name to know.

14:44:22 <femke> given that nearness is not purely a spatial concept, ie its about accessibility - so temporal also, has anyone looked into a spatio-temporal ontology or an accesibility ontology?

14:44:51 <DanC> I spoke with him (Adam Pease). doing lots of good work on SUMO at teknowledge.

14:44:57 <geodanbri> Hmm: does a point with lat/long but not height define a set of points from earths centre into space, or cutting through earth at two points?

14:45:06 <geodanbri> s/define/...?

14:45:06 <DanC> at the break-out in miami, one of Hobbs's slides was: just use cyc? or SUMO?

14:45:13 <mortenf> one point, at the surface?

14:45:13 <alberto> I like geo wgs84_pos voc simplicity (i.e. going to work) but we need to decide about lines/polygons (yes/no) and recommend some default properties to annotate the geo:Point with a label and URL pehaps - that would be enough for a trivial mapping application to overly RDF descr over a map

14:45:33 * DanC gathers interest in DAML spatial is waning

14:45:48 <geodanbri> (alberto, would you be prepared to take a look at that and report back to us?)

14:45:53 <mortenf> really?

14:46:06 <geodanbri> DanC, that was useful. Do you have much sense on 'where next' re daml spatial?

14:46:10 <DanC> one frustrating thing about the DAML meeting is that all the presentation materials were .ppt on laptops. No http sharing. :-{

14:46:16 <sandro> Really, DanC? Post-breakout did everyone go "wow, this is too hard" ???

14:46:34 <DanC> I meant waning in this IRC meeting...

14:46:37 * zool dropped off there gah

14:46:39 <sandro> Ah!

14:46:45 <chrisg05> idea: geometry should be in a separate RDF vocabulary not mapping

14:46:51 <chrisg05> see second paragraph of www.mapbureau.com/rdfmap for an argument

14:46:52 <alberto> yes geodanbri, I could just write a couple of examples using current geo voc with some labels e.g. FOAF with lat/long

14:46:56 <zool> re daml-spatial sounds inteersting if people can get publically involved

14:47:04 <zool> and it#'s not closed-doors funded projects

14:47:06 <DanC> as to where next... I'm not 100% convinced there's a critical mass in the daml-spatial forum.

14:47:22 <geodanbri> Do they have any tools or web services we could use?

14:47:38 <DanC> Hobbs has agreed to serve as editor, and there are some interesing applications, but "just use SUMO" or "just use cyc" seems as good a bet as "wait for DAML spatial"

14:47:44 <JibberJim> or even just examples using their vocabs

14:48:10 <geodanbri> alberto, I was thinking more re "but we need to decide about lines/polygons (yes/no)", though labelling of points interest too.

14:48:14 <sandro> Indeed: this is where the DAML-Presentation are supposed to be linked, and only mine, Mike Deans, and Roger Costellos are so far. http://www.daml.org/meetings/2003/04/pi/agenda.html

14:48:22 <geodanbri> the onion design sort of connects points to more easily labelled things

14:48:24 <dirkx> Or just grandfather some existing standard in for the 'encoding part' - USGS has them, OpenGIS has them - as do half a dozen other GIS related initiatives.

14:48:32 <DanC> the "just use SUMO" option is much more interesting to me since talking with Pease in Miami. cyc is still pretty monolithic; the SUMO folks have done the work of cutting it into 11 modules.

14:48:33 <dirkx> The encidong is the harder part - the what is "easier"

14:48:48 * libby would like to be able to show how you coudl start with somethign very simple and then add more complex ontologies that say something similar

14:49:27 * DanC finds it hard to stick to DAML spatial with all the other interesting dicussion going on

14:49:29 <geodanbri> Yes, what libby said. wgs_84 vocab was the simplest thing i could think of for making a start in that direction...

14:49:37 <alberto> geodanbri, perhaps foaf groups could be "mapped" to a geo:Polygon (area not point) - I will sketch some examples with points and polygons

14:49:38 <danja> any likelihood of SUMO in OWL in the near future?

14:49:45 <geodanbri> DanC, I think we've probably got all we can from Spatial...

14:49:57 <geodanbri> Let's move back to the skipped agenda item...

14:50:10 <earle> G:[http://ontology.teknowledge.com:8080/rsigma/SKB.jsp?req=SC&skb=SUMO&id=773|SUMO definition of 'city']

14:50:10 <dc_rdfig> Added comment G5.

14:50:28 <DanC> SUMO is already in OWL... well, daml, I guess.

14:50:31 <geodanbri> ==== Agenda item 'F': cal/geo usecase 2: describing a journey

14:50:36 <geodanbri> (if that's OK folks)

14:50:56 <geodanbri> Zool... could you say a few words introducing this?

14:51:08 <zool> k sorry

14:51:23 <zool> F:last week we briefly discussed naming routes as either a series of points

14:51:23 <dc_rdfig> Added comment F3.

14:51:36 <zool> F:or a series of journeys starting and ending at points

14:51:36 <dc_rdfig> Added comment F4.

14:51:55 <DanC> G:see [S. Reed on cyc spatial relations|http://www.daml.org/listarchive/daml-spatial/0019.html] and [A. Pease on SUMO stuff|http://www.daml.org/listarchive/daml-spatial/0004.html]

14:51:55 <dc_rdfig> Added comment G6.

14:52:01 <zool> F:i said i'd have a go at sketching out the former and that is at http://space.frot.org/show_route/test

14:52:01 <dc_rdfig> Added comment F5.

14:52:22 * geodanbri takes a look

14:52:24 <mortenf> example (sorta old): http://xml.mfd-consult.dk/foaf/morten-0301-ams.rdf

14:52:27 <zool> i knjow the calendaring stuff is way off

14:53:09 <geodanbri> Is <rdf:Description rdf:about="_:id0915722"> intended to be a bnode or a URI?

14:53:18 <mortenf> uses cyc

14:53:39 <geodanbri> Sorry, straight in there with the nitpics! this looks pretty handy...

14:53:53 <zool> geodanbri, a uri really so that statements can be made about it elsewhere; foaf:Depicts, stuff like that

14:53:56 <geodanbri> ...could decorate the waypoints with photos and lat/long quite nicely

14:53:57 * DanC missed the connection between G and openGIS; was that supposed to go under I?

14:54:01 <zool> that's where i wanetd to annotate with groups of people

14:54:32 <geodanbri> DanC, lets try do that under I if we get there....

14:54:33 <DanC> F:[sketch|http://space.frot.org/show_route/test]

14:54:33 <dc_rdfig> Added comment F6.

14:54:39 <zool> ah, G sorry yes... i have shit net access and have to be working really

14:54:45 <JibberJim> but it's difficult to annotate the moving portion of the rdf with cost/method/photos etc. - you only get to annotate the places.

14:54:55 <geodanbri> We backed up to the skipped 'F'

14:55:23 * zool backgrounds sorry have to fix this pub quiz machine

14:55:39 <alberto> (is there a mail list for RDF/GEO or shall we use www-rdf-calendar ?)

14:55:53 <geodanbri> use www-rdf-interest for now, I think. Unless specifically calendar related...

14:56:17 <libby> zool, interesting stuff

14:56:19 <DanC> F:space:name... hm... see esw:PropertiesForNaming

14:56:19 <dc_rdfig> Added comment F7.

14:56:35 <DanC> F:space:location... hmm... httpRange-14

14:56:36 <dc_rdfig> Added comment F8.

14:56:41 <geodanbri> OK we have looming schedulling issue. Meeting as begun finishes in 5 minutes. However as announced, it can be believed to start in 5 minutes. Is there interest in continuing into the next hour?

14:56:55 <chrisg05> yes

14:57:00 <libby> yes

14:57:01 <zool> please dont make me resubscribe to www-rdf-interest

14:57:09 <bijan> Oh dear.

14:57:10 * DanC is willing to stay for H

14:57:12 <JibberJim> space:name could be a supProperty of foaf:name ?

14:57:20 <DanC> agenda + new mailing list?

14:57:24 <geodanbri> shall we go until half past?

14:57:31 <zool> JibberJim: sure, the schema is still totally vague

14:57:32 <alberto> are space:name and space:location used to annotate the "point"

14:57:46 <libby> I think there's definitely an interest in a mailing list, wonder if there's something existing could use

14:57:51 <geodanbri> I would support an rdf-geo mailing list. That would be worth extending the meeting to discuss, imho.

14:58:01 <JibberJim> (if it was zool, my geo svg would consume it and plot it on the map. although all the in same place as it only works on world scale.

14:58:10 <DanC> BLURB: agenda J: new rdf/geo mailing list?

14:58:11 <dc_rdfig> J: agenda J: new rdf/geo mailing list? from DanC

14:58:25 <DanC> J:yes, please

14:58:25 <dc_rdfig> Added comment J1.

14:58:27 <geodanbri> any objections to extending the meeting until half-past, continuing the agenda in http://rdfig.xmlhack.com/

14:58:34 <libby> so zool, there's an implicit ordering to the route but not an explicit one...

14:58:36 <DanC> half-past: 2nded

14:58:42 <libby> (I don;t know what's best)

14:58:48 * geodanbri looks around, sees no objections

14:58:53 <alberto> if yes, we could have space:name = geo:label (or rdfs:label) and space:location = geo:url (or rdfs:see:Also) or something

14:59:01 <danja> might also be a good place for danbri to testbed the list -> rdf stuff

14:59:03 <geodanbri> resolved, #rdfig stays geocentric for another 1/2 hour.

14:59:10 <dirkx> alberto: it almost needs to be 'space:cooverage" rather than space:location :-)

14:59:17 <libby> how does it compare with your travel stuff jibberjim?, danc? could you chump links?

14:59:22 <geodanbri> OK, we're still on 'F', journey description.

14:59:47 <geodanbri> Zool, did you say everything you had time to say? Shall we compare/contrast w/ Jim and DanC's now...

14:59:47 * mortenf is leaving, back later...

14:59:52 <DanC> well, I could hunt up itinerary links for F, or I could present it under H

14:59:54 <libby> bye mortenf

14:59:54 <geodanbri> cheers mortenf

15:00:01 <alberto> dirkx, that would be just for "labelling" a point with a URL + feature/label (mainly for display at the start)

15:00:01 <geodanbri> H seems good

15:00:02 <zool> geoanbri, only dspeculations

15:00:11 * danja likes journey description, first good Windows app I saw was a route planner

15:00:34 <geodanbri> Jim, wanna show us your latest experiments?

15:00:40 * libby would liek to try it out with a bus route (an alleged journey description ;)

15:00:42 <DanC> H:[http://www.w3.org/2000/10/swap/pim/travel|Semantic Web Application Integration: Travel Tools] is now "finished" in a certain sense

15:00:42 <dc_rdfig> Added comment H1.

15:00:48 <JibberJim> Mine libby is similar, but I don't have a series of waypoints, I would have a series of journeys between 2 points, this allows me to annotate the journey, and the points. I think the reason me and Zool differ is she's interested in walking around London, and I'm interested in flights and buses across continents.

15:01:11 <DanC> I use cyc:subEvents to relate a trip to a flight, e.g.

15:01:11 <libby> got a link jim, I've lost it again

15:01:22 <Kake> Re bus routes, would be lovely to get the data from the www.busmap.co.uk guy but I have failed to explain to him what raw data is, maybe someone else could have a go.

15:01:33 <libby> kake, heh

15:01:36 <JibberJim> [http://jibbering.com/travels/places.rdf|A journey from London to Plettenberg Bay] in RDF.

15:01:40 * geodanbri congrats DanC on declaring finished-ness (a rare thing w/ semweb tools?:)

15:02:16 <JibberJim> F:[http://jibbering.com/travels/places.rdf|A journey from London to Plettenberg Bay] in RDF.

15:02:16 <dc_rdfig> Added comment F9.

15:02:26 <libby> thanks jim

15:02:31 <DanC> I'd like to push to H and then come back to F

15:02:40 <JibberJim> F:[Also available viewed in SVG|http://jibbering.com/travels/places.svg]

15:02:40 <geodanbri> Jim, how does your markup differ from Zool's?

15:02:40 <dc_rdfig> Added comment F10.

15:03:01 <geodanbri> we've already punted F once, and it was scheduled....

15:03:21 <JibberJim> As I say, Zools smallest atom, are single points in space, mine are journeys between 2 points.

15:03:25 * DanC stands by

15:03:35 <geodanbri> sorry i missed that

15:03:43 <zool> i had a few speculations about recurring routes in time, modelling e.g. a bus or train

15:03:45 <libby> are the two compatible - could you join the two together jibberjim, zool?

15:04:02 <zool> i wonder whteher that use case throw light on usefulness or otherwsie my and jims approach

15:04:26 <zool> i could think on coalesing aye

15:04:31 <DanC> "that use case"?

15:04:35 <geodanbri> That would be very cool...

15:04:39 <DanC> that?

15:04:41 <geodanbri> yep, which use case?

15:04:50 <geodanbri> that: compare/contrast

15:04:55 <libby> the timetables one?

15:04:58 <zool> well, the use case being, i want to catch a bus to a certain place by a certain time, what bus do i catch?

15:04:59 <JibberJim> I think they are 2 different use cases. Zool's for the many waypointed journey - a walk - and mine for when the journey aspect of the route is as important as the points.

15:05:06 * zool sorry bit hectic env atm

15:05:22 <libby> I'd like to have a go at the timetable usecase

15:05:30 <libby> maybe closer to yours zool

15:05:43 <geodanbri> libby, can you formulate a sensible action item for y'self to that effect?

15:05:48 <dirkx> zool/jim: virtually all route systems I am aware of treat a journey as an ordered sequence of waypoints. Along with information on the way between the waypoints; typically directional. Both zool/jim their models fit this and/or can be casted (ignoring direction)

15:05:52 * zool nods timetables... another 'context' type problem... with an idea that is a train line or a bus route... and an individual bus which is a subAspect of that., hoevere stated

15:06:08 <DanC> F:[my ltest travel itinerary|http://www.w3.org/2003/04dc-mia/itin-mia.rdf], containing enough iCalendar info to import into evolution and enough geo info to plot on a map

15:06:09 <dc_rdfig> Added comment F11.

15:06:15 <JibberJim> yes dirkx, I think both can be easily modeled together.

15:06:18 <DanC> I think JibberJim's tools work on my data

15:06:20 <libby> are we on F?

15:06:46 <geodanbri> Agenda review: we have 24 mins, and H (DanC's travel tools) and I GML and J (rdf/geo list) to cover.

15:06:46 <libby> blimey danc! very impressive

15:07:08 * zool reminds that i am just sketching really

15:07:11 <zool> bbl

15:07:13 <dirkx> jibberJim: though the issue left is the -ordered- aspect and the directional one. Those are not always easy/

15:07:20 <geodanbri> Let's move on to H. I withdraw 'I' (GML thing) as I've reported Simon's comments above, and will send mail on the subject.

15:07:29 <libby> F:ACTION libby translate a sample bus timetable to zool/jim's formats, feedback to some list (www-rdf-calendar maybe?)

15:07:29 <dc_rdfig> Added comment F12.

15:07:39 <geodanbri> ACTION: danbri send mail (to www-rdf-interest) on RDF geo / GML relationship

15:07:39 <dc_rdfig> Label ACTION not found.

15:07:48 <geodanbri> Hope that's OK. running out of time again otherwise...

15:07:53 <geodanbri> =======

15:07:58 <geodanbri> ===== H: DanC update on travel tool writeup

15:08:09 <DanC> ok, somebody look at http://www.w3.org/2000/10/swap/pim/travel and tell me what version you find

15:08:14 <JibberJim> F:[http://jibbering.com/trip/display.svg?url=http%3A%2F%2Fwww.w3.org%2F2003%2F04dc-mia%2Fitin-mia.rdf|SVG View of DanC's route]

15:08:14 <dc_rdfig> Added comment F13.

15:08:15 <geodanbri> thanks zool, jim...

15:08:44 <DanC> F:jibberjim's tools make me wish my desktop were SVG-happier

15:08:44 <dc_rdfig> Added comment F14.

15:08:49 <geodanbri> I see an unusually colourful and written up doc about travel.

15:09:08 <geodanbri> oooh, you have itin to ical! I want that...

15:09:11 <DanC> (revision info is at the bottom; danbri seems to be there; anybody else there?)

15:09:21 <geodanbri> (pls look'n'ack...)

15:09:50 <JibberJim> date stamped 2003-04-15

15:10:14 <mattb> $Date: 2003/04/15 21:49:22 $

15:10:15 <Kake> Revision 1.29

15:10:26 <DanC> ok, so I'm kinda happy with the thesis I came up with: "The Semantic Web approach to application integration emphasizes data about real-world things like people, places, and events over document structure."

15:10:26 <geodanbri> OK, we're on the same page!

15:10:49 <geodanbri> I like that, it finds a niche apart from XML's domain.

15:11:07 <geodanbri> ...without suggesting they are entirely separate worlds... just separate emphasis

15:11:17 <DanC> now, re F and such, see the teletype stuff under "working with legacy data" and the N3 translation? under the first figure.

15:11:42 <geodanbri> yup

15:11:47 <DanC> (confirm that you see it by excerpting something, pls)

15:11:49 <PeteHaglich> DanC's thesis is consistent with our application approach

15:11:57 <geodanbri> [

15:11:58 <geodanbri> 07 APR 03 - MONDAY

15:11:58 <geodanbri> AIR AMERICAN AIRLINES FLT:3199 ECONOMY

15:11:58 <geodanbri> OPERATED BY AMERICAN AIRLINES

15:11:58 <geodanbri> LV KANSAS CITY INTL 940A EQP: MD-80

15:11:58 <geodanbri> DEPART: TERMINAL BUILDING B 01HR 36MIN

15:11:59 <dc_rdfig> Label DEPART not found.

15:12:00 <geodanbri> AR DALLAS FT WORTH 1116A NON-STOP

15:12:02 <geodanbri> ]

15:12:24 <DanC> I use k: for cyc terms

15:12:43 <geodanbri> Is that syntax per booking system? (of which there are hmm ~5 big ones?)

15:13:02 <DanC> note that I use k:startingDate to relate the flight to a calendar date, but I use t:departureTime to relate the event to a string in hh:mm format.

15:13:24 <DanC> yes, that syntax is per booking system. SABRE, I suspect, though I don't have any hard evidence.

15:13:51 <DanC> I've also converted aa.com's HTML itineraries to this vocab using XSLT. not sure where I put that.

15:14:29 <DanC> so step one is "get the data into RDF with perl and duct tape"

15:14:35 <geodanbri> so for other airlines (I want my itineraries, from W3C/ERCIM in France), we would just need to replace grokTravItin.pl ?

15:14:59 <DanC> yes, you'd need other perl/duct-tape

15:15:18 <geodanbri> can do. Hmm I think charles can-did actually, in Ruby. I'll dig it out...

15:15:20 <DanC> I think the build/buy section is most relevant to F

15:15:55 <JibberJim> F:You can also use my [http://jibbering.com/trip/genrdf.html|HTML javascript page to generate DanC's RDF], in appropriate browsers.

15:15:55 <dc_rdfig> Added comment F15.

15:16:22 <geodanbri> Yes, the build/buy dillema is a very familiar one to me...

15:16:28 <DanC> thoughts on the "mixing vocabularies in RDF is easy" para?

15:16:36 <geodanbri> the diagram, are those uml modeling conventions?

15:16:45 <DanC> yes, that's UML, as supported by dia

15:17:13 <DanC> note the class="noprint mechanics" blurb under the figure, including dia2owl.xsl

15:17:22 <geodanbri> the final sentence re migration being cheap, i'm not so sure about.

15:17:37 <geodanbri> Relatively cheap compared to DTDs, perhaps, as can be more incrementally done.

15:17:40 <libby> I find mixing very tricky in some circumstances, e.g tracking down where I can use certain properties and classes

15:17:45 <chrisg05> J:yes

15:17:45 <dc_rdfig> Added comment J2.

15:17:49 <JibberJim> even my javascript parser supports subPropertyOf danbri...

15:17:55 <zool> :)

15:18:09 <DanC> ok, so the "cheap" bit should be supported with evidence.

15:18:11 <geodanbri> It conflicts with anothe SW 'best practicy' thing folk often try to do, which is have their RDF documents also be processable w/ plain XML tools

15:18:26 <geodanbri> the XML tools are more easily suprised by the future ;)

15:18:54 <DanC> I think constraining RDF to work with plain XML tools is perhaps a necessary evil, but not a best practice. esp not with the emergence of stuff like rdf twig.

15:18:58 <geodanbri> Oh, I just remembered. All my flight booking data is in a W3C SQL store, complete with airport codes.

15:19:12 <libby> often mixing seems to be a bit like mapping ontologies, i.e. tricky and involving research

15:19:13 * geodanbri will find out who hacked on it...

15:19:16 <DanC> SQL: dbview to the rescue!

15:19:16 <dc_rdfig> Label SQL not found.

15:19:18 <libby> rdf twig?

15:20:01 <DanC> rdf twig: by Norm Walsh. sourceforge project, chumped a couple weeks ago. XSLT extension functions to get at RDF in natural triple/graph form.

15:20:20 <geodanbri> yes, more like pragmatic practice until the golden age of Semantic Web really kicks in (and old style XML withers away...)

15:20:32 <geodanbri> (like Capitalism was going to, for marxists?)

15:20:36 <DanC> ok, "Integration with mapping tools" is probably the section most relevant to today's meeting.

15:20:46 <libby> ooh

15:21:29 <geodanbri> Let's read that. We have 9 mins to cover mailing list item too...

15:21:34 * geodanbri reads, fast

15:21:40 <DanC> hmm... mike dean's daml airport data uses its own notion of lat/long. opportunity for convergence, there.

15:22:03 <JibberJim> yes DanC, but I offer up the same date using geo lat/lon.

15:22:08 <JibberJim> s/date/data/

15:22:08 <geodanbri> a link explaining 'log:semantics' would be very useful. I always found the name confusing...

15:22:11 <geodanbri> (or is it covered above)

15:22:17 <DanC> JibberJim, where?

15:22:20 <JibberJim> - http://jibbering.com/rdf/airports.1?LHR

15:22:25 * zool missed the 'mapping tools' link

15:22:40 <geodanbri> log:semantics: a relationship between a Web document and a representation of it as RDF?

15:22:43 <JibberJim> sorry - http://jibbering.com/foaf/airports.1?LHR

15:22:43 <libby> it's very cool danc

15:22:55 <DanC> yes, I should have added a link to the part of the tutorial that covers log:semantics: http://www.w3.org/2000/10/swap/doc/Reach

15:23:11 <JibberJim> oh no, I did update the old one too - either of those...

15:23:21 <DanC> http://jibbering.com/foaf/airports.1?LHR ... hmm... httpRange-14 again. (i.e. is that both an airport and a document?)

15:23:21 <dc_rdfig> K: http://jibbering.com/foaf/airports.1?LHR from DanC

15:23:26 <DanC> oops.

15:23:31 <geodanbri> Yup, I think the extra effort on the writeup well worthwhile. It's what Liddy Neville was saying at the W3C Tech Plenary... RDF folks often stop just short of the extra writeup that makes the point... (we take too much for granted)

15:24:12 <JibberJim> No DanC - I don't claim that document is an airport - there's an airport:Airport in there.

15:24:20 <DanC> hm... well, if anybody's got questions as they read along, fire away. otherwise, I suppose I should yield the remaining time.

15:24:28 <geodanbri> anybody got more quick questions for DanC? I'd like to make sure we get to mailing list thing...

15:24:34 <zool> i'm talking about 'collaborative mapping on the semantic web' next week at etcon, it would be good to tie some more of this stuff togethe, emphasise the fundamental interconnectedness of things

15:24:36 <DanC> ooh... well played, JibberJim

15:24:55 <geodanbri> OK, lets move on

15:25:00 <geodanbri> thanks DanC!

15:25:01 <zool> re http-Range14, i did have that included, but forgot about it in that demo

15:25:05 <geodanbri> ==== agenda J: new rdf/geo mailing list?

15:25:12 * DanC is conflicted... I just sent this tutorial in .pdf to the workshop organizers to be printed out for the WWW2003 attendees, but now I want to change it.

15:25:25 <libby> cool, zool, will you tell us what you're thinking of saying?

15:25:28 <geodanbri> DanC, you raised this. Wanna talk about the costs/benefits of 'yet another forum'?

15:25:33 <zool> J:yes, let's

15:25:34 <dc_rdfig> Added comment J3.

15:25:47 <Kake> J:Yes please

15:25:47 <dc_rdfig> Added comment J4.

15:26:07 <DanC> evidently the benefits are clear to quite a few of those present...

15:26:14 <JibberJim> I would be willing to subscribe to an rdf-geo list - I don't follow rdf-calendar or rdf-interest any more.

15:26:17 <zool> i know a lot of people who would be interested in the rdf-geo forum but perhaps a little worried about it being wholly RDF-centric

15:26:37 <zool> nothing to stop them subbing then buggering off though

15:26:44 <geodanbri> There's always a tension between having too many interested squeezed into a general forum, versus spreading folks too thinly. In the past I think we've gone too thin in a couple places: www-rdf-rules/www-rdf-logic aren't interestingly distinct to me.

15:26:49 <DanC> as to the cost: Somebody in the W3C team has to "own" it. Minimally, deal with bounces and the like. More usefully, set expectations about scope and direction and meet them.

15:27:38 <DanC> fear of "RDF" has been expressed around www-rdf-rules too.

15:27:41 <geodanbri> In this case, I see some merit, as RDF/geo is one of those areas where (a) people have long suspected killer apps (b) has no obvious topic-specific fora *that I'm aware of anyway* (c) seems to have at moment anyway a lot of folk's active attention.

15:27:42 <alberto> I think it would be good to fork from www-rdf-calendar to a more rdf-geo one (at least for the GEO/GIS stuff)

15:28:07 <geodanbri> The 'do we make it RDF specific or not' thing has also arisen before. With calendar, rules etc.

15:28:17 <zool> F:you can play with routes by talking to mudlondon@jabber.earth.li, visiting somewhere ('start' to see random list), saying 'en route', and following the instructions...

15:28:17 <dc_rdfig> Added comment F16.

15:28:20 <DanC> the ideal for me would be a forum for times/places/people/documents/money

15:28:39 <geodanbri> I think in www-rdf-calendar we have not been unfriendly to non-RDF representations, and in fact thru emphasis on testing etc have probably done some good for ical .ics format too...

15:28:44 <alberto> about adding geo-referenced info to RDF descriptions i.e. geo-annotations or something

15:28:51 <geodanbri> but I do value the scoped-ness that making it RDF-oriented brings.

15:28:57 * zool nods finding the geo-temporal intersection much more inteersting than either alone... but rdf-calendatr seems wrong place for that

15:29:22 <geodanbri> yep, rdf cal seems to have a niche in the ical-ish space of personal scheduling, PIM etc.

15:29:23 <libby> J:yeah!

15:29:24 <dc_rdfig> Added comment J5.

15:29:25 <alberto> i.e. starting from various vocabualries our there and try to get something "basic" which everybody can implement

15:30:03 <geodanbri> DanC, we do have the 'semantic-web@w3.org' list in reserve... do you think using that for vocab/app development, versus having domain-specific lists, could work better?

15:30:10 <DanC> at the DAML meeting, there was a slide on "reference knowlege". i.e. folks building a class scheduling ontology don't want to start from scratch. they'd like to have some basic times/places/people ontology to build on.

15:30:41 <alberto> I would like to hardcode my RDQL queries with some sensible geo:xyz properties :-)

15:30:44 <geodanbri> I think a general semweb vocab/apps collaboration list could prove useful, but would certainly be non-trivial to run.

15:31:13 <geodanbri> Every time we do some focussed thing, eg. Cal, Geo, people, we risk treading on toes of non-RDF XML initiatives in that space.

15:31:13 * DanC mulls over semantic-web@w3.org for time/place/documents/people/money

15:31:26 <geodanbri> This would give us a single toe-treading fora ;)

15:31:36 <libby> maybe your rss mailing list thing could help geodanbri, e.g. in makijng sure peopel see all of what's happening without having to subscribe to all

15:31:58 <DanC> getting the W3C mailing list search service to return RDF is high on my wish list.

15:32:02 <geodanbri> related: there is a proposal being worked on in Semantic Web coordination group to propose the continuation of the RDF Interest Group as a 'Semantic Web Interest Group'

15:32:22 <geodanbri> ...Guus Schreiber considering co-chairing it with me, and wants a focus on apps, onto development etc.

15:32:24 <JibberJim> DanC - in http://www.w3.org/2003/04dc-mia/itin-mia.rdf you don't say it's you doing the travelling anywhere do you? meaning that we can't do pathcross stuff.

15:32:48 <DanC> don't I? I meant to use k:travellers or some such...

15:32:49 <geodanbri> see http://www.w3.org/2001/sw/interest/charter.html for not-yet-merged old charter + Guus's ideas at the bottom

15:33:00 * geodanbri looks at the clock. We're over.

15:33:08 <geodanbri> DanC, suggested next steps re mailing list?

15:33:09 * zool still doesnt know how to represent groups of FOAFS in a clean way

15:33:10 <alberto> space + time vocs sound good :-)

15:33:27 <geodanbri> So we definitely note some interest in a new list.

15:33:29 <mattb> zool: groups of people who co-journeyed?

15:33:36 <alberto> zool, rdf:type="Collection" ?

15:33:38 <zool> mattb, yes

15:33:41 <DanC> <r:Description r:about="itin-mia.n3#thisTrip"> ... <k:passengers r:resource="http://www.w3.org/People/Connolly/home-smart#me"/>

15:33:42 <JibberJim> you could just have 5 people taking the journey - they don't need to be a group.

15:33:49 <geodanbri> Right now we're out of time, perhaps discussing another IRC meeting comes first?

15:34:02 <DanC> I think I like semantic-web@w3.org

15:34:04 <geodanbri> Libby, do you want to try the scheduling consensus thing?

15:34:05 <zool> alberto, hmm

15:34:15 <libby> eh?

15:34:17 <geodanbri> I like semantic-web@w3.org idea too.

15:34:24 <geodanbri> (arranging of next meeting...)

15:34:26 <geodanbri> J:

15:34:26 <dc_rdfig> blurb

15:34:27 <dc_rdfig> agenda J: new rdf/geo mailing list?

15:34:28 <dc_rdfig> (1:DanC) yes, please

15:34:29 <dc_rdfig> (2:chrisg05) yes

15:34:30 <dc_rdfig> (3:zool) yes, let's

15:34:31 <dc_rdfig> (4:Kake) Yes please

15:34:32 <dc_rdfig> (5:libby) yeah!

15:34:37 <alberto> e.g. <foaf:Group rdf:parseType="Collection><foaf:Person>....</></foaf:Group>

15:34:50 <mattb> for this use case, just some triples of the form [someperson] [partipicated in] [http://space.frot.org/show_route/test] ?

15:34:54 <DanC> re groups, consider cyc:groupMembers

15:34:56 <libby> who would like another meet on this topic?

15:35:00 <earle> J:definitely

15:35:00 <dc_rdfig> Added comment J6.

15:35:08 <Kake> Me, as long as it's OK to mostly just lurk.

15:35:09 <mattb> so they are grouped implicitly by their partipication in the journey

15:35:14 <DanC> and yes, cyc:groupMembers meshes well with parseType="Collection", i.e. first/rest

15:35:16 <alberto> but then you would get ordering - rdf:Alt or rdf:Bag is probably more general...

15:35:23 <DanC> J:consider semantic-web@w3.org

15:35:23 <dc_rdfig> Added comment J7.

15:35:30 <geodanbri> chrisg05, zool, Kake, Libby: would your wish for a geo-focussed list be met by a more general Semantic Web vocabulary collaboration forum, perhaps semantic-web@w3.org with a focus on vocab testing, apps etc.

15:35:34 * mattb likes to avoid containers where possible

15:35:42 * zool nods, so many potential approaches to grouping; orderingisnt smoethingi worry about; if they could all mesh like that could be interesting

15:35:44 <libby> geodanbri, mine could

15:35:46 * JibberJim echos mattb

15:35:48 <geodanbri> J:I would like a list, preferring semantic-web idea currently but an rdf-geo one too would be ok by me.

15:35:48 <dc_rdfig> Added comment J8.

15:35:54 * zool likes to avoid cyc vocab , prolly unresonably

15:35:58 <alberto> J: yes definitely interested

15:35:59 <dc_rdfig> Added comment J9.

15:36:01 <libby> jang: sw ok by me

15:36:02 <zool> geodanbri, sure

15:36:04 <DanC> not general vocab testing, danbri. times/places/people/documents/money

15:36:05 <geodanbri> I would like to meet again...

15:36:08 <libby> J:sw list ok by me

15:36:08 <dc_rdfig> Added comment J10.

15:36:10 <Kake> geodanbri: I was hoping mainly for something small and focussed while I am still getting the hang of this.

15:36:12 * mattb likes the idea of semweb discussion list, good to be usage- not impementation-focused

15:36:19 <geodanbri> those specifically? not documents or images?

15:36:20 <Kake> But obviously that's a selfish point of view :)

15:36:25 <JibberJim> somewhere which discussed people/places etc. would be very welcome.

15:36:29 <mattb> not scare people with "i must learn RDF before i can participate in discussion"

15:36:33 * geodanbri nods re small and focussed

15:36:35 <JibberJim> images would be nice...

15:36:36 <libby> we have kata and geodanbri interested in meeting again

15:36:43 <DanC> documents is in the list, and yes, images are documents (often) about/by people

15:36:45 <geodanbri> we could have subject conventions, eg. [geo]

15:37:05 <libby> s/kata/kake/

15:37:17 <JibberJim> this group exists, we can just subscribe?

15:37:18 <chrisg05> is semantic-web@w3.org to replace or generatlize rdf-interest?

15:37:27 <DanC> well, I don't want to scare anybody, but I do want folks to learn RDF.

15:37:28 <geodanbri> We have some action items recorded. I plan to be back here to report on my progress in two weeks time...

15:37:29 <alberto> zool, right collections are a bit of an artifact I agree but allow to "group" descriptions nicely I think - anyway, just an idea...

15:37:34 <libby> anyone else like to meet again? maybe a list would be enougfh at this stage - there's a lot going on

15:37:37 <geodanbri> Give or take an hour or so timezones blah...

15:37:41 * DanC looks for actions re J

15:37:45 * earle hacks out http://downlode.org/rdf/mapterms.xml - does this make any sense?

15:37:59 * earle is prepared for 'no, you're wrong in many ways'

15:38:01 <DanC> I'd be willing to post my time/place/people/documents/money call-to-arms to J

15:38:17 <geodanbri> DanC, do you fancy taking a 1st cut at proposing a scope for using semantic-web@w3.org as an RDF IG forum for ...

15:38:31 <geodanbri> yeah what you said. Plus documents and movies and MP3s.

15:38:32 <PeteHaglich> earle: Don't agree Countries are subclasses of Continents

15:38:42 <alberto> earle, is that related to TGN ?

15:38:50 <DanC> J:ACTION DanC: call for discussion of time/place/people/documents/money (incl images) to semantic-web@w3.org

15:38:50 <dc_rdfig> Added comment J11.

15:38:54 <earle> PeteHaglich: right, I was wondering about that myself

15:39:10 <geodanbri> Do we adjourn without a next meeting? I'm not hearing much oomph for scheduling another.

15:39:11 <JibberJim> I don't agree regions are subClasses of country either.

15:39:17 <earle> OK

15:39:18 <PeteHaglich> earle: Looks like blurring between physical containment and subclassing

15:39:24 <JibberJim> (ie southern Africa is a legit region.)

15:39:42 <geodanbri> <libby> we have kata and geodanbri interested in meeting again

15:39:45 <earle> PeteHaglich: this is a first go at writing this kind of thing for me, I expected it to be wrong in places

15:39:56 <geodanbri> kata =~ Kake?

15:39:57 <PeteHaglich> sure, not trying to be too critical

15:40:00 <JibberJim> I'll meet again, depending on places.

15:40:01 * earle smiles

15:40:03 * earle hmms.

15:40:09 <Kake> geodanbri: Yup, she meant me :)

15:40:15 * DanC thinks we'd better hurry up with the semweb IG proposal

15:40:34 <JibberJim> (I'm scheduled to be in Grahamstown on rough plans, there was free internet there last time I was there.

15:40:38 <Kake> So what do we do about J, subscribe to semantic-web@w3.org?

15:40:40 <geodanbri> OK, Libby, Kake, DanBri, ?Jim... how does 2 weeks today sound? same timeslot.

15:40:50 <libby> okey

15:40:55 * DanC wonders about the process implications of adding IRC to (email, phone) as chartered communications mechanisms.

15:40:58 <Kake> Probably good for me, depending on work chaos.

15:41:05 <geodanbri> J:Next steps: DanC to circulate a proposal, in #rdfig chump too.

15:41:05 <dc_rdfig> Added comment J12.

15:41:23 <chrisg05> J:An rdf/geo mailing list might attract more interest from the GIS community than general SW list

15:41:23 <dc_rdfig> Added comment J13.

15:41:31 <geodanbri> BLURB:Resolved: some of us to meet again here in 2 weeks, same timeslot.

15:41:31 <dc_rdfig> L: Resolved: some of us to meet again here in 2 weeks, same timeslot. from geodanbri

15:41:46 <DanC> geodanbri, I hope you'll forward my proposal/call-for-disucssion to rdf-interest

15:41:48 <geodanbri> Adjourned.

15:41:48 <libby> 'same timeslot'?

15:41:49 <geodanbri> ======

15:41:50 <earle> can a class have multiple subClassOfs?

15:41:59 <chrisg05> l:I'll attend

15:41:59 <geodanbri> Yes, I was hoping you could clarify the timeslot libby.

15:42:06 <geodanbri> s/l/L/

15:42:17 <jang> earle: yes

15:42:17 <geodanbri> DanC, I'd be happy to.

15:42:24 <Kake> 3pm is better than 4pm for me, since I'm generally heading for the Tube around 4:30

15:42:30 <geodanbri> And I agree re SW IG chartering...

15:43:01 <libby> last time was supposed to be 1400UTC. that's pretty early in US west coast, but ok here

15:43:05 <libby> in the uk

15:43:05 <geodanbri> DanC, do you have any specific thoughts on IRC as chartered comm mechanism?

15:43:10 <alberto> J:Next steps: some examples using geo ws84_pos vocabulary with geo:Point and geo:Polygon + labelling

15:43:10 <dc_rdfig> Added comment J14.

15:43:19 <geodanbri> ...maybe its accessibility unfriendly? i18n unfriendly?

15:43:24 * JibberJim isn't sure about geo:Point

15:43:29 <geodanbri> ...its australia unfriendly at least.

15:44:01 * libby proposes 1400UTC for next meet on 30th April 2003

15:44:07 <libby> yep re australia

15:44:24 <DanC> on IRC/process: all internet citizens can be expected to use email. (and HTTP, I guess). business folks know how to deal with the phone; IRC is a whole new ballgame, w.r.t. firewall politics etc...

15:44:38 <alberto> JabberJim, geo:Point is the basic class you have now in the ws84_pos voc - would like to play with some geo:Polygon for area coverage and stuff like that

15:44:38 <DanC> ... on the other hand, the phone is expensive, esp. for european folks.

15:44:52 <JibberJim> Yeah, I don't even like it there alberto.

15:45:15 <DanC> ... and IRC isn't as bad as the phone when english isn't your native toungue, as far as I can tell.

15:45:15 <JibberJim> I think an airport, or a person should be free to have a lat/lon without needing the geo:Point.

15:46:00 * JibberJim 's Phone Bill is excessive enough... I don't mind paying for IRC...

15:46:00 <geodanbri> Yes, firewalls and IRC can be a pain. I wonder if Jabber will help on that front, last time I strawpolled here, wasn't much usage of it...

15:46:01 <alberto> JabberJim, that would be what GML is for, but GML is not RDF-ied in the current version

15:46:09 <geodanbri> ...amongst these IRC users -- biased sample!

15:46:31 <DanC> I think jabber is more i18n happy, not to mention XML-happy. IRC hacking is kinda fun, but sort of a black art.

15:46:35 <chrisg05> btw an earlier version (1) of GML had an RDF version, but that was dropped

15:46:50 <alberto> chrisg05, yes :-(

15:46:54 <geodanbri> Are there any realtime audio tools (voice over IP) that folk know of?

15:47:16 <DanC> voip: my office phone is voip now. http://www.vonage.com/

15:47:23 * JibberJim struggles with IRC bandwidth...

15:47:26 <libby> here is 2 weeks today, 1400 UTC http://www.timeanddate.com/worldclock/fixedtime.html?day=30&month=4&year=2003&hour=14&min=0&sec=0&p1=0

15:47:29 <alberto> VocalTec used to have some IP phone stuff....not sure anymore

15:47:30 <geodanbri> thanks everyone for coming btw. I don't think I've tried chairing an irc chat before, it's fiddly and confusing! Hope to do better next time...

15:47:46 <DanC> it acts like a POTS phone, though. I dunno if you can call my office phone from the big-I internet.

15:47:52 <alberto> thanks danbri!

15:47:58 <libby> thanks geodanbri, you did a great job!

15:48:12 <alberto> bye everyone here - leaving now...

15:48:19 <DanC> chairing is a real service.

15:48:26 * earle fiddles with mapterms.xml some

15:48:34 <libby> bye alberto

15:48:35 <geodanbri> I'm wondering what "£#rdfig" will be like when we have 24x7 free audio comms

15:48:42 <geodanbri> cu alberto, thanks

15:48:59 <DanC> sometimes I think we can do these meetings without a chair, but sometimes it helps a lot to have *one* person authorized to answer 'where are we on the agenda?'

15:49:19 <geodanbri> even if they are somewhat making up the answer...

15:49:31 <DanC> of course.

15:49:44 <libby> :)

15:50:25 <libby> bit tiring even just keeping up

15:50:34 <mattb> yeah, it helps to have someone keeping an eye out

15:50:41 <libby> this time, never mind chairing

15:50:45 * mattb finds these w3 chats befuddling, especially while trying to do the dayjob

15:50:58 <DanC> if they weren't making up the answer, I'd be automating it, or at least noodling on how to automate it. but my considered opinion is that it's not feasible to automate chairing.

15:51:04 <earle> JibberJim: I altered the vocabulary a bit. What do you think?

15:51:44 <libby> danc, re scheduled chats and the wiki, I'd like to ahve one place to grab or generate the RDF versions of chats, for a bot I'm writing. that sound feasible?

15:52:08 <DanC> "the RDF version of chats" <- explain?

15:52:39 <libby> I had a chat w geodanbri about it, and he reckoned a table of times would be ok for people who don;t or can;t write the RDF version of a schedule i.e. start and end times for a chat on a given topic

15:53:06 <libby> and for peopel or groups who do want to generate an RDF file like that, then a single place to find all the links would be cool

15:53:22 <geodanbri> the 'where do i look to see what is lined up in #rdfig scheduled chats' problem

15:53:29 <geodanbri> s/the/a/, as ever

15:53:47 <libby> yeah

15:54:12 <JibberJim> I'm most unsure about region, it can apply to parts of continents (or areas which cross continents - e.g. Turkey) or regions of cities etc.

15:54:21 <JibberJim> Equally cities may be in more than one country.

15:54:28 <libby> sked, all

15:54:30 <DanC> this is perhaps feasible, but it looks quite hard. Are you sure it's valuable? who really has this problem that isn't sufficiently served by MailingLists?

15:54:30 <skedulike> time now is 2003-04-16 15:54 UTC. Upcoming meetings concerning #rdfig:

15:54:30 <skedulike> 2003-04-23 15:00UTC Calendar meeting #rdfig

15:54:43 <earle> JibberJim: that's true, and also difficult.

15:54:52 <geodanbri> and these things change over time... regions i mean.

15:55:00 <libby> well, I find myself trawling through mailing lists looking for whatever it was that I said

15:55:07 <geodanbri> libby, cool!

15:55:10 <libby> and what TZ it was on and so on

15:55:12 <PeteHaglich> earle: I think you need a containment property.

15:55:22 <geodanbri> its only as good as the data we feed it, of course

15:55:29 <DanC> libby, if you automate this, you realize we'll have to make you SW IG co-chair.

15:55:30 <libby> so I've got a problem....

15:55:40 <earle> PeteHaglich: yes. Pointer for newbie appreciated.

15:55:45 <earle> (any example will do...)

15:55:53 <libby> danc, aw you old charmer you

15:56:08 <DanC> DanC has changed the topic to: Semantic Web wish factory. http://rdfig.xmlhack.com/ (geo meeting winding down/adjourned)

15:56:13 <JibberJim> I think it might be easier to escape from the idea that they're subclasses of anything

15:56:16 <geodanbri> thx

15:56:24 <PeteHaglich> earle: I have a pragmatic geographic location ontology I can point you at, just saw a small error I need to correct.

15:56:28 <earle> JibberJim: sure, OK.

15:56:33 * geodanbri degeocentrises

15:56:34 <earle> PeteHaglich: ah, that sounds interesting.

15:56:35 <geodanbri> geodanbri is now known as danbri

15:56:39 <JibberJim> and then we can just have inRegion, inCountry etc. predicates which allows us to infer it.

15:56:42 <libby> so danc, danbri, shall I make a new wiki page, try it out?

15:56:58 <DanC> new wiki page? I don't think I'm following.

15:57:00 <libby> or maybe it coudl go in scheduledTopicChats

15:57:10 <danbri> it eats and parses data from text in the wiki?

15:57:22 <danbri> eg. in a Wiki table syntax...

15:57:28 <libby> maybe the wiki isn;t the right forum, I dunno - then chats coudl be scheuled all over the place

15:57:45 <libby> yep, that's what we discussed. my bot only finds RDF descriptions of chats at the moment

15:57:47 <PeteHaglich> earle: but my pragmatic location ontology is in DAML+OIL, not just rdf(s)

15:57:57 <danbri> I would like a 'here are all the RDFIG chats that danbri knows about' facility, and not do that via my personal calendar...

15:58:10 <libby> - but probbaly most people won;t be bothered to write ouot the rdf descriptions of chat times

15:58:11 <DanC> the most "real" thing about a schedule is the agreement formed between the people. editing the wiki page won't change what people have agreed to.

15:58:17 <dajobe> libby: re next meeting, didn't you agree 1400UTC, bot said 1500UTC ?

15:58:22 <earle> PeteHaglich: all the better, I think. I'm really just playing about.

15:58:37 <libby> dajobe, that was the calendar one I think, which was 1500

15:58:55 <dajobe> ok

15:59:08 <DanC> danbri, why do you expect that something besides your personal calendar would be a reliable source for info about chats you know about?

15:59:12 <libby> so what are you getting at danc? my feeling is that it's a problem with duplicating information all over the place, and I'd liek to avoid that

15:59:17 <Kake> Must concentrate on work now. Thanks for good ideas, bye all.

15:59:33 <libby> bye kake, thanks

15:59:43 <JibberJim> battery dead, and internet cafe closing, also leaving - goodbye.

15:59:51 <DanC> yes, it's a problem with duplicating information all over. And there's no easy solution to it. it's called the consensus problem in C.S. text books, I believe.

15:59:52 <libby> bye jim! see ya

16:00:00 <danbri> Cos it'd have a chance of being more reliably populated. I have gaps in my ical cal at moment. I could if there was an ical .ics feed from ScheduledChats have that show up in my personal cal as a remote feed.

16:00:02 <danbri> thanks Jim!

16:00:27 <libby> maybe ical would be good - easier to generate maybe than rdf cal. dunno

16:00:36 <libby> I guess it's not quite the issue

16:00:38 <PeteHaglich> earle: http://isx.com/~phaglic/ontologies/public_onts/isxlocusont.daml is my "pragmatic" DAML+OIL location ontology

16:00:40 * danbri still finds ical unreadable

16:00:49 <libby> the format doesn;t matter, what's authoritative matters

16:00:57 <DanC> but of j.radom edits ScheduledTopicChats, that doesn't mean danbri knows about it. unless there's some agreement/contract that danbri reviews ScheduledTopicChats periodically

16:01:14 <libby> true

16:01:20 <libby> but he probbaly would...

16:01:22 <DanC> you only get "authoritative" data by consent of the governed.

16:01:42 <danbri> yes, not sure publically editable and 'known to be known' go together

16:01:56 <libby> if all in one place (somehow) easireer to spot non-consented to chats

16:02:10 <DanC> i.e. w3c process says that meetings are scheduled 1 week in advance; anybody who joins a W3C WG agrees that 1 week notice from the chair is sufficient to reserve the group's time.

16:02:17 <danbri> maybe i don't need to know, just need to know how to find out where/when folks have consented to meet

16:02:20 <libby> danbri argued that a harvester woudl be the best thing

16:02:38 <danbri> the relevant data is yep inevitably going to be scattered

16:03:21 <earle> PeteHaglich: I just realised who you are! I was looking at your stuff this morning :)

16:03:24 <danbri> On the process front, the RDF IG currently has very few members per W3C process. Only invited experts and Member reps 'count' nowadays. Most of the people here today (??) probably weren't...

16:03:53 <libby> yeah, one issue is that tthe governed in irc is differnet from that on the mailing lists

16:04:22 <PeteHaglich> earle: I'm someone who should be lurking more during my first chat here.

16:04:24 <libby> for example

16:04:25 <DanC> the best solutions to "where's the truth?" problem I've seen are (a) meeting records available via HTTP that record decisions to meet in the future, and (2) meeting calls in email, schedule 1 week ahead, agenda 24hrs ahead.

16:04:26 <earle> PeteHaglich: I was looking at your public house ontology for possible use with Grubstreet (http://grault.net/grubstreet/)'s development effort

16:05:03 <earle> although I was a little worried about ""Use restricted to the U.S. Government, Horus Project members, and DAML PIs.""

16:05:07 <PeteHaglich> earle: Funny, that was a 15 minute "throwaway example" that I decided to post to DAML ontology list

16:05:11 <PeteHaglich> doh!

16:05:12 * earle smiles

16:05:23 <PeteHaglich> earle: I should remove that

16:05:30 * earle not surprised to run into you here...

16:05:52 <libby> could monitor the the w3c mailing lists, trying to spot meeting announcements

16:06:09 <libby> ...esp if they were in a standad format, much like the conference announcements case

16:06:27 <DanC> well, yes, you could, but why? i.e. you already monitor the lists you care about. why go to any particular effort?

16:06:51 <danbri> re monitoring the mailing lists, I do want to try doing that for more far-flung events, conferences and workshops.

16:07:00 <earle> PeteHaglich: one thing I noticed with the pubs example is that the public ontologies listed in it are now password protected on your site, so digging around I found them in the public_onts directory

16:07:03 <danbri> conferences are announced pretty inefficiently at moment...

16:07:15 <libby> I'd like to automate it, both types

16:07:28 <libby> danbri, was just digging out the stuff you wrote yesterday

16:07:47 <danbri> which stuff, i forgot already!

16:07:51 <PeteHaglich> earle: Thanks, I'll have to research that further and remove comments that aren't germane, like the one you found

16:07:52 <DanC> I repeat that this is a hard, though not completely intractible problem. (this = scalable calendaring and scheduling). What exactly do you want to automate?

16:07:55 <libby> announceOmatic

16:07:58 <earle> PeteHaglich: cool, thanks.

16:08:09 <collord> are there DAML ontologies for GML entities?

16:08:14 * earle gets rid of his map terms example and uses PeteHaglich's stuff instead

16:08:24 <danbri> I want to know the dates and locations of conferences and workshops announced to www-rdf-* lists.

16:08:26 <dajobe> how about cc:ing event emails to various lists to www-archive+rdfigevent@w3.org and then scraping that archive

16:08:52 <danbri> I'm happy to have us impose conventions, yes like that dave, and/or attachment or syntax rules

16:09:26 <earle> PeteHaglich: I may well use it in the rdf format I'm putting together for Grubstreet pages (test example at http://grault.net/grubstreet/hen_and_chickens.rdf)

16:09:27 <danbri> 1st step i think is to get a better feel for what's currently happening

16:09:58 <danbri> dajobe, is logger still searchable or it is grepping thru too much stuff?

16:10:09 <PeteHaglich> earle: OK, I'll clean up englishpubont so that it references the public locus ontology

16:10:18 <libby> I agree danc. I've been thinking about a conference calendar bot for a while, and had a chance to implement some of it

16:10:18 <dajobe> not searchable to any useful degree at the moment

16:10:50 <libby> I guess an irc channlel is a bit like a room to be booked

16:11:04 <earle> PeteHaglich: much obliged!

16:11:06 <DanC> conferences/workshops: one problem/issue is: getting people to fomalize the meeting announcements.

16:11:06 <danbri> if the bot is purely an aggregation point / ui to aggregations, exposing 1 or more datasets, that's v do-able.

16:11:36 <mattb> earle: typos in your example rdf: geo namespace URI ends with # not /, lat/long are lowercase ?

16:11:43 * mattb nitpicks

16:11:55 <DanC> you could have a bot subscribed to the list that notices words like "announce" and "conference" and such and mails a reply ala "that sounds like a conference announcement; please fill out the form at http:...."

16:11:58 <libby> yep, can be just that danbri. perhaps too ambitious to assume would find all relevant data. though no chance at all if the data is not machine readable in some way

16:12:09 <earle> mattb: I've just got rid of it, don't worry about it. :)

16:12:22 <PeteHaglich> Libby: one slight difference. Chances are that this channel will always be available. However meeting announcements are necessary to ensure that people are actually here.

16:12:24 <danbri> a couple of probs w/ formalised announcments: (i) until we make some better demos, its more work for no percieved gain (ii) conferences often extend deadlines, so we'd need to be ready to harvest revisions too

16:12:29 <DanC> or you could allow other folks on the mailing list to formalize the info and re-send it.

16:12:42 <earle> mattb: nitpicking appreciated tho

16:12:58 <danbri> re bot subscribed to list, I just got login to access lists.w3.org raw materials, get a feel for what's there.

16:13:03 <DanC> better demos: yes, we need to show value. Some eye candy (svg map of upcoming events...) might get things moving.

16:13:04 <earle> mattb: oh wait, you mean the grubstreet one? ta

16:13:09 * earle fixes

16:13:11 <danbri> I got sidetracked into RSS feeds...

16:13:35 <libby> PeteHaglich, yep at the moment, channel is usually free, but more and more chats are being scheduled and no simple say currently of checking there's no clash

16:13:43 <earle> mattb: fixed, thanks.

16:13:43 <danbri> RSS works in progress at http://www.w3.org/2003/04/list-synd/

16:13:56 <danbri> not sure if crontabbing worked...

16:14:00 <libby> yeah, svg map of conferences would be neat

16:14:13 <PeteHaglich> Libby: Thankfully, channels are free and you can always create a new one. Sort of like "Infinite Hotel"

16:14:41 <libby> I see what you mean yep. we havn;t been doing that, but definitely an option

16:14:55 <DanC> perhaps it's best to skip email altogether... constraining the way people send email seems very, very high cost. And the events I care about all get into *some* RSS feed.

16:15:02 <danbri> yes, that is one issue with our current setup. Our support tools (logger and dc_rdfig bots) can't be cheaply pointed at supplimental channels...

16:15:22 <libby> ah, true, I'd forgotten that

16:15:27 <libby> maybe solvable though

16:15:32 <dajobe> yes

16:15:34 * mattb recalls that being a todo on the chump list two years ago :/

16:15:41 * DanC yes which? forgotten what?

16:15:45 <danbri> Yes, encouraging the conference site to have RDF or ical or RSS+events linked from its homepage seems good way to go. Just announce the conf. And sniff each URL sent to the list maybe?

16:15:58 <libby> danc, yes re the channel tools

16:16:03 <danbri> yes/forgot: bot cost of new channels

16:16:24 <PeteHaglich> Unless you set up dedicated channels and use the bots to hold them open and op them.

16:16:27 <dajobe> re email format - I wasn't proposing any foramt. The action of cc:ing would be enough to just get the URIs of the messages that purported to announce the events

16:16:28 <DanC> I think the maximum bang/buck is getting time+place info into RSS feeds.

16:16:28 <libby> ical would be a start, and low cost

16:17:01 <danbri> libby, is the rdf/cal vocab in a state where you could show how RSS+events could be done in that style, using the rdf/ical translation?

16:17:04 <libby> danc, do you think the rdf icalendar format could be shoehorned in?

16:17:07 <libby> heh

16:17:15 <danbri> i do

16:17:22 <libby> I was just asking danc that ;)

16:17:26 <danbri> i have an ical aware tool on my laptop now

16:17:30 * danbri now pro-ical

16:17:30 <libby> on the face of it, I don;t see why not

16:17:35 <danbri> (except its syntax)

16:17:40 <DanC> I wonder if iCalendar/rdf is too constraining to be cost-effective in the RSS context. but maybe.

16:17:52 <libby> maybe only need a small part of it usually

16:18:05 <DanC> yes, but the small parts might be too constrained too.

16:18:08 <libby> location is important though (darn, full circle again)

16:18:30 <libby> searching the summaries or descriptions is a pain and not very accurate

16:18:39 <libby> subjects good too - maybe dc:subject?

16:18:46 <libby> foaf:interest?

16:19:10 <libby> how do you mean constraining danc?

16:19:17 <DanC> ok, let's give it a try, anyway... dajobe, you run a popular RDF-related RSS feed, right? danbri, libby, shall we aim at upcoming conferences (ExML, ISWC, WWW2003) or upcoming chats? (geo, calendar)?

16:19:20 <libby> i.e. not enough information?

16:19:51 <libby> do we have to choose? ;)

16:19:59 * danbri suggests conferences

16:20:17 <DanC> no, I mean that by giving the ical:dtstart of something, you're saying it's an ical:Vevent, which has exactly one organizer. but the thing you want to talk about has 0 organizers or lots or something.

16:20:23 <libby> yeah, probbaly most benefit

16:20:38 <libby> ah, ok, yep, coudl well be an issue

16:20:54 <libby> so I have this: http://www.w3.org/2001/sw/Europe/200211/swconferences/events1.html

16:21:06 <libby> it's a bit out of date

16:21:32 <libby> it's hand generated and can be converted to rss+events, though rss+events is rather crufty

16:21:32 <danbri> things it says are no longer true, or there are true things it could say that it doesn't?

16:21:47 <libby> it'd be simple to change the xslt to generate icalendar RDf though

16:21:58 <libby> true things it shoudl say that it doesn;t

16:22:22 <libby> mostly anyway

16:22:33 <DanC> another issue: how to maintain the data? by hand in CVS, with few authorized writers? I think you guys are aiming for something with more writers. I like eikon's "put the URI of your .rdf here and I'll aggregate it" style. (also the geourl style).

16:22:45 <dajobe> my rdf guide has tended not to have time-based stuff in it, since that means regular editing & pruning

16:23:05 <danbri> yes, more writers. aggregators should keep track of provenance. eg in http://www.w3.org/2001/sw/Europe/200211/swconferences/events1.html

16:23:15 <dajobe> (amusingly, this is where I started with metadata way way back)

16:23:15 <libby> yep, I like that too, if people can be persuaded to write RDF (which they may be able to be, esp if as xhtml like the w3c homepage etc)

16:23:21 <danbri> ...if the K-CAP info came from the website of K-CAP, that is worth noting.

16:23:28 <danbri> and 'last-refreshed' info too.

16:23:43 <DanC> ok, danbri/libby, are you willing to babysit a long-running crawler/aggregation service?

16:23:51 <danbri> similar issues that we get w/ FOAF descriptions of people. lots of writers, but try to note (pgp or whatever) authoritiative srcs

16:24:09 <libby> e.g.a ref to an email

16:24:27 <libby> I've got a crawler already, so no hardship

16:24:33 <DanC> where?

16:24:35 <danbri> long running as in Cwm eats up all my memory and I go have lots of cups of tea, versus service level agreement kind of long running... (months/years)

16:24:43 <DanC> months/years

16:25:06 <DanC> at least: a cron job, if not a process that stays running for months

16:25:16 <danbri> I aspire to, yes. I hosed rubyrdf by migrating to Ruby 1.8 and breaking the xml parser glue, else I'd say 'thats easy...' more glibly.

16:25:22 <danbri> yup

16:25:23 <libby> the foaf one crawls once a week

16:25:28 <libby> s/the/my/

16:25:29 <DanC> where, libby?

16:25:30 * mattb interested in making a pubsub crawler

16:25:33 <dajobe> I can do long term, if logger & chat logs isn't enough evidence

16:25:50 <libby> hm, it doesn;t really have a homepage danc

16:25:53 <mattb> crawling/aggregating seems a resource that could/should be shared

16:25:58 * danbri checks http://www.hackdiary.com/ for evidence of mattb's harvestering

16:26:06 * danbri nods re sharing

16:26:09 <libby> pubsub?

16:26:09 * mattb ahems, shuffles feet

16:26:28 <mattb> libby: publish+subscribe, tell me what you want and i send you messages when i have something that matches that description

16:26:41 <libby> ooh

16:26:44 <danbri> .google harvest gatherer broker

16:26:46 <datum> harvest gatherer broker: http://www.icewalkers.com/Linux/Software/514500/Harvest.html

16:26:48 <DanC> yes, your long-term credentials are well established, dajobe.

16:26:48 <mattb> a service that just tells you foaf-related triples when it finds them, or geographically-based ones, or something

16:26:56 <libby> nice

16:27:01 <libby> like a query?

16:27:06 <mattb> yeah, that sort of thing

16:27:13 <mattb> a query would be one example of subscription criteria

16:27:20 <mattb> new triples that match the query get pushed to you

16:27:32 <mattb> i don't like the idea of everyone running their own scutters, seems inefficient

16:27:44 <mattb> centralised service can cache, respect if-modified-since and so forth

16:27:53 <mattb> hence recent hackdiary snippet on that subject

16:27:57 <DanC> there should be a healthy marketplace of scutters. not too many, but not a monopoly either.

16:28:03 <danbri> yup. aggregators can do value adding things, check pgp sigs, compute subprops, inverses etc.

16:28:04 <libby> didn;t see that - chump it....

16:28:18 <libby> yeah, mines a bit simple :)

16:28:25 <DanC> libby, your weekly thingy... does it produce something I can see via http?

16:28:30 <danbri> I'll be happy the say folk start complaining about all the RDF harvesters pestering their servers!

16:28:39 <libby> I did this recently, was going to migrate to RDF/xml resultset format http://swordfish.rdfweb.org/discovery/2001/08/codepict/ntriplequery.html

16:28:40 <mattb> for preference, i think a scutter should run continuously and publish changes, not rescutter the entire web once a week

16:28:44 <mattb>http://www.hackdiary.com/archives/000028.html

16:28:44 <dc_rdfig> M: http://www.hackdiary.com/archives/000028.html from mattb

16:28:49 <mattb> M:|Conditional GETs in HTTP

16:28:50 <dc_rdfig> Titled item M.

16:28:56 <libby> danc, it runs the codepiction database

16:29:00 <mattb> M:how to make your crawler more polite to the servers it hits

16:29:00 <dc_rdfig> Added comment M1.

16:29:01 <libby> makes sense mattb

16:29:02 <dajobe> can somebody state the requirement again? A site that stores aggreagate triples, the urls of the content are submittable by anyone, the content is ???, what's the output?

16:29:05 <DanC> I gather the RSS world is seeing the "too many crawlers" stuff for real.

16:29:29 <DanC> "the codepiction database"... pointer?

16:29:29 <danbri> yup, that's true. but they're not very RDF-y... just a bit RDF-y ;)

16:29:29 <mattb> i don't see a crawler as doing the storage itself

16:29:31 <danbri>http://harvest.sourceforge.net/harvest/ChangeLog

16:29:32 <dc_rdfig> N: http://harvest.sourceforge.net/harvest/ChangeLog from danbri

16:29:32 <mattb> just publishing things it finds

16:29:41 <danbri> N:|ChangeLog from the Harvest project

16:29:41 <dc_rdfig> Titled item N.

16:29:42 <mattb> if someone wants to make a store by subscribing to those finds, they can

16:29:46 <danbri> N:It seems to still be alive!

16:29:46 <dc_rdfig> Added comment N1.

16:29:47 <libby> danc, this is maybe the most useful interface at the moment(but experimental) http://swordfish.rdfweb.org/discovery/2001/08/codepict/ntriplequery.html

16:29:59 <libby> the codepiction search is here

16:30:00 <libby>http://swordfish.rdfweb.org/discovery/2001/08/codepict/

16:30:00 <dc_rdfig> O: http://swordfish.rdfweb.org/discovery/2001/08/codepict/ from libby

16:30:02 <libby> oops

16:30:22 <danbri> N:See [http://harvest.sourceforge.net/|Harvest: A Distributed Search System] for more context.

16:30:22 <dc_rdfig> Added comment N2.

16:30:32 <libby> I woudln;t claim that it was a particularly good harvester or database or code, though it's been up for a year or two with occasional breaks

16:30:38 <DanC> ok, libby, but it's disclaimers like that which constrast shaply with what I'm after when I say "are you willing to babysit a long-running process?"

16:31:01 <libby> eikeon's looks pretty big, and he's keen on the calendar stuff

16:31:04 <DanC> you said "I already have a crawler, so it's no trouble." babysitting a service that people depend on *is* trouble.

16:31:04 * danbri decides to revisit his harvester code

16:31:27 * mattb has some JFDI time next week, will dust off and demo the pubsub harvester prototype written the week before last

16:31:29 <danbri> you snuck the 'that people depend on' bit in there...

16:31:38 <libby> it's no trouble to keep it going as it has been, which is reasonably realiable, but not 24/7

16:31:50 <DanC> didn't mean to be sneaky. "babysit" was intended to connote "customer service"

16:31:59 <libby> ah :)

16:32:30 * danbri read it more as 'put up with noisy troublesome kids' sorta thing :)

16:32:41 <libby> heheh

16:32:50 <dajobe> maybe we should have a code-off

16:32:54 <danbri> OK, I'm going to try being a customer of my own harvester.

16:32:56 * libby no good at that either ;)

16:33:05 <danbri> danbri, I hear you have an RDF harvester. Is it vapourware?

16:33:23 <danbri> danbri, thanks for asking. Why no, it has a wide and varied user community such as... your good self.

16:33:26 <mattb> heh

16:33:35 <mattb> right, i promise to next week publish some harvesting code

16:33:36 <DanC> LOL

16:33:41 <mattb> i've been putting it off because the smushing isn't quite right

16:33:44 <mattb> so i'll remove the smushing

16:33:54 <danbri> danbri, if so, where does one find it? I hear there may even be two versions out there somewhere...

16:33:55 <dajobe> is that the week you are moving house mattb?

16:33:57 <mattb> cos i'm very happy with the rest

16:33:58 <libby> cool mattb

16:34:00 <mattb> dajobe: no, that's this week

16:34:05 <danbri> mattb, that'd be great :)

16:34:16 <danbri> .google ruby rdf harvester

16:34:17 * mattb invites anyone witnessing to hold him to his promise :)

16:34:17 <datum> ruby rdf harvester: http://rdfweb.org/2002/09/ayf/intro.html

16:34:21 <DanC> I'm likely to write code to synchronize my personal calendar with a pub/sub event marketplace

16:34:28 * mattb just lurves pubsub

16:34:40 <dajobe> pub/sub event marketplace - great

16:34:42 <mattb> messaging infrastructures have got to be part of the future somehow

16:34:45 <mattb> i just haven't worked out how

16:34:57 <mattb> they are great in apps inside the firewall, controlled environments

16:35:01 <danbri> ok, i'll try make http://rdfweb.org/2002/09/ayf/ayf.rb run

16:35:02 <mattb> not sure about internets

16:35:19 <DanC> timbl's been talking about pub/sub marketplaces for a while, I think... a subscription includes a query pattern, right?

16:35:28 <mattb> yeah, some description of what you're after

16:35:50 <mattb> java messaging service uses SQL-where-clause style syntax

16:35:56 <mattb> which matches on the headers defined on the messages

16:36:05 <mattb> which themselves are rather like rfc822 headers

16:36:12 <libby> how do the pubsubs get pushed mattb?

16:36:23 <DanC> hm... are we boiling the ocean here, danbri? I don't see you getting an RDF description of semweb conferences this afternoon.

16:36:34 <mattb> libby: that's part of the Big Question, at least over the public internet

16:36:44 <DanC> for my own purposes, I list them in XHTML on my homepage and scrape them with XSLT into RDF.

16:36:51 <dajobe> :)

16:37:00 <mattb> libby: my prototype POSTs a notificaiton document to your URL of choice giving a description of new content and a URL you can retrieve it from if you choose

16:37:11 <libby> handy :)

16:37:19 * DanC remembers KnowNow, the wisen workshop...

16:37:47 <mattb> yeah, knownow sourceforged an early version of their apache-based http messaging infrastructure

16:37:59 <danbri> danbri@fireball:~/public_html/rdfig/latest$ grep -i 'Subject: cfp' *|wc

16:38:01 <mattb> the prototype has a java server and a perl client

16:38:07 <mattb> the perl client is only about 30 lines, maybe less

16:38:08 <danbri> finds 29 hits in www-rdf-interest archive

16:38:54 <danbri> I'll try to streamline the harvester bit, and will load up whatever RDF cal stuff you folks point me at...

16:39:08 <danbri> (it's almost all apt-get able, on a good day)

16:40:10 * danbri moves ruby1.8 out of the way

16:40:18 <danbri> Finished in 14.786149 seconds.

16:40:18 <danbri> 21 tests, 54 assertions, 0 failures, 0 errors

16:40:24 <danbri> thats one way to pass the tests :(

16:40:44 <libby> do the harvesters have to be very realiable? could you not just switch to another if one falls over?

16:41:11 <danbri> bit of both. several instances of same harvester w/ same dataset might have vulnerabilities

16:41:27 <danbri> ...eg. query planner might lock up under same conditions

16:42:46 <danbri> (like clone armies catching a cold...)

16:42:57 * libby investigates xslt and current list of sw conferences

16:42:58 <danbri> but yes, i would hope we can achieve some redundancy

16:43:05 <libby> heh, re clones ...

16:43:20 * DanC checks mail, finds some web services stuf...

16:43:44 <DanC> hey, this pub/sub stuff and crawlers is web services, right?

16:43:54 <mattb> heh, yup

16:43:57 <mattb> doesn't have to be SOAP to be WS

16:44:07 <mattb> HTTP GET is a web service

16:44:27 <DanC> amen. preach it, brother! 1/2 ;-)

16:44:35 <DanC> mattb, have we met?

16:44:49 <mattb> DanC: nope, but i think we have FOAFs in common

16:44:54 <mattb> i used to work with edd dumbill for example

16:44:59 <DanC> ah. indeed.

16:45:21 <mattb> i'm speaking at the developer day at www2003...

16:45:24 <DanC> do you keep a homepage? I keep one at http://www.w3.org/People/Connolly/

16:45:32 <DanC> ooh! a pathcrossing...

16:45:37 <mattb> i have http://www.hackdiary.com for scribbles

16:45:39 * danbri wishes foafbot were around still

16:45:52 <mattb> and http://www.picdiary.com for photos, with rdf-based metadata toys behind the scenes

16:45:58 <danbri> mattb, are you at XML Europe too? I was wondering about making a dash there...

16:46:01 <dajobe> danbri: arrgh

16:46:03 <mattb> and a ton of vapourware on IRC

16:46:12 <danbri> aargh re foafbot or xmle?

16:46:21 <dajobe> as it happens, both, but for different reasons

16:46:26 <mattb> danbri: yup, i'm speaking just after zool :)

16:46:40 * danbri should come and heckle

16:46:42 <DanC> mmm... yummy... Screenscraping HTML with TagSoup and XPath. I suppose I'll just have to get java-happy tools on my machine(s).

16:47:13 * DanC really wants an apt source for a good JDK

16:47:24 * danbri too

16:47:29 <mattb> DanC: you have the blackdown apt sources already?

16:47:31 <DanC> silent PC... more yummy stuff.

16:47:58 <mattb> [[ deb http://www.mirror.ac.uk/sites/ftp.blackdown.org/java-linux/debian unstable non-free main ]] enables apt-get install j2sdk1.4

16:48:02 <DanC> seems to be commented out:

16:48:03 <DanC> # blackdown java

16:48:04 <DanC> # deb ftp://metalab.unc.edu/pub/linux/devel/lang/java/blackdown.org/debian/ woody non-free

16:48:18 <mattb> change unstable to woody in my pastage if necessary

16:48:21 * danbri grins again at mattb & libby codepiction, http://www.picdiary.com/bloons/img_1806.jpg

16:48:55 <DanC> ooh! apt-get install j2sdk1.4 want it!

16:49:22 <mattb> gives you the blackdown sdk rather than the Official Sun Release

16:49:28 <mattb> but Works For Me

16:49:35 <DanC> logger_1, pointer?

16:49:35 <DanC> See http://ilrt.org/discovery/chatlogs/rdfig/2003-04-16#T16-49-35

16:49:35 <mattb> blackdown do a good job

16:50:29 * DanC runs sid; not sure why I had woody in my sources

16:51:03 <DanC> most of the JDKs I've installed for the last year or so crash my web browser.

16:51:27 <mattb> yeah, it's often worth removing the java plugin from /usr/lib/mozilla/plugins or wherever

16:51:35 * DanC is soooo happy the galeon/mozilla conflict has been resolved

16:51:49 <mattb> otoh the latest j2sdk1.4's plugin doesn't seem to have been compiled against the latest moz libs and so has unresolved symbols

16:51:54 <mattb> maybe it's the sid move to gcc3.2 causing it

16:51:56 * DanC has gone hunting in '/usr/lib/mozilla/plugins or wherever' too many times

16:53:06 <DanC> this is the sort of thing where the integration of debian really would pay; the integration you just don't get unless the software is really DSFG-free

16:54:05 * mattb heads for home

16:54:13 <mattb> good to chat...

16:54:19 <DanC> thanks for the jdk+debian clues, mattb

16:54:20 <collord> is there a way to describe relationships besides hierarchies within elements e.g. of spaces.owl?

16:54:21 <libby> bye mattb

16:54:36 <danbri> no joy with "apt-get install j2sdk1.4" and mattb's deb settings

16:54:38 <DanC> spaces.owl?

16:54:39 <danbri> 'Couldn't find package j2sdk1.4

16:54:54 * danbri goes back to rdf harvestering

16:55:04 <DanC> my 'apt-get install j2sdk1.4' seems to be winning. do you run sid?

16:55:13 <DanC> did you apt-get update?

16:55:15 <danbri> testing (is that 'woody')

16:55:17 * danbri did

16:55:24 <dajobe> sid=unstable

16:55:32 <dajobe> sarge=testing

16:55:34 <danbri> i s/unstable/testing/ initially, then s/unstable/woody/

16:55:36 <danbri> ah

16:55:43 <collord> sorry, i guess spaces.owl is specific to one project... http://space.frot.org/rdf/space.owl

16:56:01 <collord> in general though i think it's a valid question of RDF pertaining to GIS

16:56:16 <collord> how to describe a road connecting two points?

16:56:43 <collord> if there is only a single parent hierarchy allowed

16:56:45 <DanC> the cyc ontology has path stuff. that's where I'd start

16:56:50 <collord> ok

16:56:56 <DanC> single parent hierarchy?

16:57:12 <collord> well, can an element have more then one parent?

16:57:25 <collord> subClassOf multiple things

16:57:47 * collord needs to rtfm a bit more, sorry

16:58:05 <DanC> sigh... XML syntax getting in the way... yes, you can say { :C1 rdfs:subClassOf :C2. :C1 rdfs:subClassOf :C3}

16:58:08 * earle goes home, thanks for interesting discussion all

16:58:18 <earle> earle is now known as grault

16:58:28 <collord> ah, thanks

16:59:20 <DanC> mutiple inheritance ... deserves a page in the wiki, I suspect. along with stratification.

16:59:37 * DanC isn't quite inspired to do it now.

17:00:31 * DanC looks at space.owl thru N3 glasses... looks reasonable...

17:00:44 <DanC> :First-order_Administrative_Division a owl:Class;

17:00:44 <DanC> rdfs:label "a First-order Administrative Division";

17:00:44 <DanC> rdfs:subClassOf :Administrative_Area .

17:01:12 <DanC> Hypsographic???

17:01:52 <DanC> hmm... this looks like trouble: :Promontory(-ies)

17:02:41 <DanC> this one too :[SE1]

17:04:07 * collord looks at cyc.daml and has a stroke

17:04:44 <DanC> stroke?

17:04:48 <DanC> why?

17:05:05 <collord> it's huge

17:05:23 <DanC> cwm can eat it up in a few seconds. it can't be that big.

17:05:44 <DanC> though yes, I'd prefer more bite-sized chunks.

17:06:14 <DanC> "is cyc too big? Can I use just part of it without buying into all of it?" -- http://esw.w3.org/topic/BuildOrBuyTerms

17:06:44 <dajobe> I'm asking myself the same about owl ;)

17:06:59 <DanC> which part of owl seems too big?

17:07:10 <sandro> what does it mean to use a URI when you don't owl:import the ontology it comes from? Or when you do? No one knows!

17:07:34 <DanC> well, lots of people know. they just know different things ;-)

17:07:44 <sandro> same difference. :-)

17:07:46 <dajobe> DanC: the parts I don't need, which is most of it. That's fine for me.

17:08:02 <DanC> which parts do you need?

17:08:20 <dajobe> I was wondering what an "owl:Ontology" is too, but I'm still following the pointers, not got a question to ask so far

17:08:46 <dajobe> inverseFunctionalProperty if that's the right one

17:09:04 <sandro> I'm fairly suspicious of the lack of interoperable implementations of OWL to date.

17:09:34 <sandro> although maybe I just have looked hard enough (or programming enough!)

17:09:49 <sandro> s/have/haven't/

17:09:59 <DanC> you've seen implementations that don't interoperate? or you just don't see any OWL implementations at all?

17:10:12 <DanC> any RDF+prolog system is OWL-happy enough for me.

17:10:30 <DanC> (well, it needs some sort of loop prevention thingy)

17:10:38 <sandro> I don't see OWL implementations that do enough reading & writing OWL to show interoperating happening.

17:10:43 <collord> i'll ask a somewhat simple question... i'm writing a service for locating and describing wireless networking resources... what ontology or even technologies like OWL should i care about?

17:10:47 <dajobe> I tried 2 of the OWL validators, they gave different warnings/errors

17:10:54 <dajobe> maybe that's still because there are no final test cases

17:11:22 <DanC> I should perhaps have pushed harder to give a label to the horn-clause-happy subset of OWL. But perhaps we'll get enough feedback to re-open that idea.

17:11:38 <danbri> OK I'm moving the scutter .rb library into rubyrdf, its too expensive to try to do it separately. will keep obviously app-specific foafy stuff separate tho.

17:12:47 <sandro> That's what Benjamin Grosof called OWL Feather, right? And it's pretty much what his paper with Ian Horrocks calls "Description Logic Programs", I think. The center bit in his venn diagram. (on the slides I can't link to.)

17:13:35 <dajobe> if I can find the text that tells me what owl:import does, I may use that. I'm not sure when it happens

17:13:47 <DanC> hmm... well, I'm not to interested in the constraints imposed by DL stuff. I'd prefer to just deal with the whole horn part of the venn diagram [that I can't link to either]

17:14:18 <DanC> documentation of owl:imports is hard to find?

17:14:33 <DanC> or you've found it but it doesn't answer your questions?

17:14:48 * DanC thinks owl:imports isn't ready for prime time; is on record as such

17:15:24 <sandro> So you really want a Horn vocabulary, not an OWL one, DanC? I'm optimistic we'll get one out of Common Logic.

17:16:52 <DanC> well, eventually I want a horn-rule vocabulary. but even when I have that, I'll want owl:FunctionalProperty in my standard library. I won't want to write the rules over and over. so I still want the owl vocabulary.

17:18:16 <sandro> Understood. I used to think it was goofy to lump together vocabulary terms, and this reminds me of that feeling. There's little reason to consider owl:FunctionalProperty a part of OWL.

17:19:32 <DanC> huh? little reason to consider owl:FunctionalProperty a part of OWL?

17:20:29 <DanC> owl is an ontology-building vocabulary. the vocabulary for ontology-building is FunctionalProperty, inverse, etc. no?

17:22:08 <sandro> But where do you draw the lines about what's part of OWL or not? And why draw lines? Presumably for social reasons....

17:22:40 <DanC> ???

17:23:01 <DanC> for the same reason you draw lines around the C standard library, or SQL apis, or whatever.

17:23:05 <DanC> no?

17:23:32 <DanC> yes, they're sorta social reasons... but only in the sense that all of engineering is a social phenomenon.

17:26:57 <sandro> there's a difference between the placement of strftime in libc (which is like FunctionalProperty in OWL) and "for" in the C language. I think there may even be a bright line betweent the two, but I'm not seeing it at the moment.

17:27:45 <sandro> something about a modular part of the system or a non-modular part.

17:31:08 <sandro> And I think it's important for the documentation of strftime to be mostly self-contained, with occasional references to related documents, whereas the documentation for "for" can say pretty much "a keyword in C".

17:32:14 <sandro> It's totally reasonable to use FunctionProperty without anything else in OWL, and strftime without the rest of libc; it's not reasonable to use "for" apart from the rest of C.

17:53:32 <libby> is it mad, do you think, to have every item in an rss channel also being a vevent?

17:54:05 <DanC> it's clearly wrong to say that every item in every rss channel is a vevent

17:54:16 <DanC> but there may be some rss channels for which all the items are vevents

17:54:40 <DanC> clear use of quantifiers is essential to productive discussion ;-)

17:54:55 <libby> heh

17:55:00 <libby> I meant the latter

17:55:07 <libby> I have such a channel

17:55:45 <libby> but in any case it's probbaly still not right, as it woudl have to have a vcalendar in there too somewhere

17:56:15 <DanC> I recall some confusion about whether RSS items were documents, news events, news events about documents, news-event-documents about other documents, or something else.

17:56:56 <libby> sigh

17:57:32 <libby> also I don;t know if dates without times need timezones. that'll complexify it....

17:57:56 <DanC> the ical date value type involves no times nor timezones

17:58:07 <libby> thanks

17:58:10 <libby> good

17:59:08 <libby> I think this is wrong though: http://sern.ucalgary.ca/ksi/K-CAP/K-CAP2003/ isn;t a vevent, but a page about an event

18:00:22 <DanC> what's wrong?

18:00:51 <libby> my use of it as the identifier for a vevent

18:01:01 <DanC> oh; calling that a vevent would be wrong, yes.

18:01:25 <libby> <item rdf:about="http://sern.ucalgary.ca/ksi/K-CAP/K-CAP2003/">

18:01:25 <libby> <title>K-CAP 2003</title>

18:01:25 <libby> <description>K-CAP 2003, the Second Internetaional Conference on Knowledge Capt$

18:01:25 <libby> <link>http://sern.ucalgary.ca/ksi/K-CAP/K-CAP2003/</link>

18:01:26 <libby> <rdf:type rdf:resource="http://www.w3.org/2002/12/cal/ical#Vevent"/>

18:01:29 <libby> yeah

18:03:29 <DanC> { :item1 a rss:item; rss: description "K-CAP 2003, the Second Internetaional Conference..."; rss:link "http://sern.ucalgary.ca/ksi/K-CAP/K-CAP2003/"; foaf:topic [ a ical:Vevent; ...

18:04:04 * DanC heads off to a 1pCT appointment...

18:04:07 <libby> ah, I was wondering what could go there. is that how you are using foaf:topic?

18:04:10 <libby> thanks

18:04:14 <libby> see ya

18:05:53 <kao> Is there any way in N3 to do something like anonID for [] nodes?

18:05:55 <danbri> wohoo, fixed my crappy harvester!

18:05:57 <libby> foaf:topic makes sense to me "A topic of some page or document."

18:06:09 <danbri> it was dropping out after 1st 0 sized model it found

18:06:33 <danbri> a foaf:topic relates a document to a thing that document is about.

18:06:48 <libby> yep, cool

18:07:27 <DanC> kao, you can write _:nodeID

18:07:30 <DanC> is that what you mean?

18:07:40 <kao> not quite

18:07:45 <kao> i know that from n-triples

18:07:46 <danbri> 158 files this time before it crashed on bad xml... hmm needs robustifying

18:07:52 <danbri> on the + side, it's apt-getable

18:07:54 <kao> i thoiugh about the [] anons

18:07:57 * danbri -> movie

18:08:26 <kao> what i need is something like [:x ... ]

18:08:35 <kao> that gives a fixed it like _:x

18:08:44 <kao> so i can reference this later

18:09:18 <kao> i found this :- but dont know that that is used for... :-(

18:23:02 <torsti_> vittu

18:26:47 <torsti_> homoot

18:28:49 <nada> hi jen

18:33:18 <libby>http://www.w3.org/2001/sw/Europe/200211/swconferences/events1.html

18:33:20 <dc_rdfig> P: http://www.w3.org/2001/sw/Europe/200211/swconferences/events1.html from libby

18:33:34 <libby> P:|some Semantic Web related conferences

18:33:34 <dc_rdfig> Titled item P.

18:34:30 <libby> P:parsable into [RDF calendar|http://www.w3.org/2002/12/cal/] combined with RSS 1.0.

18:34:30 <dc_rdfig> Added comment P1.

18:34:43 <libby> P:experimental - comments welcome

18:34:43 <dc_rdfig> Added comment P2.

18:39:23 <libby> P1:parsable into [RDF calendar|http://www.w3.org/2002/12/cal/] combined with RSS 1.0 using [an xslt file|http://www.w3.org/2001/sw/Europe/200211/swconferences/home2rssical.xsl].

18:39:23 <dc_rdfig> Replaced comment P1.

18:41:21 <libby> O:with code examples - great!

18:41:21 <dc_rdfig> Added comment O1.

18:41:46 <torsti_> #maikkulan_mafia

18:49:49 <libby> O1=""

18:49:52 <libby> O1:""

18:49:52 <dc_rdfig> Deleted comment O1.

18:49:59 <libby> O2:""

18:49:59 <dc_rdfig> Comment O2 not found.

18:50:02 <libby> O1:""

18:50:02 <dc_rdfig> Deleted comment O1.

18:50:07 <torsti_> #maikkulan_mafia

18:50:08 <torsti_> #maikkulan_mafia

18:50:08 <torsti_> #maikkulan_mafia

18:50:14 <libby> O:|misschump, sorry

18:50:14 <dc_rdfig> Titled item O.

18:50:36 <libby> N:with code examples - great!

18:50:37 <dc_rdfig> Added comment N3.

18:51:06 <libby> M:next meet [2003-04-30, 1400 UTC|http://www.timeanddate.com/worldclock/fixedtime.html?day=30&month=4&year=2003&hour=14&min=0&sec=0&p1=0]

18:51:06 <dc_rdfig> Added comment M2.

18:51:27 <libby> M:with code examples - great!

18:51:29 <dc_rdfig> Added comment M3.

18:51:54 <libby> L:next meet [2003-04-30, 1400 UTC|http://www.timeanddate.com/worldclock/fixedtime.html?day=30&month=4&year=2003&hour=14&min=0&sec=0&p1=0]

18:51:54 <dc_rdfig> Added comment L1.

18:52:20 <libby> N3:""

18:52:20 <dc_rdfig> Deleted comment N3.

18:52:30 <torsti_> A2:replacement_text

18:52:30 <dc_rdfig> Comment A2 not found.

18:52:43 <libby> M2:""

18:52:43 <dc_rdfig> Deleted comment M2.

18:53:05 <libby> torsti_ I'm correcting some mistakes I made, please don;t arbitrarily change things

18:53:06 <libby> ta

18:58:47 <torsti_> libby ?

18:59:57 <torsti_> A:look

18:59:57 <dc_rdfig> Added comment A1.

19:00:18 <torsti_> A:look http://pants.heddley.co

19:00:18 <dc_rdfig> Added comment A2.

19:00:21 <torsti_> A:look http://pants.heddley.com

19:00:21 <dc_rdfig> Added comment A3.

19:25:27 <torsti_> A:keyword

19:25:27 <dc_rdfig> Added comment A1.

19:25:32 <torsti_> A:keyword A1

19:25:32 <dc_rdfig> Added comment A2.

19:25:35 <torsti_> A:keyword join

19:25:35 <dc_rdfig> Added comment A3.

19:27:01 <torsti_> dc_rdfig A3

19:32:51 <eDtivrsky> eDtivrsky is now known as Radical_Edward

19:34:56 <torsti_> kukkuu

19:35:00 <torsti_> ketää elossa++????

19:42:07 <mea_culpa> mea_culpa is now known as Radical_Edward

19:42:28 * torsti_ slaps dc_rdfig around a bit with a large trout

20:34:08 <unul> what is the difference between a complete and a consistent rule/strategy ?

20:39:57 <ericP> i assume it's something like the diff betwen soundsness and completeness in logic

20:40:22 <ericP> soundness : derive only what you should derive

20:40:36 <ericP> completeness : derive _all_ of what you should derive

20:42:09 <ericP> something that calculate transitive subProprtyOf of and subClassOf relations from rdfs could be sound and complete with respect to rdfs while only sound with respect to owl.

20:42:33 <ericP> if i'm using the terms correctly.

20:43:20 <unul> ericP: what?

20:45:59 <unul> ericP: are you saying that complete it's a non-optimal?

20:46:07 <unul> and consistent is ?

20:47:41 <ericP> not sure where either completeness or consistency would be non-optimal, but i'd say that consistency is more important

20:48:21 <ericP> if you are inconsistent, you will derive facts that the authors did not intend. ie false premises

20:48:35 <ericP> if you inconmplete, you will merely miss some facts.

20:48:41 <unul> ohhhhh

20:48:44 <torsti_> #maikkulan_mafia

20:48:46 <torsti_> join

20:49:20 <unul> so inconsistency it's like finding the wronh solution

20:49:24 <unul> *wrong

20:49:27 <ericP> indeed

20:49:52 <unul> and incomplete like not finding the solution at all

20:50:07 <unul> but at least when you find the solution it's the right one

20:50:09 <ericP> also with rdf, you can never know you are complete with respect to all logics the author had in mind

20:50:14 <ericP> yup

20:50:23 <unul> rdf ?

20:50:44 <unul> what's that ?

20:50:45 <ericP> rdf doesn't have a system for listing all of the required logics.

20:51:00 * ericP tries to think of a funny acronym for RDF

20:51:15 <dajobe> reasoning done funny

20:51:16 <unul> i've joined this chanel only because it was on #ai's topic

20:51:21 <ericP> i like

20:51:23 <dajobe> read the topic then

20:51:25 <dajobe> :)

20:52:08 <unul> :)

20:55:43 <unul> ericP: thanks for the info

20:55:49 <unul> bye everyone

22:09:52 <torsti_> miten kellää ei oo oppeja

22:50:59 <torsti_> kierans

22:51:13 <kierans> Hi

22:51:44 <torsti_> hi

22:52:53 <kierans> So many people here, never anything being talked about....

22:52:56 <torsti_> where is all ops in the channel

22:53:17 <kierans> *shrug*

22:53:26 <torsti_> ?

23:05:24 <torsti_> A:link

23:05:24 <dc_rdfig> Added comment A1.

23:06:21 <torsti_> A:link torsti_

23:06:22 <dc_rdfig> Added comment A2.

23:06:39 <torsti_> A:join #Maikkulan_mafia

23:06:39 <dc_rdfig> Added comment A3.

23:06:50 <torsti_> A:op

23:06:51 <dc_rdfig> Added comment A4.

23:07:24 <torsti_> A:super owner torsti_

23:07:24 <dc_rdfig> Added comment A5.

23:07:26 <deltab> what are you doing?

23:08:05 <torsti_> runxs