Semantic Web Interest Group IRC Chat Logs for 2003-02-12

00:41:44 <SethR>http://lists.w3.org/Archives/Public/www-rdf-comments/2003JanMar/0229.html

00:41:45 <dc_rdfig> A: http://lists.w3.org/Archives/Public/www-rdf-comments/2003JanMar/0229.html from SethR

00:42:16 <SethR> A: Pat Hays explains de-re and de-dicto in relation to RDF reification

00:42:16 <dc_rdfig> Added comment A1.

01:07:09 <danbri> A:Aka why I never use RDF's reification vocab.

01:07:09 <dc_rdfig> Added comment A2.

01:46:42 <SethR> A: [http://lists.w3.org/Archives/Public/www-rdf-comments/2003JanMar/0231.html |where Seth muddels that the WG gofed]

01:46:42 <dc_rdfig> Added comment A3.

01:47:44 <burtonator> danbri: in answer to you NewsMonster question for the other day - yes... it will reload the RDF database if it changes on disk.

01:48:36 <SethR> A3: [http://lists.w3.org/Archives/Public/www-rdf-comments/2003JanMar/0231.html|where Seth muddels that the WG goofed]

01:48:36 <dc_rdfig> Replaced comment A3.

01:49:39 <sbp`-> sbp`- is now known as sbp`

01:49:47 <SethR> yo sean

02:05:00 <LTjake> hello.

03:01:57 <dmiles> any ODBC based RDF things?

03:02:14 <dmiles> like an ODBC driver?

03:02:24 <dmiles> that proxies out SQL queries?>

03:12:31 <Canoedist> hello

03:13:32 <Canoedist> is anybody there?

03:24:49 <jordan> jordan is now known as uhh-moving

07:37:09 <arnarl> hi

07:51:33 * Mutiny is gone... zzzzzzz <since(02:51am)> <(lon pon)syntax>

09:44:26 <burtonator> Are all RDF vocabs going to use the application/rdf+xml media type?

09:44:31 <burtonator> isn't this a bad idea?

09:44:41 <burtonator> I am forced to download the file BEFORE I can determine what it includes

09:59:27 <mortenf_> burton, there's a title attribute?

09:59:52 <mortenf_> also, there's not really a concept of doctype in RDF.

10:07:25 <JibberJim> dmiles, were you looking for an ODBC which accepted squish or something, and mapped into whatever SQL backend?

10:07:37 <JibberJim> interesting idea... but no I don't have one...

10:25:40 <danbri> libby had something like this for Java/JDBC

10:26:34 <danbri> ie http://ilrt.org/discovery/2001/02/squish/ "Each application uses Java Squish accessed through the JDBC API from a JSP or Java Servlet, using either an in-memory or a Postgres SQL backend as the RDF database.

10:26:36 <danbri> "

10:36:47 <burtonator> mortenf_: the title applications is a kludge

10:37:01 <burtonator> what we need is a media type like we have for RSS 1.0

10:37:05 <burtonator> application/foaf+xml

10:37:24 <danbri> <burtonator> Are all RDF vocabs going to use the application/rdf+xml media type?

10:37:24 <danbri> <burtonator> isn't this a bad idea?

10:37:24 <danbri> <burtonator> I am forced to download the file BEFORE I can determine what it includes

10:37:31 <danbri> the granularity is different w/ RDF

10:37:38 <danbri> often an RDF file will have bits of everything

10:37:48 <danbri> bit of DC, bit of RSS, bit of FOAF, bit of Wordnet...

10:38:02 <burtonator> danbri: but generally it is application specific

10:38:20 <burtonator> I mean why don't I put all my RDF in one 700G index.rdf file on my main website

10:38:22 <burtonator> ?

10:38:32 <burtonator> obviously that would be a valid RDF graph but it isn't a good idea

10:38:57 <danbri> yeah, i'm not denying there's an issue...

10:39:02 <burtonator> ok

10:39:11 <burtonator> Mozilla is using URNs for the datasource name

10:39:11 <danbri> ...just mime typing isn't ideally suited

10:39:21 <danbri> yeah, that's kinda retro

10:39:25 <burtonator> then one would have ot give it a title

10:39:34 <burtonator> but in order to do that we should use a URN/URI

10:39:39 <burtonator> at which point we should use a mime type.

10:40:14 * mortenf_ is back

10:40:56 <mortenf_> good points, but it seems a lot of hassle to register mime types for every possible rdf application / vocabulary

10:41:25 <dajobe> yes

10:41:33 <dajobe> and people don't register generally at all

10:42:57 <burtonator> right... so the solution seems to be to complain I guess ;)

10:43:23 <mortenf_> yeah, that always works on one level! ;)

10:43:44 * burtonator is tired... but.. isn't there a vendor or x- mime type we can use

10:43:46 <burtonator> I can't remember...

10:44:53 <mortenf_> I suppose there is, but then there would be no real value, since it's not registered, meaning id's could overlap...

10:45:34 <burtonator> You mean the fact that someone else could use application/x-foaf+xml ?

10:45:41 <burtonator> I think the chance of that is essentially zero

10:45:59 * burtonator thinks the concept or a centralized registry is sooo ICANN

10:46:03 <mortenf_> yeah, but i'm thinking beyond foaf.

10:46:34 <dajobe> rdf core's registering a type so you don't have to ;)

10:46:34 <burtonator> I don't understand...

10:46:46 <burtonator> danbri: what was the software you created that graph with? It wasn't isaviz?

10:47:02 <burtonator> dajobe: what?

10:47:16 * burtonator wants to update the mod_link spec to include FOAF

10:47:20 <dajobe>http://www.ietf.org/internet-drafts/draft-swartz-rdfcore-rdfxml-mediatype-01

10:47:20 <dc_rdfig> B: http://www.ietf.org/internet-drafts/draft-swartz-rdfcore-rdfxml-mediatype-01 from dajobe

10:47:31 <dajobe> B:|application/rdf+xml Media Type Registration

10:47:31 <danbri> rdfauthor

10:47:31 <dc_rdfig> Titled item B.

10:47:45 <mortenf_> burton, the point being that other vocabularies face the same "problem".

10:48:13 <burtonator> dajobe: this is the problem... application/rdf+xml is too generic

10:48:17 <burtonator> might as well be text/xml

10:48:31 <burtonator> or application/semantic-web

10:48:33 <burtonator> means nothing

10:48:41 <burtonator> waste of bits ;)

10:48:42 <dajobe> this argument is circular

10:48:48 <dajobe> at what point do you want to know?

10:48:49 <danbri> re "burtonator thinks the concept or a centralized registry is sooo ICANN"

10:48:53 <dajobe> before the http fetch?

10:48:56 <dajobe> at the fetch?

10:48:57 <dajobe> after?

10:48:58 <danbri> doesn't adding it to mod_link create the same centralism?

10:49:00 <dajobe> before xml parsing?

10:49:00 <burtonator> before

10:49:03 <dajobe> after rdf/xml parsing?

10:49:17 <dajobe> so you are in the realm of url heuristics

10:49:23 <dajobe> just look for '.foaf' in the url and be done

10:49:26 <burtonator> danbri: it is just like your <link> proposal

10:49:37 <burtonator> at the time I am given the rdf:resource

10:49:49 <burtonator> .foaf doesn't mean anything

10:49:54 <burtonator> what if it is produced from a CGI?

10:49:56 <dajobe> why not look for a triple <new resource> rdf:type <something>?

10:50:00 <burtonator> with no .foaf extension

10:50:13 <danbri> yeah, filename extensions are a distraction

10:50:23 <danbri> people may as well use foaf.rdf but it shouldn't matter

10:50:26 <burtonator> dajobe: that is what I am saying we need

10:50:35 <dajobe> well without precognition, you won't know what you will get back until you try

10:50:50 <burtonator> or foaf.png and then serve a foaf RDF file just to piss everyone off

10:51:14 <dajobe> (I'm adding png w/ embedded rdf parsing support soon)

12:11:09 <libby> hm, how do I find oiut what characters can go in a nodeID for a blank node identifier?

12:13:15 <dajobe> ahem

12:14:03 <dajobe> see http://www.w3.org/TR/2003/WD-rdf-testcases-20030123/#nodeID

12:14:05 <dajobe> then name

12:14:27 <dajobe> but that's just a suggestion for an easy life

12:15:21 <libby> ooh, cheers dave

13:27:59 <danbri> dave, I presume it is ok for the same name string to be used in both a nodeID and an ID, so long as on different XML elements... right?

13:28:09 <danbri> (annoying/stupid maybe, but not invalid?)

13:29:39 <dajobe> allowed & stupid - yes, yes

15:18:58 <libby>http://lists.w3.org/Archives/Public/www-rdf-calendar/2003Feb/0007.html

15:18:59 <dc_rdfig> C: http://lists.w3.org/Archives/Public/www-rdf-calendar/2003Feb/0007.html from libby

15:19:19 <libby> C:|IRC calendar meeting agenda Wednesday Feb 12th, 17:00 UTC

15:19:19 <dc_rdfig> Titled item C.

15:27:26 <libby> C:[RDF calendar workspace|http://www.w3.org/2002/12/cal/]

15:27:26 <dc_rdfig> Added comment C1.

16:02:53 <_joshua> good morning

16:26:53 <danbri> .google ben guardian foaf friendster

16:26:54 <datum> ben guardian foaf friendster : http://www.benhammersley.com/archives/003468.html

16:32:26 <LotR> hola

16:40:36 <libby> afternoon lotR

16:44:34 * LotR looks at the agenda and sees very little of interest

16:46:05 <libby> hey andyS

16:46:23 <libby> well you dont have to come then LotR :)

16:46:40 * LotR nods

16:46:54 <AndyS> Hi Libby

16:51:47 <danbri> hi all

16:51:54 <DarkPenguin> Hi Dan

16:52:46 <libby> I didn;t mean it LoTR, stay ;)

16:52:52 <libby> recurrence is on the agenda...

16:54:11 <DanCon> hmm... today's calendar chat didn't make it into my WearableGizmo

16:55:37 <danbri> hi amy!

16:55:44 <amy> heya Danbri :)

16:56:18 <danbri> nice to see you here

16:56:34 <amy> nice to be here

16:56:42 <libby> hi amy, dancon

16:56:48 <amy> hi Libby!

16:56:57 <libby> why';s that dancon? somethuign wrong w the icalendar?

16:59:32 <DanCon> no, just oversight or something.

17:00:00 <DanCon> i.e. I don't have the "any appointment I agree to in IRC goes in my calendar" agent finished yet. 1/2 ;-)

17:00:24 <libby> heh, no?

17:00:32 <libby> ok, guess we better start

17:00:39 <DanCon> I might have left before the time was nailed down or something too; I think I was getting on a plane.

17:00:48 <libby> ----- rdf calendar meeting ------

17:00:56 <libby> yeah, you were off somehwere

17:01:19 <libby> so - anyone like to add anythoing to the agenda?

17:01:24 <DanCon> C:spiffy agenda

17:01:25 <dc_rdfig> Added comment C2.

17:01:28 <libby> :)

17:01:36 <danbri> C:

17:01:36 <dc_rdfig>http://lists.w3.org/Archives/Public/www-rdf-calendar/2003Feb/0007.html

17:01:37 <dc_rdfig> IRC calendar meeting agenda Wednesday Feb 12th, 17:00 UTC

17:01:38 <dc_rdfig> (1:libby) [RDF calendar workspace|http://www.w3.org/2002/12/cal/]

17:01:39 <dc_rdfig> (2:DanCon) spiffy agenda

17:02:03 <libby> hm , mayeb I shoudl have copied it into irc

17:02:38 * danbri waves to ericP

17:02:42 <libby> I had lots of fun makeing a java version of ical2rdf. nice to do a bit of programming

17:02:45 <libby> hey ericP

17:02:49 <ericP> heya, folks

17:02:59 <timbl-away> timbl-away is now known as timbl

17:03:02 <DanCon> DanCon has changed the topic to: RDF Calendar chat. 1700Z Wed 12Feb http://rdfig.xmlhack.com/

17:03:20 <ericP> i forget, do { :a hasColor "green". :b hasColor "green". } have the same object?

17:03:32 * DanCon is tempted to install libby's java ical2rdf

17:03:42 <DanCon> yes, ericp.

17:03:50 <libby> the bulk of it isn;t my stuff, which means it works :)

17:04:12 * gk apologizes for being late, hasn't had time to check out final agenda

17:04:32 * danbri sidetracked into trying to find out why mozilla suddenly stopped working on his desktop

17:04:32 <ericP> dancon, tx

17:04:43 <DanCon> hmm... how about checking the java one into http://www.w3.org/2002/12/cal/ somewhere, libby?

17:05:12 <libby> for the 4 testcases I could find it works ok:http://www.w3.org/2002/12/cal/test/

17:05:43 <DanCon> hmm... we could use a test harness... something so folks can do make test

17:05:45 <libby> I was wondering about that, but it depends on GPLed code that belongs to someone else. would it make sense just to put the ical2rdf.java file in cvs?

17:06:12 * timbl wonders if we have started and are on agenda 1.1

17:06:17 <libby> hey gk

17:06:23 <DanCon> if the code is GPLed, then we're licensed to republish it

17:06:37 * timbl ... or whether we are still on intros and stuff

17:06:38 * DanCon notes timbl's point of order, sorry...

17:07:02 * danbri wasn't sure if we publish GPL'd code on www.w3.org (can discuss that later)

17:07:16 <danbri> [[

17:07:17 <danbri> 1. iCalendar2RDF issues

17:07:17 <danbri> time: 15-20 minutes

17:07:18 <danbri> ]]

17:07:20 <libby> yeah, would like to clarify that later

17:07:21 <danbri> we're on that still?

17:07:33 <libby> yeah, not started it really

17:07:36 <danbri> 1.1 vcal thing being something that was thrown up by your implementation work?

17:07:52 <libby> sortof

17:07:52 <DanCon> agenda + OK to check GPL'd code in under http://www.w3.org/2002/12/cal/ ?

17:08:07 <DanCon> oops... no zakim bot to queue agenda requests. :-{

17:08:07 <libby> yeah, please

17:08:12 <libby> add to C?

17:08:20 * gk still waiting for email to download, so can see agenda

17:08:29 <danbri> So you have java code that eats ical and emits RDF in the same way that DanC's Perl code works.

17:08:37 <DanCon> C:agenda + OK to check GPL'd code in under http://www.w3.org/2002/12/cal/ ?

17:08:37 <dc_rdfig> Added comment C3.

17:08:43 <libby> gk, we are on [[

17:08:59 <danbri> And you have it working ok on 4 files, hence the vcalendar-as-container thing...

17:09:04 <danbri> and prospect for more test data?

17:09:16 <libby> that's right

17:09:33 <danbri> this is sounding healthy. Two automatic ical to RDF gadgets behind more or less the same...

17:09:55 <gk> Hi Libby. This one? 1.1 The relationship between the Vcalendar and components.

17:10:08 <libby> testdata,I've asked Ol, and the icalendar guys. Ol is prparing stuff; icalendar calsch guys nearly siad yes but didn;t lie the idea of making it public

17:10:23 <libby> that's rioght gk, sorry, I trued to copy and paste that, didn;t work

17:10:38 <DanCon> hmm... test case to look at for 1.1 The relationship between the Vcalendar and components ?

17:10:53 <libby> so, 1.1 came up partly because it'ss actually a pain when paring vcalendar data *not* to enclose compoments in it

17:11:06 <libby> and because it's usefult o be able to seaparate calendars

17:11:24 <gk> Tnx. We're currently using ical:hasEvent as relationship property, per http://www.ninebynine.org/RDFNotes/Calendaring/iCal-proposal.html

17:11:41 <timbl> Choice: either [] a Calendar; cal:event [ :date ..... or [] a Calendar; cal:includes [ a Event; :date ....

17:11:45 <libby> danCon, sorry, I didn;t check in the end whether you are allowed multiple vcalendars bper dopcument in RFC2445

17:11:48 <danbri>http://iCalShare.com/

17:11:49 <dc_rdfig> D: http://iCalShare.com/ from danbri

17:11:56 <danbri> D:|icalshare

17:11:57 <dc_rdfig> Titled item D.

17:12:01 <timbl> "has" is yucky.

17:12:01 <DanCon> ah; here's a test case I found relevant to 1.1rel: http://www.gxsnmp.org/CVS/evolution/libical/test-data/2445.ics

17:12:01 <danbri> D:Source of more test data?

17:12:01 <dc_rdfig> Added comment D1.

17:12:09 <libby> datum: ooh, great source of testdata

17:12:14 <danbri> D:Probably mostly generated by MacOS X tools?

17:12:15 <dc_rdfig> Added comment D2.

17:12:22 * DanCon tends to doorbell...

17:12:27 <timbl> Noun phrases are better. You can add an optional "has" in N3 is you need to.

17:12:41 <timbl> I suggest replace hasEvent with event.

17:12:43 <danbri> yes, I prefer not to use 'has' too.

17:12:45 <gk> I thought "has" might not be liked. I don't care, but was trying to make the point clear.

17:13:02 <libby> oh great dancon

17:13:05 * DanCon back

17:13:16 <libby> great testcase for that

17:13:21 <danbri> 'has' is sorta implicit. 'isFooOf' is a worse convention, as it reads backwards.

17:13:23 <timbl> I think the community has gone with the nouns in by the far the iggest way.

17:13:23 <gk> So suggestion is: ical:event a rdf:Property; rdfs:range ical:Vevent . ?

17:13:28 <timbl> biggest

17:13:46 <danbri>http://www.gxsnmp.org/CVS/evolution/libical/test-data/2445.ics

17:13:46 <dc_rdfig> E: http://www.gxsnmp.org/CVS/evolution/libical/test-data/2445.ics from danbri

17:13:46 <timbl> Basically, yes gk.

17:14:03 <danbri> E:|ical test relating to 1.1rel issue in today's agenda.

17:14:03 <dc_rdfig> Titled item E.

17:14:09 <libby> I sorta agree with DanCon that not totally clear what a calendar does. just a grouping mechanism

17:14:09 <timbl> We could discuss whether the "V" will be a curious annachornism in a decade.

17:14:16 <gk> I'm entirely happy with that.

17:14:41 <timbl> grouping mechansim: if so, then a generic grouping vocab should be used or made.

17:14:46 <danbri> yes, v- and e- and i- are all a bit late nineties...

17:14:51 <libby> I think if we are replicating icalendar in rdf as close as possible, then it makes sense to have that grouping mechanism

17:14:58 <timbl> if not a grouping mechansim, then does it has semantics?

17:15:25 <DanCon> PROPOSED(1.1q1): to use { ical:event a rdf:Property; rdfs:range ical:Vevent . } to relate calendars to events, e.g. in the case of multiple calendars in one file.

17:15:53 <danbri> I've run into this before w/ converting XML stuff to RDF, and often end up thinking of entire RDF docs as my grouping mechanism. The granularity differs though...

17:15:54 <timbl> I agree that it mkes sense to be able to suck in 3 files and have the calendars intact.

17:15:56 <libby> hey ol

17:15:56 <timbl> SECONDED

17:16:10 <DanCon> hmm... what to use to relate calendars to other components, e.g. journal entries?

17:16:13 <Ol> hi

17:16:14 <libby> what about things like vtodos etc?

17:16:15 <gk> Libby, our recurring schedule stuff motivates multiple Vevents related to a Vcalendar object ...

17:16:16 <LotR> hey Ol

17:16:23 * Ol late, sorry

17:16:33 <libby> glad you could make it

17:16:37 <gk> ... e.g. in data at end of http://www.ninebynine.org/RDFNotes/Calendaring/Users.n3

17:16:38 <danbri> yes, merging icals-in-rdf should be a useful rather than information-losing act.

17:16:49 * DanCon asks "I like that" and "seconded" to be qualified by proposal names, e.g. yes, I 2nd 1.1q1

17:17:00 * libby agrees danbri

17:17:06 <danbri> so /me 3rds DanCon's proposal

17:17:19 <danbri> I 2nd 1.1q1

17:17:21 <timbl> I seconded 1,1q1

17:17:29 <libby> but for vtodos etc, doesnt really need to eb adiffeent property

17:17:30 * DanCon is about to make another proposal, rendering "DanCon's proposal" ambiguous

17:17:39 <gk> Sorry, lost the plot, what's 1.1q1?

17:18:13 <DanCon> PROPOSED(1.1q2): to use { ical:contains a rdf:Property; rdfs:range ical:Component . } to relate calendars to events, e.g. in the case of multiple calendars in one file.

17:18:14 <gk> Seen it now ... I support it as far as it goees

17:18:23 <libby> [[

17:18:25 <libby> 4.4 iCalendar Object

17:18:25 <libby> The Calendaring and Scheduling Core Object is a collection of

17:18:27 <libby> calendaring and scheduling information. Typically, this information

17:18:29 <libby> will consist of a single iCalendar object. However, multiple

17:18:31 <libby> iCalendar objects can be sequentially grouped together.

17:18:33 <libby> ]]

17:18:35 <libby>http://www.ietf.org/rfc/rfc2445.txt

17:18:35 <dc_rdfig> F: http://www.ietf.org/rfc/rfc2445.txt from libby

17:18:43 <libby> oops

17:18:56 <DanCon> F:|icalendar spec.

17:18:57 <dc_rdfig> Titled item F.

17:19:15 <DanCon> F:see esp 4.4 iCalendar Object re 1.1 on today's agenda

17:19:15 <dc_rdfig> Added comment F1.

17:19:35 <DanCon> any takers on 1.1q2, i.e. ical:contains?

17:19:37 <libby> doesnl;t seem to say much else

17:19:54 <gk> I can support 1.1q1 or 1.1q2, have inertial preference for q1, but not strongly so

17:20:05 <DarkPenguin> Me too

17:20:33 <libby> I second 1.1q2

17:20:40 <DanCon> hmm... maybe ical:component rather? I'm not too crazy about nouns for properties when they're not naturally role nouns. But maybe component is a role noun.

17:20:56 <timbl> I note in passing that specifying the class of a Vevent is unnecssary in 1.1q1 when it follows from the range of the event property.

17:20:58 <gk> Question: will the use of names clearly related to RFC2445 (e.g. Vcalendar, etc) ease acceptance by that community?

17:21:06 <DanCon> 1.1q2 (ical:contains) looks easier to implement, to me.

17:21:09 <timbl> component is a role noun.

17:21:54 <DanCon> PROPOSED(1.1q3): to use { ical:component a rdf:Property; rdfs:range [ owl:unionOf (ical:Vevcent ...) . } to relate calendars to events, e.g. in the case of multiple calendars in one file.

17:22:19 <libby> I'll support either 1.1q2 or 3 - prefer 3 becauuse less generic

17:22:42 <timbl> I wonder why you limit it so much - why not leave 'component" as a more generic thing others could reuse?

17:22:48 * danbri nods

17:23:03 <gk> q3 is *less* generic?

17:23:13 <danbri> could just use rdfs:member, the superproperty of _1 _2 _3... I'm not sure how much value adding a range constraint does for us here.

17:23:13 <DanCon> why not more generic: xp:youAren'tGonnaNeedIt

17:23:17 <libby> well component is a work used in icalendar

17:23:35 <libby> word rather

17:23:43 <timbl> If you aren't going to need the range constrint, leave it out.

17:24:01 <timbl> You might find later yu made it wrong.

17:24:42 <LotR> or a new compoment type shows up

17:24:56 <DanCon> PROPOSED(1.1q4): to use { ical:component a rdf:Property } to relate calendars to events, e.g. in the case of multiple calendars in one file. { :cal1 ical:component :ev1. :cal2 ical:component :ev2 }

17:25:19 <gk> For the purposes of discussion, is it useful to talk of, say, intendedRange?

17:25:53 <DanCon> ss:rangeIntersects is the name I've implemented; close to intendedRange

17:26:33 <libby> I second 1.1q4

17:26:37 <timbl> I'm tending toward rdfs:member unless it has bugs we are aware of.

17:27:02 <DanCon> { [] :P [ a :C ] } log:implies { :P :rangeIntersects :C }. -- http://www.w3.org/2000/10/swap/util/sniffSchema.n3

17:27:16 <timbl> It would mean that the calendar becomes presumably a set.

17:27:16 <gk> Dan, sounds sensible. SO we might paraphrase 1.1q4 as { ical:component a rdf:Property; ss:rangeIntersects ical:Vevent }

17:27:18 <DanCon> hmm... member...

17:27:22 <DarkPenguin> I gravitate towards 1.1q4

17:27:41 <DanCon> yes, we might paraphrase 1.1q4 as ... .

17:27:48 <danbri> I'd be happy with a custom property and a range of i:Component

17:28:09 <timbl> ical:component rdfs1:superProperty rdfs:member?

17:28:17 <gk> 1.1q4 works for me

17:28:42 <DanCon> PROPOSED(1.1q.member): to use rdfs:member to relate calendars to events, e.g. in the case of multiple calendars in one file. { :cal1 rdfs:member :ev1. :cal2 rdfs:member :ev2 }

17:28:58 <gk> tim, to paraphrase your earlier comment, if we don't sue that info why say it now ... we might be wrong ?

17:29:06 <gk> s/sue/use/

17:29:21 * DanCon considers taking a preference poll.

17:29:24 <timbl> yes.

17:29:30 <timbl> gk, yes.

17:29:32 <DarkPenguin> good call

17:29:46 <gk> Yes to poll

17:29:59 <timbl> Also, I am thinking of syetsma dn langauges which will have generci support for sets just as they are starting to have support for collections now.

17:30:09 <DanCon> I prefer 1.1q4.

17:30:14 <timbl> s/collections/ordered collections/

17:30:27 <gk> I prefer 1.1q4 (for now)

17:30:38 <DarkPenguin> 1.1q4 for me too

17:30:43 <danbri> does q4 ignore range issues for now?

17:30:48 <danbri> I like 1.1q4

17:31:00 <danbri> s/like/prefer/

17:31:12 <DanCon> q4 doesn't constrain the range, except to note it intersects ical:Vevent

17:31:17 <gk> danbri, effectively, except for the implied ss:rangeIntersects

17:31:33 <timbl> I prefer member but can settle for q4 -- afetr all, standard terms for things are so useful its always good to have more thanone, right? ;-)

17:31:43 <DanCon> any objections to q4?

17:32:15 <libby> nope

17:32:41 <DanCon> RESOLVED(1.1q4): to use { ical:component a rdf:Property; ss:rangeIntersects ical:Vevent } to relate calendars to events, e.g. in the case of multiple calendars in one file. { :cal1 ical:component :ev1. :cal2 ical:component :ev2 }.

17:32:47 <libby> hooray!

17:32:59 <gk> Great!

17:33:09 <libby> thanks DanCon

17:33:15 <timbl> next?

17:33:18 <danbri> logger_1, pointer?

17:33:18 <danbri> See http://ilrt.org/discovery/chatlogs/rdfig/2003-02-12#T17-33-18

17:33:35 <danbri> C:[http://ilrt.org/discovery/chatlogs/rdfig/2003-02-12#T17-33-18|RESOLVED(1.1q4): to use { ical:component a rdf:Property; ss:rangeIntersects ical:Vevent } to relate calendars to events, e.g. in the case of multiple calendars in one file. { :cal1 ical:component :ev1. :cal2 ical:component :ev2 }.

17:33:36 <gk> Next is: 1.2 Vcalendar rdf:about=""

17:33:36 <dc_rdfig> Added comment C4.

17:33:47 <danbri> C4:[http://ilrt.org/discovery/chatlogs/rdfig/2003-02-12#T17-33-18|RESOLVED (1.1q4)]: to use { ical:component a rdf:Property; ss:rangeIntersects ical:Vevent } to relate calendars to events, e.g. in the case of multiple calendars in one file. { :cal1 ical:component :ev1. :cal2 ical:component :ev2 }.

17:33:47 <dc_rdfig> Replaced comment C4.

17:33:56 <DanCon> BLURB:making proposals and deciding them in IRC.

17:33:57 <dc_rdfig> G: making proposals and deciding them in IRC. from DanCon

17:34:25 <DanCon> G:I offer a 300 point bounty for implementing the protocol we just used to decide 1.1q4, say in RRSAgent

17:34:25 <dc_rdfig> Added comment G1.

17:34:38 <libby> [[

17:34:38 <libby> 1.2 Vcalendar rdf:about=""

17:34:38 <libby> For testing I've removed this, because it generated different graphs

17:34:38 <libby> depending on the filename, and I wanted to be able to test if the RDF

17:34:39 <libby> graphs were identical ot not using Jena's comparison tool.

17:34:41 <libby> Dan Connolly wondered if it was bogus:

17:34:44 <DanCon> logger_1, pointer?

17:34:44 <DanCon> See http://ilrt.org/discovery/chatlogs/rdfig/2003-02-12#T17-34-44

17:34:45 <libby>http://lists.w3.org/Archives/Public/www-rdf-calendar/2003Jan/0009.html

17:34:45 <dc_rdfig> H: http://lists.w3.org/Archives/Public/www-rdf-calendar/2003Jan/0009.html from libby

17:34:49 <libby> ]]

17:35:03 <DanCon> G:see [discussion|http://ilrt.org/discovery/chatlogs/rdfig/2003-02-12#T17-34-44]

17:35:03 <dc_rdfig> Added comment G2.

17:35:11 <libby> re 1.2, given our last decision, we shoudl probably remove it

17:35:40 <DanCon> hmm...

17:35:44 <gk> I don't understand the issue, other than the mention of relating vCalendars to events now seems to be redundant

17:35:46 <danbri>http://esw.w3.org/t/view/ESW/RdfCalendar

17:35:47 <dc_rdfig> I: http://esw.w3.org/t/view/ESW/RdfCalendar from danbri

17:35:58 <danbri> I:|RdfCalendar Wiki page, for notes and scribbling...

17:35:58 <dc_rdfig> Titled item I.

17:36:16 <libby> H:|Dan Connolly's message to www-rdf-calendar mentionning Vcalendar rdf:about=""

17:36:16 <dc_rdfig> Titled item H.

17:36:48 <DanCon> I'd sure like to relate the self-document to the calendar somehow. But I suppose the implicit log:includes might be enough... hm...

17:37:05 <danbri> Is Vcalendar a subclass of a notional XMLRDFDocumentType; itself a subclass of WellFormedXML10Doc...?

17:37:23 <libby> I removed it for convenience in testing, but could put it back if a good reason for it

17:37:54 <danbri> reasoning about classes of RDF doc could be particualrly useful for discovery purposes: find me those Vcalendars that were created by co-workers, etc etc...

17:37:57 <gk> I don't see Vcalendar as a subclass of a document type ... am I missing something?

17:38:18 <DanCon> subclass: no, by about="", I meant a generic resource, not any particular representation.

17:38:22 <danbri> gk, i don't think it has been discussed quite in those terms (esp not explicitly subclassed).

17:38:36 <DanCon> you don't think a calendar is a document, gk?

17:38:39 <timbl> When I have been doing financial stuff, i uses a "source" property, which I attach to a transaction which records, for me, the resource I originally found it in (bank statement, etc). But that is somethingw which is me-relative. And it only becomes needed and valid when the information is picked up and moved.

17:38:44 <danbri> more the rdf graph than any specific XML/RDF serialization of it, DanC?

17:38:52 <gk> Our Vcalendar data is buried in a larger document ... it depends on what you mean by a "document"

17:39:13 <libby> if there can bemore than one vcalendar in a documen, then they are not identical...

17:39:22 * danbri suspects a rathole and we could probably come back to this later, with a more worked out model...

17:39:30 <DanCon> I mean document in the most ordinary sense... the calendar on my wall is a document.

17:39:43 * gk would be content to first understand the question

17:39:57 <DanCon> no, it's not a rathole. It's observable in libby's testing.

17:40:48 <DanCon> more than one: yes, that's compelling...

17:41:04 <gk> OK, does this file, which contains at least three Vcalendar descriptions, itself contain VCalendar documents? http://www.ninebynine.org/RDFNotes/Calendaring/Users.n3

17:41:11 <danbri> so there are members of the class VCalendar that are not XML documents? That should be test-case-able...

17:41:35 <danbri> Yes, good test case:

17:41:42 <Ol> mapping the 'vcalendar' concept of 2445 to the idea of a document is maybe somewhat dangerous, vcalendar is just a general wrapper in 2445 syntax.

17:41:46 <DanCon> PROPOSED(1.2q.drop): whereas relationship beteween files and ical:Calendars is not 1-1, drop Vcalendar rdf:about="" from our icalendar<->RDF mapping.

17:42:00 <danbri> Is the thing named by that URI a vcalendar?

17:42:10 * danbri nods, agrees with Ol

17:42:21 * DanCon wonders which uri?

17:42:50 <danbri>http://www.ninebynine.org/RDFNotes/Calendaring/Users.n3

17:42:50 <dc_rdfig> J: http://www.ninebynine.org/RDFNotes/Calendaring/Users.n3 from danbri

17:42:51 * gk still doesn't understand why Vcalendar rdf:about="" implies a Vcalendar *document*

17:42:54 <danbri> that one

17:43:07 <danbri> J:An N3 calendar test case from gk

17:43:07 <dc_rdfig> Added comment J1.

17:43:23 <DanCon> "" is a short form of the name of whatever document it's in, right gk?

17:43:24 <gk> J:The calendar data is at the end of the file

17:43:25 <dc_rdfig> Added comment J2.

17:43:51 <gk> Dan, gottit, thanks! That's not a convention I'm used to using yet.

17:44:01 <danbri> gk, maybe it doesn't. Basically if we are creating an RDF class called ical:Vcalendar, we need to be clear to ourselves and others about which things in the world are members of that class. I believe currently we're a bit confused.

17:44:11 <libby> I agree with 1.2q.drop

17:44:21 * DanCon expected libby to 2nd drop -- ah, there.

17:44:24 <libby> I think that lots of people are conbfused about the realtiosnhip

17:44:30 * libby thinking :)

17:44:33 <gk> Me to (i.e. agree 1.2q.drop)

17:44:35 <danbri> ...but yeah as danc points out, there's an 'i think therefore i am'-ism; rdf:about="" only happens inside RDF documents (of some kind...)

17:44:56 <DanCon> no, it occurs in HTML too.

17:44:59 <danbri> I agree with 1.2q.drop too

17:45:13 <DanCon> any objection to 1.2q.drop?

17:45:41 <DanCon> RESOLVED(1.2q.drop): whereas relationship beteween files and ical:Calendars is not 1-1, drop Vcalendar rdf:about="" from our icalendar<->RDF mapping.

17:45:50 <libby> :)

17:46:10 <timbl> :)

17:46:45 <libby> ok, agenda item 2?

17:46:47 <libby> [[

17:46:48 <libby> 2. recurring events

17:46:48 <libby> time: 30 minutes?

17:46:48 <libby> suggested by Graham Klyne

17:46:52 <libby> Ronan has done some work on converting my network access schedule

17:46:54 <libby> data to the new schema, and we have put together some notes and proposals,

17:46:58 <libby> available at [1]. Maybe this can seed some discussion of recurring

17:47:00 <libby> events? (Some background comments and links are at [2].)

17:47:01 * DanCon noodles on propagation times...

17:47:02 <libby> [1] http://www.ninebynine.org/RDFNotes/Calendaring/iCal-proposal.html

17:47:04 <libby> [2] http://www.ninebynine.org/RDFNotes/Calendaring/Intro.html

17:47:06 <libby> ]]

17:47:08 <libby> gk, want to take the floor?

17:47:18 <gk> OK...

17:47:21 <danbri> ronan=dark penguin?

17:47:25 <DarkPenguin> yep

17:47:29 <danbri> hi :)

17:47:48 <gk> The document at [1] summarizes Ronan's analysis of the current schema (which I understand is based on sniffing current data) ...

17:47:56 <DanCon> any .ics data relevant to this agendum? (2)

17:48:26 <libby> Dancon, you had some with recurring evcents didnt you?

17:48:34 <gk> ... he noticed some values used in our schedule stuff don't appear in the schema. These are listed.

17:49:01 * DanCon tries to think of non-confidential recurring .ics data...

17:49:05 <gk> Dan, there is a sample at http://www.ninebynine.org/RDFNotes/Calendaring/iCalendarExample.txt

17:49:08 <minddog> hey i'm trying to get the rdf syntax down, but w3 specs just give me a headache, know some better sites?

17:49:20 <libby>http://www.w3.org/2002/12/cal/test/cal01.rdf

17:49:20 <dc_rdfig> K: http://www.w3.org/2002/12/cal/test/cal01.rdf from libby

17:49:22 <DanCon> iCalendarExample.txt -- thx

17:49:34 <libby> K:some RDF recuing data

17:49:34 <dc_rdfig> Added comment K1.

17:49:39 <danbri> hi minddog! We're in middle of a meeting (on rdf calendaring, see http://rdfig.xmlhack.com/)

17:49:45 <libby> K:|some RDF recurring data

17:49:45 <dc_rdfig> Titled item K.

17:49:45 * danbri has a think re other docs

17:49:55 <libby> what about your onw on striping

17:49:56 <libby> ?

17:50:07 <libby> K1:""

17:50:07 <dc_rdfig> Deleted comment K1.

17:50:25 <danbri> minddog, see if http://www.w3.org/2001/10/stripes/ helps at all...

17:50:28 <minddog> oh neato thanks =)

17:50:33 <libby> K:[another example|http://www.ninebynine.org/RDFNotes/Calendaring/iCalendarExample.txt]

17:50:33 <dc_rdfig> Added comment K1.

17:50:47 <minddog> i'm the guy working on the c# implementation of redland/rdfparser

17:51:25 * danbri remembers :) (I have Euler/Csharp running in dotgnu on my machine now, so am gonna be in touch re RDF APIs soon)

17:51:25 * DanCon wonders if minddog has tried http://www.w3.org/TR/rdf-primer/ ; if that's no good, I recommend http://www.w3.org/2000/10/swap/Primer , though it doesn't cover RDF/XML syntax.

17:51:48 <gk> The proposal suggests: Vtimeentry, Vdatetime, Vdate , Vtime ,Vrecur , hasEvent. The last is now settled by 1.1q4.

17:52:17 <danbri> on recurrance, do you use Vrecur as a sort of template/generator for instances of the event class?

17:52:18 <minddog> yup mdupont got euler going =) going to use that for n3 in the implementation

17:52:31 <DanCon> Vtimeentry? does that correspond to any existing .ics vocab?

17:52:57 <DanCon> ical2rdf.pl does grok RRULE:FREQ=WEEKLY;BYDAY=MO,TU,WE,TH,FR

17:53:15 <gk> The sample data at http://www.ninebynine.org/RDFNotes/Calendaring/Users.n3 has examples of: Vevent dtstart dtend rrule Vtime Vrecur freq interval byday

17:53:41 <gk> danbri, not a template, but an RDF class

17:54:08 <gk> danc, Vtimeentry came from the hybrid schema, I think (Ronan...?)

17:54:18 <DarkPenguin> yes.

17:54:19 <DanCon> ical2rdf.pl get this [[

17:54:21 <DanCon> <Vevent>

17:54:21 <DanCon> <dtstart rdf:parseType='Resource'>

17:54:21 <DanCon> <dateTime>2002-12-01T19:00:00</dateTime>

17:54:21 <DanCon> </dtstart>

17:54:21 <DanCon> <dtend rdf:parseType='Resource'>

17:54:22 <DanCon> <dateTime>2002-12-01T21:00:00</dateTime>

17:54:24 <DanCon> </dtend>

17:54:26 <DanCon> <rrule rdf:parseType='Resource'>

17:54:28 <DanCon> <freq>WEEKLY</freq>

17:54:30 <DanCon> <byday>MO,TU,WE,TH,FR</byday>

17:54:32 <DanCon> </rrule>

17:54:34 <DanCon> </Vevent>

17:54:36 <DanCon> ]] from RRULE:FREQ=WEEKLY;BYDAY=MO,TU,WE,TH,FR

17:54:52 <danbri> [[

17:54:52 <danbri> ical:rrule

17:54:52 <danbri> [ a ical:Vrecur ;

17:54:52 <danbri> ical:freq "WEEKLY" ;

17:54:52 <danbri> ical:interval "1" ;

17:54:52 <danbri> ical:byday "MO,TU,WE,TH,FR" ]

17:54:54 <danbri> ] ;

17:54:56 <danbri> ]]

17:55:01 * danbri muses

17:55:16 <libby> damn, mine barfed

17:55:30 <DanCon> where in Calendaring/Users.n3 should i look, gk?

17:55:33 <timbl> This is an attempt to carrythe syntax , right, not model the semantics? just checking.

17:55:51 <libby> yep

17:55:59 <gk> DanC, look at the end.

17:56:06 <timbl> It mith eb reasonable to have a separate syntactic and semantic translation.

17:56:16 <danbri> gk, re the snippet i quoted: that's attached to a partial Vevent description, and systems that know the rules can generate other fully-described Vevents?

17:56:25 <danbri> (check my thinking for me, pls!)

17:56:27 <gk> DanC, so dateTime is a property with literal value. That would answer by question about using rdf:value.

17:57:04 <timbl> { ?r ical:freq "WEEKLY" } => { ?r a cal:WeeklyRecurrence }

17:57:26 <gk> danbri, wondering why you say *partial*. Checking...

17:57:48 <DanCon> I wrestled over "WEEKLY" vs ical:weekly

17:58:07 <DanCon> ical2rdf.pl implements some properties as symbolic enumerations

17:58:25 <DanCon> but not others.

17:58:35 <DanCon> uncomfortably arbitrary

17:59:23 <danbri> partial, cos they're not on particular days.

17:59:25 * DanCon still looking for example of Vtimeentry usage

17:59:32 <danbri> or do i musundetsand [ a ical:Vtime ;

17:59:32 <danbri> rdf:value "T220000" ] ;

17:59:50 <timbl> Hence the rpoposal to separate the syntactic "as clase as possoble to ical syntax" vocabulary from the better modelling.

18:00:26 <danbri> the generator takes an event that is 8pm-10pm and asserts instances of vCalendar for each day of the week that is a Saturday?

18:01:02 <gk> Well, what we did is (strictly) an abuse of iCalendar -- a recurring event is specified as a first occurrence and recurrence interval. But that didn't really help this application. In my mind, I was thinking that the absence of date specification allowed *any* date that satisfied the other properties.

18:01:23 <DanCon> well, where the ical RFC constrains strings to a particular enumeration, I think makin them URIs is still pretty much syntax conversion. But where arbitrary strings are allowed, they need to stay strings.

18:02:36 <gk> FWIW, the original syntax I used was based on copying/adapting some recurring schedule data using the hybrid schema that I found using foafnaut. The original was a college class schedule, I think.

18:03:09 <DanCon> ical:Vtime ... yes, that's an abuse of the http://www.w3.org/2002/12/cal/ical namespace policy, to wit: that the semantics of the terms are grounded in (1) conversion to .ics syntax, (2) consulting .ics spec and implementations

18:03:23 <danbri> gk, this reminds me of when i was playing with modelling geo stuff, and I would often omit the altitude when I said <g:Point g:lat="50" g:long="200" /> or whatever...

18:03:44 <danbri> ...ie using an underconstrained description that corresponds to potentially many actual points.

18:03:56 <danbri> The markup says 'there exists a point and it has a lat of, and a long of...'

18:04:11 <danbri> ...and you sorta do that and then decorate it with a list of rules for determining instances?

18:04:32 * gk nods, but also contemplates DanC's comment

18:04:55 <DanCon> underconstrained is fine... but terms that can *never* arise from converting a .ics file have no right to be in the http://www.w3.org/2002/12/cal/ical# namespace

18:05:26 <gk> A *strict* conversion from Vtime to .ics would require adding in a date before the first use of the system.

18:06:29 <gk> Dan, I see where you're coming from, I think. The fact that iCalendar doesn't use bare times like this means that we shouldn't invent them?

18:06:31 <danbri> we're having two separate threadlets here, DanCon. I think your point re namespace policy is reasonable, but can we not think about gk's point (hypothetically imagine he used a cal-utils aux namespace...)?

18:06:35 <DanCon> PROPOSED(2q0): to represent RRULE as ical2rdf.pl,v 1.6 2003/01/22 21:17:12 does, using rrule, freq, byday

18:07:15 <DanCon> can we not: yes, we can. "underconstrained is fine"

18:07:29 <danbri> ok

18:07:31 <danbri> hi jo

18:07:42 <zool> hey danbri

18:07:52 <gk> DanC, (a) also uses ical:dateTime; (b) what's the range of ical:rrule?

18:08:29 <DanCon> (a) refers to what?

18:08:50 <gk> (a) was just the first of two points.

18:09:15 <DanCon> range of ical:rrule intersects domain of freq, byday. But I don't see any need to name the class.

18:09:24 <gk> DanC, I think your conversion has a serious bug, omitting interval when it's "1". I think that takes us into non-mon territory.

18:09:49 * danbri looking to support some proposal for recurrance rules so i can go away and test it; am not feeling qualified to discuss detail at this stage.

18:10:10 * libby neither.

18:10:26 <DanCon> hmm... I didn't license any non-mon inferences when I left the interval out...

18:10:28 <gk> DanC, apart from the interval issue, I'm happy to support your proposal.

18:10:39 <DanCon> ... but I agree that making defaults explicit is likely to be better.

18:11:52 <danbri> yes, explictness in defaults seems good

18:11:54 <DanCon> I don't think we can resolve this agendum in the alloted time. Any volunteers to take an action to make a proposal by email? (it's ok to check in your proposal code/test data before it's agreed.)

18:12:30 <DanCon> or maybe gk can make a concise proposal?

18:12:33 <DanCon> now

18:12:44 <gk> Suppose you have an interval=2 property when the event is (say) every other day. By the RDF subgraph lemma, the same graph *without* the interval property must be true whenever the original is true. Then I think that the most we know when omitting the interval property is that there is some interval for which the subgraph is true, and cannot assume an interval of 1.

18:13:07 <DanCon> cannot assume an interval of 1: quite@!

18:13:15 <DanCon> quite!!!

18:13:19 <timbl> If I am rigt in understanding that there is a default of 1, then I think in a way it is syntax conversion to add the default.

18:13:23 * danbri remembers webont design had similar issues

18:13:46 <gk> tim, yes, I think the syntax conversion must add in the default.

18:14:02 <danbri> (illformedthought) should we using the first/next/last container representation to indicate closable world?

18:14:18 <DanCon> it "must" on social grounds; i.e. if it doesn't, folks will *think* the nonmon assumption is licensed.

18:14:24 <timbl> cwm could handle it but it would be a pain - it would need to represent the formula of the vevent and test it for having a value.

18:14:53 <DanCon> ah... on technical grounds too: the rdf->ical mapping can't do anything wihtout an explicit interval.

18:15:08 <timbl> precisely.

18:15:35 <gk> danbri, I considered this in my application, and was happy that the semantics were OK with "open" container form. But I'm not sure this would always work for other applications. This issue of open/closed containers did bite me on another part of the same application.

18:15:51 <DanCon> OK, i object to 2q0.

18:16:28 * danbri scrolls back looking for 2q0

18:16:35 * DanCon checks the time vs. the agenda...

18:16:48 <danbri> ah <DanCon> PROPOSED(2q0): to represent RRULE as ical2rdf.pl,v 1.6 2003/01/22 21:17:12 does, using rrule, freq, byday

18:17:33 <gk> PROPOSED(2q1): to represent RRULE as ical2rdf.pl,v 1.6 2003/01/22 21:17:12 does, using rrule, freq, byday, and also with interval always specified. If ics interval is omitted, translation must add interval="1"

18:17:58 <timbl> BTW how does one know that an event is not a recurring event?

18:18:06 <DanCon> narrow the proposal to just the RRULE in http://www.ninebynine.org/RDFNotes/Calendaring/iCalendarExample.txt , I suggest

18:18:20 <DarkPenguin> there is no rrule propetyif the event does not recur

18:18:22 * DanCon takes a call

18:18:46 <timbl> Ah ... another default problem.

18:18:46 <danbri> tim, do events recur, or are they just used as a way of describing patterns of other events?

18:19:05 <gk> Tim, good question; DarkPenguin, that runs against the same problem I raised for interval.

18:19:09 <timbl> don't philosophate me here danbri ;-)

18:19:29 <danbri> heheh

18:19:45 <danbri> it's a serious question re the identity conditions for the class 'event'

18:19:56 <gk> I'm also reminded there's an issue of when recurring events terminate. I think this needs more thought.

18:20:16 <libby> I think that the icalendar people had trouble with recurrence

18:20:17 <danbri> not essential, but really really useful when modelling w/ classes to have some sense of strategies for declaring two members of the class as identical.

18:20:26 <danbri> i'm not suprised, its a toughie...

18:20:47 <libby> yeah, probbaly not solvable today.... ;)

18:21:01 <libby> it's great that we've got an app whihc is using this stuff

18:21:06 <gk> OK, I'd be happy to think forwards using DanCs ical2rdf example as a starting point.

18:21:12 <soccos> soccos is now known as soccos|away

18:21:20 <minddog> has anyone done an implementation of rdf with abstract data types

18:21:24 <timbl> A recrring event is a pretty well defibend thing. There is a relaetd set (possibly infinte) of non-recurring events which is just related that's all, not equivalent.

18:21:56 <gk> IIRC, the hybrid schema had a different class for recurring events.

18:22:13 <timbl> That resolves the problem.

18:22:39 <danbri> on timbls account, these are mutually disjoint classes? hmm he didn't quite say that.

18:22:43 <timbl> You could of course have your cake and eat it: the syntactic conversio has a Ventent class, and the semantic one specializes them to rec and non-rec.

18:23:01 <danbri> ...mutually disjoint seems right; all else makes my head hurt

18:23:18 * danbri didn't parse much of that

18:23:52 <timbl> RecurringEvent is related to SetOfNonRecurringEvents by a sort of "burst" or "split" relationship.

18:24:31 <timbl> It is very useful for test cases, as the setofnonrecurringevents is the easiest way to specify what you mean by a recurring evnt.

18:24:35 <gk> Yup... hybrid schema had REC-VEVENT as well as VEVENT.

18:25:06 <gk> ... and it's a subclass of VEVENT. Umm, that doesn't help us.

18:25:26 <libby> we didn;t think alot about recurrence

18:25:54 <libby> we have 5 minutres alotted time left

18:25:55 <timbl> It is also in practice worth hving a spec for a simpler calendar which dooes not handle recurring events, for import and export and wider interop. So a well-defined mapping, even though lossy in that it it loses the original recurring event, is valuable.

18:26:10 <gk> If anything, VEVENT should be a subclass of REC-VEVENT. (Libby, you did great job, gave us something to work with.)

18:26:46 <timbl> ? otehr wayaround?

18:27:24 * libby would like to decide next meeting time and date before the end of the meeting; also talk about putting stuff on w3.org if possible

18:27:25 <timbl> VEVENT - some event, may or may not have recurrence.

18:27:38 <gk> Yes: a once-only event is one that recurs zero times. For equality test, we know that a VEVENT without rrecur isn't missing importamnt info.

18:27:39 <timbl> REC_VEVENT - an event which does have recurrence info

18:27:57 <timbl> SIMPLE-EVENT - one that has no recurrence info.

18:28:19 <timbl> If a recurreing event happens to only happen once, taht doesn't make it simple.

18:28:59 <timbl> When i say i can handle a simple event, it means I don'tgrok recurrence info, so don't say we are meeting next every wednesday between feb 14and feb 21.

18:29:37 <libby> I can't make next week: propose 2 weeks' time for next meet: http://www.timeanddate.com/worldclock/fixedtime.html?day=26&month=2&year=2003&hour=17&min=0&sec=0&p1=0

18:29:42 <gk> I didn't mean to make it simple. But the logic imposed by RDF subgraph entailment seems to work that was, not the other. Your disjoint proposal works too, I think (but need to consider some more)

18:30:16 <gk> Two weeks good for me

18:30:26 <timbl> ok for 2 weeks. i can't make next week either not that i am critical

18:30:57 <libby> cool

18:31:05 * libby wonders if DanCon can make 2 weeks time

18:32:40 <gk> I'd like to revisit for some future agenda (not necerssarily next) if it's OK to have times without dates in a Vevent.

18:32:53 <libby> ok, thanks gk

18:33:04 <libby> I think we have examples of that

18:33:42 <libby> I'll move remaining items to nnext agenda andplease get in touch withme if you want other stuff

18:33:50 <danbri> I can do two weeks time.

18:33:58 <libby> we'll go for 2 weeks time

18:34:02 <danbri> Libby, will you send round the timeslot in ical?

18:34:29 <danbri> for extra points, send round a proposal for a standing fortnightly meeting at this slot using repeating rule proposal(s)... ;)

18:34:29 <libby> C: next meeting [26th Feb 2003 17:00 UTC|http://www.timeanddate.com/worldclock/fixedtime.html?day=26&month=2&year=2003&hour=17&min=0&sec=0&p1=0]

18:34:29 <dc_rdfig> Added comment C5.

18:34:34 <libby> yep, will do

18:34:46 <libby> no, not the repeating thing!

18:34:51 <libby> gk can do that :)

18:34:59 <libby> thanks everyone

18:35:16 <libby> ol,thanks for sending around the relative uri message

18:35:59 * libby feels like killed an interestign conversation w boring meeting times discussion :(

18:36:42 <danbri> cool, thanks guys that was interesting.

18:37:07 * danbri reads back to ponder where recurrence got left

18:37:19 * minddog invented a data manipulation ipc with rdf

18:37:28 <minddog> data-type rather

18:37:49 <ericP> dajobe, gk, others, could yo explain the <rdf-wrapper> bit in http://www.w3.org/TR/2003/WD-rdf-concepts-20030123/#section-XMLLiteral ?

18:38:45 <ericP> is this <rdf-wrapper> element supposed to surround what was parsed with parseType="Literal" ?

18:39:14 <danbri> to make text strings into wellformed xml docs, i guess

18:39:17 <zool> recurrence strikes me persistence of properties associated with an event, organiser, participants, venue etc... maybe a threshold of similarity/association can be set

18:39:32 <danbri> interesting question, ericP...

18:39:42 <danbri> (I hadn't noticed that bit of the spec. <blush/>)

18:40:12 <ericP> so i don't think they stuff in the model needs to be a well-formed XML document

18:40:37 <ericP> say it's stored in a persistent store, then returned as an answer to a query,

18:41:01 <ericP> the query engine can just package it up again in another parseType="Literal"

18:41:15 <libby> zool, I think apps working out identity of events very interesting/difficult :)

18:41:22 <gk> Thanks Libby.

18:41:49 <minddog> <rdf-wrapper><[CDATA[unparsed data ..dfasdfhiow45hu4]>></rdf-wrapper> as a workaround

18:41:57 <libby> thank you gk: recur tricky put v intersting

18:41:58 <ericP> this treatment of XMLLiteral data works fine for annotea

18:42:15 <danbri> ericP, that's good to know. Can you send something to www-rdf-comments to that effect?

18:42:20 <danbri> (we normally only get complaints ;)

18:42:40 <danbri> also did you see my query w.r.t. whether Annotea has a use case for rdf:bagID syntax ?

18:42:44 * ericP considers re-wording so it doesn't look too whiney

18:43:08 <ericP> danbri, query, nope, but it doesn't

18:43:36 <danbri> oh i thought you used it to round-trip attribution-tagged rdf between client/server...

18:44:15 <ericP> nah, i use rdf:ID

18:44:16 <gk> Eric, as I recall, <rdf-wrapper> stuff is to make a well-formed document so that canonical form can be described by simple reference to XML C14N specification. It feels a bit kludgy, but I couldn't see a neater way to do it.

18:44:39 * danbri wishes http://www.idleworm.com/nws/2002/11/iraq2.shtml were in SVG

18:45:15 <danbri> gk, is rdf-wrapper in the Syntax spec anywhere?

18:45:25 * DanCon returns from call, wonders where we are on the calendar agenda; adjourned?

18:45:28 <ericP> gk, i'd recommend saying that each child of the propertElt with parseType="Literal" is a exc-c14n document

18:45:42 <danbri> adjourned, i believe.

18:45:48 <DanCon> thx

18:46:07 <gk> Danbri, not as I recall. It's part of the mapping from abstract graph to the domain of interpretation.

18:46:27 <danbri> it has angle brackets in it, so seems syntaxy...

18:47:30 <ericP> if i find such a literal as a result of a query, will it have <rdf-wrapper> around it?

18:47:38 <danbri> depends on the query api

18:47:40 <gk> Eric, I'd be happy to suggest that to Jeremy, but is that enough? What about the space between the children? I don't know XML subtleties well enough to know.

18:47:41 <ericP> so that i need to XPath my real data out?

18:47:55 <danbri> presumably we won't see <rdf-wrapper> in ordinary rdf serializations?

18:48:18 <ericP> gk, inter-children space, hmm, thinking...

18:48:19 <gk> danbri, absolutely; not in RDF/XML serialization.

18:49:33 <ericP> what if rdf-wrapper were made more abstract, ie, some element, so that folks would never try to put it in or take out out with XPath.

18:49:36 <ericP> ?

18:50:02 <amy> bye all

18:50:17 <gk> As I understand it, it's just a device to help the formal specification, and (maybe) the description of equivalend RDF graphs under XML canonicalization. The presence of XML literals in the RDF somewhat middies that issue, I think.

18:50:38 <ericP> assuming my current interp is correct, that it is a temporary construct to make sure the canonicalization starts one down from the root element

18:51:11 <DanCon> DanCon has changed the topic to: RDF & Semantic Web hack'n'chat http://rdfig.xmlhack.com/

18:51:30 <ericP> algorithm:

18:51:36 <gk> Eric, I think there's a problem of exposition if folks are tempted to put in <rdf-wrapper> in their XPath.

18:51:43 <ericP> get byt string from API or parser

18:51:54 <ericP> stick inside <rdf-wrapper>

18:52:04 <ericP> allpy exc-c14n

18:52:16 <ericP> remove from <rdf-wrapper>

18:52:20 <ericP> stick in graph

18:52:22 <ericP> done

18:53:19 <gk> Eric, I'm very sympathetic. Don't know if that's enough. I believe the intent of what you describe matches the WG intent.

18:53:46 <ericP> i'll think about this and see if i can come up with anything

18:54:36 <zoyd> other than RDF, what else will/does make the semantic web happen ?

18:54:58 <ericP> right now, my code (serving annotea) does something like that, except it doesn't need to stick it inside <rdf-wrapper> as that fucntionality is available in my canonicalizer

18:55:45 <ericP> other than inter-child spaces, is there an affect of the <rdf-wrapper>

18:55:46 <ericP> ?

18:56:28 <ericP> (other than to confuse my poor canonicalizer's understanding of which namespaces are currently in play)

18:59:34 <gk> Eric, you just reminded me: <rdf-wrapper> is also used to bring the appropriate language tag into play. I suppose that might be done on each child element. But now I feel the complexification of this approach may start to bite.

19:00:47 * gk dinner calls, must run

19:01:35 <libby> bye gk

19:02:52 <danbri> cu

19:02:54 <ericP> chicken

19:02:59 <ericP> i mean, cheers

19:03:07 <danbri> cheers ff

19:04:23 <libby> chicken?

19:04:24 * ericP thinks about whether this is a document or a fragment and whether nested semantics are going to provide wierd 2nd order logic affects

19:05:03 <ericP> "chicken" 'cause he was afraid of the complexification

19:05:16 <libby> oh, heh

19:05:25 <ericP> just idly picking on folks. now that he's gone, i'll start on you.

19:05:34 <libby> heey

19:05:36 * danbri cowers

19:07:15 <zool> fear induced by pragmatism...

19:16:44 <mdupont> hi

19:16:50 <leobard> hullo

19:17:02 <mdupont> hi leobard

19:17:17 <mdupont> Grüsse1

19:17:34 <leobard> Servus.

19:56:10 <soccos|away> is xcal dead?

19:56:15 <soccos|away> soccos|away is now known as soccos

19:58:01 <sandro> I wondered that myself. (but only in passing; I don't pay a lot of attention to calendaring right now)

19:58:31 <soccos> I'm wondering it the RDF effort is making it irrelevant

19:59:08 <sandro> Right -- it might. But that doesn't mean there aren't a hundred companies working on xcal software -- I have no idea.

20:00:37 * soccos wants to hear a guru speculate

20:04:12 * sandro smiles silently

20:12:10 <danbri> investing in xcal shouldn't be a dead end; we can always generate it from the RDF variant, or (I'd hope) XSLT from xcal into RDF...

20:12:54 <sandro> But are people investing in xcal these days?

20:13:08 <danbri> i have no idea!

20:13:11 <danbri> i'm not

20:13:28 <danbri> offtopic, but...

20:13:55 <danbri> anyone know some good voice over TCP/IP systems that might replace some of my (non-work) international phone usage?

20:14:08 <danbri> (the phone company just rang up to warn me of my next bill's size!)

20:14:43 <danbri> couple of hours to boston, australia, france... soon adds up. I've lost track of what's what with voice over the Internet.

20:15:38 <deltab> to talk to others with Internet access, or without?

20:15:47 <danbri> with

20:51:18 <soccos> soccos is now known as soccos|away

21:11:40 <eikeon> danbri: I have used net2phone via both Yahoo Messenger and Windows Messenger to make long distance phone calls. There account management seems a bit flaky, but the quality has been pretty good.

21:13:13 <eikeon> danbri: Also, depending on what all firewalls are between you and who you wish to talk to... I have had good luck talking via Windows Messenger... not across any of the big ponds, but across a few states.

21:33:32 <danbri> anything that works on linux? or w/ open protocols?

21:33:35 <danbri> naive? ;)

21:33:39 <danbri> (me, i mean!)

21:34:17 <Morbus> speaking of naive, ever notice how the bottled water "Evian" is naive backwards?

21:35:08 <mdupont> Morbus: hehhe

21:35:14 <mdupont> that is funny!

21:35:17 <Morbus> :)

21:35:33 <mdupont> did you know that selters water is from a small town on the lahn

21:35:40 <mdupont> near the loreley

21:35:47 <mdupont> next to the rhein?

21:36:18 <Morbus> is there a joke in there, or are we being factual? ;)

21:37:59 <mdupont> fatual

21:38:01 <mdupont> factual

21:38:09 * eikeon has not tried any voice over TCP/IP apps from the linux side... keep an XP box on my other desk instead ;)

21:38:09 <mdupont> in the states growing up

21:38:21 <mdupont> i always referred to bubbly water as selters

21:38:36 <mdupont> now i get bottled selters, from the original source

21:38:43 <mdupont> :)

21:38:48 <mdupont> go figure

21:39:15 <mdupont> it is like champange

21:39:39 <Morbus> yeah. i've not been a fan of seltzer water.

21:42:28 <DanCon> I'm interested in voice/ip/linux... notes at http://dm93.org/z2001/RemotePresence

21:44:49 <danbri> thx dan

21:44:58 <Mutiny> eikeon: what windows client do you use for VoIP, just curious.

21:46:44 <eikeon> Mutiny: I have just used the make phone call feature of MSN Messenger and Yahoo Messenger... not even sure if that is VoIP or not.

21:47:14 <DanCon> I'm really bummed that I returned that voipblaster... I got one for only ~$20 awhile ago, but returned it when I couldn't get it to work on linux. Now I gather linux support is available, but the gizmos themselves are no longer widely available for $20. :-{

21:47:19 <Mutiny> yeah, i believe MSN Messenger uses SIP

21:47:55 <Mutiny> quicknet linejacks are nice

21:48:00 <DanCon> SIP is for call set-up; any idea what MSN uses for the actual voice transmission, Mutiny?

21:48:03 <soccos|away> soccos|away is now known as soccos

21:48:19 <Mutiny> very good linux support, with opengatekeeper interface

21:49:06 <Mutiny> no idea, dancon

21:54:56 <rhoads> Anyone see a problem with including custom datatype definitions in the same XML file in which an RDF Schema is defined which references those datatypes?

21:56:57 <DanCon> "custom datatype definitions"? do you mean in XML schema syntax?

21:57:19 <DanCon> getting it to parse inside an RDF schema might be tricky.

21:57:37 <DanCon> but I suppose you could do it with parseType="Literal"

21:57:42 <rhoads> Yes. Custom types built on the XML Schema built-in types.

21:58:07 <rhoads> The datatype defs would be located OUTSIDE the <rdf></rdf> tags

21:58:42 <soccos> soccos is now known as soccos|away

21:59:18 <DanCon> er... what would you use for the root element?

22:00:18 * DanCon prefers to take the component graph represented by the XML Schema declarations and serialize as RDF/XML

22:01:35 <rhoads> mmmmm ... better go brush up on my XML.

22:04:24 <mdupont> eikeon++

22:04:26 <mdupont> thxs

22:08:32 <danbri> DanCon, do you have an example of that I could look at?

22:08:54 <DanCon> umm... not a pretty on.e

22:10:28 * danbri tries to get his usb zoomcam working w/ gnomemeeting

22:13:40 <mdupont> mdupont is now known as mdupont-away

22:17:57 <danbri> hey techtipsters! here's a gem. Try plugging your USB webcam into your computer before wondering why the driver didn't detect it.

22:23:46 <DanCon> chuckle

22:26:16 <danbri> so i have gnomemeeting working now w/ video, just local loopbback, fiddling around now w/ audio... Do you have a working gnomemeeting at your end? maybe we could try it sometime?

22:28:38 <DanCon> hmm... yes, let's...

22:28:56 <DanCon> (1) firewall issues... lemme poke a hole or whatever...

22:30:30 <DanCon> ah... seems already done...