Semantic Web Interest Group IRC Chat Logs for 2004-01-21

00:03:47 <mortenf> C:

00:03:49 <dc_rdfig> blurb

00:03:50 <dc_rdfig> some issues to be resolved for p[arts of image annotation for w3 photo project

00:03:51 <dc_rdfig> (1:libby) polygons or just rectangles

00:03:52 <dc_rdfig> (2:libby) identifying parts of images

00:03:53 <dc_rdfig> (3:libby) 'depicts' for parts of an image - or some other property

00:03:54 <dc_rdfig> (4:libby) images or multimedia

00:03:55 <dc_rdfig> (5:danbri_dna) Note: this intersects re privacy issues. Some of might not want our annually expanding waistlines documented in tight-fitting bezier curves

00:03:56 <dc_rdfig> (6:mortenf) current consensus regarding photos indicates: class "Image", property "hasRegion" [domain: Image, range: Region], class "Region" (subclassed into Polygon, Circle etc.), property "regionDepicts" [domain: Region, range: Resource].

00:03:59 <dc_rdfig> (7:libby) action jibberjim and mortenf - send a message to the w3photo mailing list - a schema and a sample document if possible

00:04:02 <dc_rdfig> (8:libby) [WWW2004 coference photos|http://rdfig.xmlhack.com/2004/01/19/2004-01-19.html#1074522663.890375]

00:04:04 <dc_rdfig> (9:GregElin) action GregElin and libby pick out 3-5 images to use for ontology examples.

00:04:06 <dc_rdfig> (10:mortenf) ACTION jim and morten also list options for describing a region, whether by SVG-path or polyline/circle/ellipse, and whether to include rect and an optional bounding box.

00:04:51 <mortenf> C:Morten sent a [http://lists.w3.org/Archives/Public/www-archive/2004Jan/0096.html|mail on image regions] to the list

00:04:55 <dc_rdfig> Added comment C11.

00:13:10 <D[a]vey> D[a]vey is now known as Davey

03:00:22 <dmiles> wow goole is alwready indexing rdf logs from 24 hrs ago

03:01:04 <dmiles> idexing.. but who knows how quick it gets in database

03:01:19 <dmiles> 25.82.68.64.in-addr.arpa name = crawler10.googlebot.com.

03:02:32 * DanC wonders what'sup

03:02:46 <dmiles> 'sup homez

03:02:50 <dmiles> ;P

03:03:44 <DanC> dmiles, since you're here, are you interested at all in RDF/HTML integration issues? heard of GRDDL?

03:03:51 <DanC> .google GRDDL

03:03:56 <DanC> phpht

03:04:04 <dmiles> i am already gooling

03:04:24 <dmiles> at http://www.w3.org/2004/01/rdxh/spec

03:04:28 <DanC> yup

03:04:57 <DanC> that's a few days old; the freshest bits are the history/rationale, specbg.html

03:05:10 <DanC> but I'm interested in feedback on rdxh/spec too

03:05:18 <dmiles> i find i am ussing variety of ontolies.. so am interested in as much transformation metadata as possible

03:05:26 <dmiles> erm ontologies

03:05:32 * DanC guessed as much

03:05:57 <dmiles> , a mechanism for encoding RDF statements in XHTML and XML to be extracted by programs such as XSLT transformations

03:06:05 <dmiles> generic thought right?

03:06:12 <dmiles> (not rdf specific)

03:06:29 <dmiles> erm /thought/though

03:06:54 * dmiles had gotten you confused that he was confused

03:07:18 <dmiles> so yes .. interested

03:07:29 * DanC isn't at all sure he's following

03:07:36 <dmiles> i have specific programs i write to do these transformations (from prolog)

03:07:46 <dmiles> "generic thought right?"

03:07:50 <dmiles> meant though

03:08:25 <dmiles> so for a RDF spec in my case i am slanted towards a transliteratuion of the prolog code that actually is used

03:09:30 <dmiles> for example.. kif2cycl([instance,X,'Class'],[instance,X,'Collection'])

03:09:36 <dmiles> erm

03:09:45 <dmiles> kif2cycl([instance,X,'Class'],[isa,X,'Collection'])

03:10:18 <DanC> well, there is a more generic transformation problem, but GRDDL is about getting RDF/XML out of XHTML and XML.

03:10:46 <DanC> more about XHTML... the generalization to XML is still sorta debated by some of my colleagues.

03:11:01 <DanC> do you see the link to Joe Lambda's homepage?

03:11:40 <dmiles> ok there, looking for his exmaple

03:12:28 <dmiles> the "Demonstration of an RDF in XHTML processor"

03:12:44 <dmiles> erm i am getting ahead

03:12:48 <dmiles> ok at page

03:14:21 <DanC> maybe it gives you a feel for why GRDDL?

03:14:39 * DanC isn't sure what he's fishing for.

03:14:52 <dmiles> i am reading 3. The GRDDL attribute in XML

03:14:59 <dmiles> now so getting more of a sense

03:15:50 <dmiles> well your sharing something that i would be quick to use/need

03:16:00 <DanC> that's nifty.

03:16:27 <dmiles> since i have my irc bot mining and snaking out all the rdf/daml/kif/cycl that it can get

03:16:36 <DanC> if you're in a mood to read the history/rationale, I'd like to know if it makes sense.

03:16:46 <dmiles> the ebnd reault will be something geared to owl

03:16:58 * DanC re-reads the specbg...

03:18:01 <dmiles> klets see.. with this spec .. i should be able to make Kif2Cycl that you could use later?

03:18:38 <dmiles> erm cycml to daml lets say

03:18:58 <DanC> hmm... it's a stretch to apply GRDDL to anything besides XML, so KIF input seems a stretch.

03:19:17 <dmiles> yeah .. so reducing ti to CycML ;P

03:19:39 <DanC> in particular, GRDDL is about labelling the input of the transformation as saying "I hereby declare transformation XYZ as meaning-preserving"

03:19:46 <dmiles> CycML is a Cyc to Cyc IPC query languge

03:20:29 <DanC> so the GRDDL concept would only apply to KIF files that are written in anticipation of being transformed to cycl

03:20:37 <DanC> hi jim

03:20:46 <dkk> kinda curious... how do rdf people feel about xml messaging protocols? rest vs soap?

03:20:46 <jimh-scribe> hey Dan - how you doing?

03:20:58 <DanC> pretty good... good bandwidth here in the hotel.

03:21:19 <jimh-scribe> kendall?

03:21:31 <jimh-scribe> sorry - I mean are you staung at the kendall?

03:21:39 <dmiles> <svg xmlns="http://www.w3.org/2000/svg"

03:21:40 <dmiles> xmlns:data-view="http://www.w3.org/2004/01/rdxh/data-view#"

03:21:40 <dmiles> data-view:interpreter="http://www.example.org/2004/01/svg2dc.xsl"

03:21:40 <dmiles> width="4cm" height="8cm"

03:21:40 <dmiles> version="1.1" baseProfile="tiny" >

03:22:11 <dmiles> emr paisted that before my question was fully formed.. just aminute

03:22:24 <DanC> yup, I'm at the Kendall

03:25:18 <jimh-scribe> has the weather made it into the teens yet?

03:25:44 <dmiles> kinda funny in anouther chanelel we are designing a AIML impimentation for OpenCyc

03:26:14 <dmiles> which is going to be simular to xsl:template>

03:27:13 <dmiles> hrmr Transformation algorithms should be represented in XSLT

03:27:30 <dmiles> i see

03:28:33 <Davey> Davey is now known as D[a]vey

03:28:39 <dmiles> the spec is pretty clear

03:29:28 <dmiles> GRDDL being a namespace ?

03:31:09 <dmiles> there insttance a dc:* room for it?

03:31:13 <dmiles> erm,

03:31:24 <dmiles> ( there isnt space at dc: ?)

03:31:54 <DanC> umm... you can look at GRDDL as a namespace, containing the name "xslt2rdf". (that name is likely to change)

03:32:38 <dmiles> guess i am asking if it would have enough properties that it'd need it's own namespace

03:32:58 <dmiles> also becasue i'd think so many people would need to use it

03:33:59 <dmiles> would be an import tag in many html

03:34:19 <dmiles> erm important

03:36:22 <jimh-scribe> jimh-scribe is now known as jhendler

03:41:22 <dmiles> spec is clear, now has me reading "Using RDF/XML with HTML and XHTML"

04:46:33 <jmdupont> hi all

04:46:44 <jmdupont> any tips on dealing with large rdf files

04:46:56 <jmdupont> i have cwm maxing out onme

04:48:19 <aharth> distributed cwm on n nodes?

04:48:39 <aharth> probably nothing that's possible till tomorrow

04:50:03 <aharth> jmdupont: just curious, what size is your rdf repository?

05:01:13 <DanC> jmdupont, what does cwm do on these large files?

05:01:40 <DanC> ... and what was it *supposed* to do? ;-)

05:03:42 <aharth> danc: do you have an idea about to what amount of rdf cwm scales?

05:04:20 <aharth> my jena/mysql stuff gets really slow at about 1 mio triples (no inferencing though)

05:08:46 <DanC> well, for some things, cwm scales to infinitely large RDF files; i.e. cwm --pipe

05:09:15 <DanC> otherwise, it seems limited by RAM. I don't have any hard numbers. I haven't run into performance problems in a long time.

05:10:00 <DanC> I know that cwm can read the cyc ontology (the released part of it, anyway) and reason about it. I expected it to fall over doing that.

05:10:32 <aharth> cyc is pretty large i guess

05:10:45 <DanC> moderately

05:11:11 <aharth> number of rdf statements?

05:11:45 <DanC> I don't have a cached copy on this machine ... looking...

05:12:07 <DanC> . http://www.cyc.com/2002/04/08/cyc.daml

05:12:11 <aharth> site error at http://www.cyc.com/cyc-2-1/cyc.daml

05:13:00 <DanC> ew... 404 @ http://www.cyc.com/2002/04/08/cyc.daml too

05:15:36 <aharth> from http://www.isi.edu/nsf/papers/hovy2.htm: Later additions may include more of the 40,000-odd terms currently in CYC

05:16:07 <aharth> hmm looks like i have to check out cwm again...

05:17:32 * DanC finds http://opencyc.sourceforge.net/daml/cyc.daml

05:17:51 <DanC> not all 40,000 terms have been released in DAML (RDF) format

05:18:26 <jmdupont> aharth that is an interesing idea

05:18:33 <aharth> a@t22:~$ curl -XHEAD http://opencyc.sourceforge.net/daml/cyc.daml

05:18:33 <aharth> curl: (18) transfer closed with 2436919 bytes remaining to read

05:18:44 <jmdupont> )

05:18:44 <jmdupont> [06:11:00] <DanC> well, for some things, cwm scales to infinitely large RDF files; i.e.

05:18:45 <jmdupont> cwm --pipe

05:18:59 <jmdupont> that is interesting

05:19:01 <jmdupont> sorry, this thing did not beep

05:19:04 <jmdupont> when you answered

05:19:16 <DanC> $ time python ~/w3ccvs/WWW/2000/10/swap/cwm.py --rdf cyc.daml --no

05:19:16 <DanC>

05:19:16 <DanC> real 0m41.531s

05:19:16 <DanC> user 0m33.277s

05:19:16 <DanC> sys 0m0.515s

05:19:52 <aharth> 2.4 meg is not too big

05:19:56 <jmdupont> wow

05:19:57 <DanC> 40KLOC in cyc.daml; cwm read it in under a minute.

05:20:00 <jmdupont> but what did it do with it?

05:20:09 <jmdupont> here are some fat files of mine

05:20:09 <DanC> read it and indexed it, I think.

05:20:26 <jmdupont> let me check

05:20:42 <jmdupont> afaicr cwm took like 20 minutes to do my queries

05:20:56 <jmdupont> i can show you some examples

05:22:26 <jmdupont> but want to look for some database that can store this data

05:22:29 <jmdupont> and do adhoc queries

05:22:37 <jmdupont> it is annoying to have to reload all the time

05:22:56 <jmdupont> it would be nice if python would beable to cache the objects

05:23:01 <jmdupont> between executions

05:23:15 <jmdupont> there should be a way to freeze and thaw them!

05:23:15 <DanC> sandro was playing with swi-prolog (I think... that or another prolog) that has a fast persistent store that works with RDF.

05:23:26 <jmdupont> i never got that running

05:23:32 <jmdupont> i should try again!

05:23:36 <jmdupont> good idea

05:23:38 <jmdupont> have a new machine

05:23:56 <jmdupont> so, not so much cruft

05:24:36 <DanC> yes, swi-prolog is probably worth another look. meanwhile, I'll chalk this up as another data point supporting the idea that cwm, being pure python, is pretty easy to deploy/install.

05:25:02 <DanC> running a query over cyc.daml:

05:25:14 <DanC> $ time python ~/w3ccvs/WWW/2000/10/swap/cwm.py --rdf cyc.daml --filter=subc.n3 -n3 >,cyc-sub.n3

05:25:14 <DanC>

05:25:14 <DanC> real 0m43.022s

05:25:14 <DanC> user 0m38.761s

05:25:14 <DanC> sys 0m0.385s

05:25:46 <DanC> the query just finds all subClassOf triples: { ?C1 s:subClassOf ?C2 } => { ?C1 s:subClassOf ?C2 }.

05:25:56 <jmdupont> :algn a owl:Property;

05:25:58 <jmdupont> owl:Domain :boolean_type,

05:26:00 <jmdupont> :enumeral_type,

05:26:02 <jmdupont> i have lots of statements like that

05:26:13 <DanC> owl:Domain? what's that?

05:26:26 <DanC> was that supposed to be rdfs:domain ?

05:26:30 <jmdupont> each field has the domain and range specified

05:26:38 <jmdupont> well it might be buggy

05:26:41 <DanC> ok

05:26:51 <aharth> thanks danc for the numbers

05:26:56 <jmdupont> but iik, there is not anything to check it, is there?

05:27:10 <jmdupont> doh

05:27:21 <DanC> yes, there are tools to check your RDF data...

05:27:26 <jmdupont> there must be an online onto. checker

05:27:28 <jmdupont> ok

05:27:31 <jmdupont> i think i need to get one

05:27:36 <jmdupont> but cwm cannot do it?

05:27:50 <DanC> yes, cwm can...

05:28:09 <jmdupont> i have tried hard to use some of these axiomized owl checkers running

05:28:14 * DanC surfing for details...

05:28:15 <jmdupont> oh? is that new?

05:28:56 <jmdupont>http://unc.dl.sourceforge.net/sourceforge/introspector/extract-structure.n3

05:28:56 <DanC> May 2002. http://lists.w3.org/Archives/Public/www-rdf-interest/2002May/0034.html <- http://esw.w3.org/topic/RDFCheckingTools

05:28:56 <dc_rdfig> A: http://unc.dl.sourceforge.net/sourceforge/introspector/extract-structure.n3 from jmdupont

05:29:17 <DanC> A:|extract-structure.n3

05:29:17 <dc_rdfig> Titled item A.

05:29:31 <jmdupont> A:Primitive and inkorrekt attempt at writing rules from the data

05:29:32 <dc_rdfig> Added comment A1.

05:29:47 <jmdupont> A:might not even be korrekt owl

05:29:47 <dc_rdfig> Added comment A2.

05:30:06 * DanC wonders if extract-structure.n3 is like sniffSchema.n3

05:30:11 <jmdupont> cool

05:30:17 <dmiles> oops ii missed the prolog talking

05:30:26 <jmdupont> is that what it is called?

05:30:28 <jmdupont> great!

05:30:30 <jmdupont> hi dmiles

05:30:38 <jmdupont> i have here a new desktop

05:30:43 <jmdupont> for installing prolog

05:30:48 <dmiles> neat neat

05:30:56 * jmdupont installs swi

05:31:22 <dmiles> i am spending time in the win32 a while

05:31:22 <DanC> A:ah... seems similar to [http://www.w3.org/2000/10/swap/util/sniffSchema.n3|sniffSchema.n3]

05:31:22 <dc_rdfig> Added comment A3.

05:31:35 <dmiles> doing the ActiveX extensions to SWI-Prolog with .Net

05:31:37 <jmdupont> A:thanks danc!

05:31:37 <dc_rdfig> Added comment A4.

05:32:03 <dmiles> i made it so SWi could call arbirtrary java code.. now i am doing that with COM/OLE

05:32:04 <jmdupont> dmiles: no way!

05:32:09 <jmdupont> send me the code!

05:32:26 <jmdupont> please

05:32:28 <jmdupont> :)

05:32:29 <dmiles> one sec.. i'll get you links to my desktop

05:32:56 <jmdupont> thanks DanC!!!

05:33:14 <jmdupont> you are, as always a great help and good and patient teacher!

05:33:36 <jmdupont> sorry that i waste your time with my stupidity

05:34:14 <DanC> not at all; each of us has something to contribute.

05:34:26 <jmdupont> :)

05:34:26 <dmiles> look at http://lotus.daml.us/pvw/

05:34:35 <jmdupont> i am going to whip my rdf files into shape

05:34:44 <dmiles> its too disorganized for a release right now ;P

05:34:56 <jmdupont> and get the ontology corrected

05:34:58 <jmdupont> this checker, is it fast?

05:35:14 <jmdupont> dmiles tell me about the interfaces

05:35:26 <jmdupont> can we embed pnet into prolog?!!

05:35:29 <jmdupont> for syscalling out?

05:35:30 <dmiles> yes

05:35:35 <jmdupont> wow!!!!

05:35:35 <jmdupont> ok

05:35:36 <DanC> hold my hand just a little, dmiles? what to look at it /pvw/?

05:35:40 <jmdupont> that should grear

05:35:41 <dmiles> well if pnet can control java

05:35:46 <jmdupont> great

05:35:49 <DanC> what to look at in /pvw/? rather

05:36:02 <dmiles> then the pnet (linux) can use a bridge to java to talk to swi

05:36:17 <jmdupont> dmiles do you use jni?

05:36:27 <dmiles> yes

05:36:28 <jmdupont> i can replace the calls to jni with pnet

05:36:29 <DanC> no, using cwm to schema-validate stuff isn't fast... or at least: it wasn't before the big cwm optimization last Oct

05:36:31 <jmdupont> infact

05:36:33 <dmiles> swijbni is the project

05:36:39 <dmiles> erm swijni

05:36:47 <jmdupont> what about porting the jni interface to pnet

05:36:51 <dmiles> but thats prolog calling java

05:36:58 <jmdupont> ....

05:37:00 <jmdupont> that would gain alot of customers!!!

05:37:02 <jmdupont> :)

05:37:03 <dmiles> jpl exists for the other way arround

05:37:06 <jmdupont> i could sell that...

05:37:26 <jmdupont> we could..

05:37:26 <dmiles> i took jpl code though and made it two-way .. thats what started swijni

05:37:39 <jmdupont> wow

05:38:01 <dmiles> DanC .. checjking

05:38:13 * DanC can't figure out which end of http://lotus.daml.us/pvw/ is up

05:38:57 <jmdupont> + http://lotus.daml.us/pvw/PrologVb/SWIProlog.cls

05:38:58 <dmiles> well i had 5 versions spred out on zips and dirs ;( so i am spending time reassembing it

05:39:07 <jmdupont> that looks like a vb class?

05:39:16 <dmiles> i had a 1/2TB HD crash a little while ago

05:39:24 <dmiles> just after i backed up my whole life on it

05:39:31 <jmdupont> Function NextSolution(ByRef rgArgs() As Variant) As Boolean

05:39:31 <jmdupont> NextSolution = thePrologQuery.NextSolution(rgArgs())

05:39:32 <jmdupont> End Function

05:39:37 <jmdupont> wow

05:40:01 <dmiles> ProQuery was written for MS to use in network confuigurator

05:40:01 <jmdupont> you said about that before

05:40:08 <jmdupont> i need your help later with dd

05:40:15 <jmdupont> i have a drive that is awol

05:40:27 <dmiles> but the code got forgotten and i made it work with modern swi

05:41:15 <dmiles> so PVWE is only me restarting the the SWITLB (swi on com)

05:41:38 <dmiles> which oh i'll mv the src into there.. thats the oputput dirt

05:41:45 <dmiles> dir for switlb

05:41:56 <jmdupont> so you can also verify the ont

05:41:58 <jmdupont> in prolog

05:42:06 <jmdupont> i mean that is mostly what i need for the introspector

05:42:38 <dmiles> btw: TLB stands for typelib libraries./.. so any windows objects/code that has a typelib will work in it

05:42:38 <jmdupont> because enormous amounts of owl can be written to describe individual programs

05:43:06 <jmdupont> it is like an always finer mesh can be created to sample the rdf from the gcc

05:43:30 <dmiles> oh .http://lotus.daml.us/pvw/AvatarInstance/morphism.cls.html

05:43:48 <dmiles> look at that.. thats how i reverse engineer the interfaces

05:44:12 <jmdupont> the more external data you bring in (bug reports, cvs commits, documnentation, mailing lists) the more information we can add to the schema, until the schema describes the data itself even more accuratly, yet is completly specialized

05:44:26 * jmdupont looks

05:44:43 <jmdupont> very cool

05:44:57 <dmiles> so it introspects.. gets a prolog vierw of the object

05:44:58 <jmdupont> that could provide some inspiration for an swi interface

05:45:12 <jmdupont> to call prolog from dotgnu

05:45:56 <jmdupont> create_datum_ithing_method.NODEKIND = NODEKIND_FUNCTION

05:46:08 <jmdupont> the nodekind is what makes a function a function and a property a property

05:46:56 <jmdupont> dmiles you mean the vb introspects over vb

05:47:01 <jmdupont> or the vb introspects over the prolog object?

05:47:08 <dmiles> ok heres the swijni http://lotus.daml.us/pvw/swijni/

05:47:13 <dmiles> cp'd into there

05:47:42 <jmdupont> no way

05:47:56 <jmdupont> this is your server?!!

05:47:58 <jmdupont> doh

05:48:00 <jmdupont> daml.us!

05:48:02 <jmdupont> ?

05:48:53 <dmiles> yeah is was nauty

05:49:01 <dmiles> erm i's

05:49:52 <dmiles> the switlb i amfinsihing tonight.. but got distracted with httpd.conf

05:50:12 <jmdupont> great

05:50:26 <jmdupont> i am going to test this

05:50:28 <jmdupont> out

05:50:32 <jmdupont> is there a prolog validator for owl ?

05:50:39 <dmiles> triple20

05:50:42 <jmdupont> can cwmprolog run the validor scripts?

05:51:04 <dmiles> triple20 is alot of c code

05:51:13 <dmiles> werm so it is swi-prolog specific

05:51:16 * jmdupont looks

05:52:26 * jmdupont gets cvs

05:52:38 <dmiles> i wanted to be standard so i took CIAO-Prolog docs on thier java interface to make http://lotus.daml.us/pvw/swijni/javart.html

05:53:42 <dmiles> ok .. i will also put together a package of swijni .. which are you most interested in using switlb or swijni ?

05:53:59 <jmdupont> tlb?

05:54:05 <jmdupont> jni!

05:57:11 <dmiles> ok.. thats lower hanging fruit.. about 2 hours and swijni bugs will be worked out

05:57:33 <dmiles> i kinda wanna try getting it working on win323 anyways

05:57:40 <dmiles> erm on windows

06:01:10 <dmiles> if your on the httpd .. i am reorganising that dir before i start

06:02:08 <jmdupont> ERROR: /home/mdupont/development/triple20/Triple20/src/concur.pl:36: Syntax error: Operator expected

06:02:08 <dc_rdfig> Label ERROR not found.

06:02:14 <jmdupont> httpd?

06:03:07 <dmiles> who's concure.pl there is Jan's and Mine

06:03:28 <dmiles> erm and one in swi-library

06:03:44 <jmdupont> from the triple20

06:03:52 <dmiles> i mean there is triple20 in openmodality (maine) that doesnt use a UI

06:04:38 <jmdupont> you have tarfiles for swijni?

06:04:42 * jmdupont yawns

06:04:56 <jmdupont> 06 in the morn

06:04:56 <dmiles> will make some in about 40 minutes

06:04:58 <jmdupont> time to zzz

06:05:14 <dmiles> one sec.. while i look up line 36

06:05:26 <jmdupont> just put it on the server

06:06:10 <dmiles> meta_predicate is defined in library quintus i think

06:06:45 <dmiles> so before you run triple20 ... ytpy use_module(library(quintus)).

06:07:18 <jmdupont> ok

06:07:56 <jmdupont> i look into this tomorrow

06:07:58 <jmdupont> jmdupont is now known as md-zzzz

06:08:50 <dmiles> DanC .. i'll have have something more organised soon.. just caught in middle of trying to resolve too many differnt versions

06:12:54 <dmiles> ERROR: (c:/program files/prolog virtual worlds/library/actx_invokes.pl:52):

06:12:54 <dc_rdfig> Label ERROR not found.

06:12:54 <dmiles> Unknown message: activex_error(actx_create_object, 'CoCreateInstance error', '0x80070002')

06:13:05 <dmiles> getting closer ;P

06:58:11 <botlars> botlars is now known as larsbot

10:21:34 <D[a]vey> D[a]vey is now known as Davey

11:42:40 <danbri> A:|http://mozquery.mozdev.org/

11:42:40 <dc_rdfig> Titled item A.

11:42:47 <danbri> oops

11:42:48 <danbri> A:

11:42:48 <dc_rdfig>http://unc.dl.sourceforge.net/sourceforge/introspector/extract-structure.n3

11:42:49 <dc_rdfig>http://mozquery.mozdev.org/

11:42:50 <dc_rdfig> (1:jmdupont) Primitive and inkorrekt attempt at writing rules from the data

11:42:51 <dc_rdfig> (2:jmdupont) might not even be korrekt owl

11:42:52 <dc_rdfig> (3:DanC) ah... seems similar to [http://www.w3.org/2000/10/swap/util/sniffSchema.n3|sniffSchema.n3]

11:42:53 <dc_rdfig> (4:jmdupont) thanks danc!

11:43:09 <danbri> A:|A link from mdupont whose title I accidentally wiped out

11:43:09 <dc_rdfig> Titled item A.

11:43:13 <danbri>http://mozquery.mozdev.org/

11:43:13 <dc_rdfig> B: http://mozquery.mozdev.org/ from danbri

11:43:18 <danbri> B:|Mozquery

11:43:18 <dc_rdfig> Titled item B.

11:43:32 <danbri> B:Mozquery is "a Mozilla/Firebird extension for RDF coders. It's purpose is to allow fast an reliable RDF code development providing a set of tools to write and process RDF files content."

11:43:32 <dc_rdfig> Added comment B1.

11:43:47 <danbri> B:"Mozquery first tool is an ASRDF to RDF XML syntax converter allowing you to quickly and accuratly write RDF files from scratch. ASRDF (Abreviated Syntax for RDF) is a terse RDF statements notation using the outliner paradigm to structure RDF graph constructs."

11:43:48 <dc_rdfig> Added comment B2.

11:46:47 <danbri> B:Warning: installing it crashed my (recent Firebird) browser, and didn't give me the 'Tools/Mozquery' menu item it promised. I tried twice. But I am using a random nightly snapshot, maybe working with a release would be better.

11:46:48 <dc_rdfig> Added comment B3.

12:21:58 <Davey> Davey is now known as D[a]vey

14:20:21 <libby> hello masaka!

14:20:28 <masaka> hey libby

14:20:57 <masaka> how was meeting?

14:21:31 <libby> yesterday's meeting?

14:21:38 <libby> (today's is in 90 mins)

14:21:57 <libby> I'm really sorry they are so awkward for you to come to masaka :(

14:22:12 <masaka> yes. I think I'll stay here about an hour

14:22:30 <masaka> no that's ok

14:22:54 <libby> BLURB:yet another image annotation meeting, 2004-01-21 1600 UTC (today), here on #rdfig

14:22:54 <dc_rdfig> C: yet another image annotation meeting, 2004-01-21 1600 UTC (today), here on #rdfig from libby

14:23:46 <masaka> I'm glad to know the agreement on regionDepicts

14:24:35 <libby> C:this time focusing on getting together sample RDF documents for use with the W3 photos annotation project

14:24:35 <dc_rdfig> Added comment C1.

14:24:42 <libby> yes, very good news

14:24:55 * libby has to step out, sorry...

14:56:40 * DanC pores over sandro's prolog performance numbers, looking for the punch line, i.e. the apples-to-apples comparison...

14:57:07 <sandro> dja find it?

14:57:39 <sandro> 1.4s vs 43s

14:57:56 <DanC> ah, ok.

14:58:58 <sandro> With parsing being 1.2 of the 1.4s, and rule execution being about 0.01s

15:02:48 <sandro> I was thinking a rule-chaining test would be interesting. I have bits of very very dusty code for turning n3 into prolog; not sure if they're worth dusting off.

15:19:53 <libby> so, masaka, have you been hanging outon the japaese sw irc channel?

15:20:39 <masaka> japanese sw irc ? I don't know

15:20:56 <libby> danbri told me about it...

15:21:34 <masaka> Unfortunately, I've not seen that

15:21:45 <mmealling> :wonders if anyone is trying to create a foaf-based alternative to linkedin.com

15:21:58 * mmealling forgets which chat system he's on

15:22:12 <dmwaters> {global notice} Hi all! it appears that we just lost a main rotation server. We're currently working on the situation, and any other information will be given in wallops

15:25:24 <libby> here are the logs masaka: http://iwai.alib.jp/irc/sw-ja/

15:27:01 <libby> kota was translating/copying some of the image discussions yestreday into it: http://iwai.alib.jp/irc/sw-ja/log/2004/2004-01-20.html

15:28:38 <masaka> thanks. i'll say hello to them

15:32:46 <masaka> My w3cphoto message is not archived, so I put a uri of an example

15:32:54 <masaka>http://kanzaki.com/works/2004/imgdsc/040114_1443.rdf

15:32:54 <dc_rdfig> D: http://kanzaki.com/works/2004/imgdsc/040114_1443.rdf from masaka

15:33:10 <libby> yes I've asked danbri and gregelin about that. most mysterious.

15:33:21 <libby> thanks for that

15:33:45 <masaka> it' because of registered email address

15:33:49 <libby> ah

15:34:47 <masaka> This example demonstrate to show Polygon's bounding rectangle with XSLT

15:35:06 <libby> D:|An example image annotation RDF file from Masahide

15:35:06 <dc_rdfig> Titled item D.

15:35:16 <libby> D:<masaka> This example demonstrate to show Polygon's bounding rectangle with XSLT

15:35:17 <dc_rdfig> Added comment D1.

15:35:35 <libby> ooh! cool!

15:35:48 <danbri> nice!

15:36:22 <libby> maybe I shoudl forward your message to the archive masaka, so that we can refer to it

15:36:22 <masaka> It's a bit tricky to get max/min values with XSLT

15:36:31 <masaka> but it works, i think

15:36:49 <masaka> Yes, that would be nice

15:37:17 <libby> ok, I will do. did it get to the people on the webphoto list?

15:37:34 <libby> danbri, did you see masaka's message?

15:37:58 <masaka> I think so. I received my message from list

15:38:07 <libby> cool

15:38:07 * danbri in telecon

15:39:49 <libby> hi bwm

15:41:20 <masaka> libby, what kind of RDF do you expect for 'post' server ?

15:41:34 <libby> oh any kind at the moment

15:41:45 <libby> as long as it is valid rdf

15:42:07 <masaka> Is RSS that points image description RDF ok ?

15:42:13 <libby> I'm hoping that today we will be able to decide what form of RDF we want for the project

15:42:19 <libby> anything :)

15:42:30 <libby> so I'm only testing it at the moment, but it should be ok

15:42:40 <libby> there's no way to delete currently

15:43:13 <masaka> I post one RSS. I'm not sure whether it works

15:43:19 <libby> let me look

15:44:09 <libby> seems to: http://sw1.ilrt.org/discovery/2004/01/www2004/scutterplan.rdf

15:44:46 <masaka> yes, it is listed.

15:45:09 <masaka> but the RSS doesn't have any seeAlso

15:45:33 <masaka> it has several rss:item's with rdf:about

15:45:56 <masaka> Can scutter follow?

15:46:05 <libby> ah, at the moment, no

15:46:11 <libby> it's very simple

15:46:42 <masaka> should I include additonal seeAlso's ?

15:47:37 <libby> well, it won;t follw seealsos either...but I'm sure other tools will, so I think it would be a good ideas, yes

15:49:19 <masaka> so, do you think it's better to post the uri of image description itself for this demo server ?

15:52:56 <libby> rather than posting the rdf?

15:53:30 * libby would prefer if it people posted the url because then I'm not responsble for making sure the data doesn;t get lost

15:53:39 <libby> ...but I do have backups

15:54:05 <masaka> no, uri of img desc rdf rather than uri of an RSS that points descriptions

15:54:40 <libby> ah, yes, that's what I was exepecting

15:54:56 <masaka> I see. I'll post some later :)

15:55:03 <libby> ok, cool :)

15:57:03 * libby tries SELECT ?dd, ?url, ?image WHERE (an:annotates ?url ?image) (rss:description ?url ?dd) USING rss for http://purl.org/rss/1.0/ an for http://www.w3.org/2000/10/annotation-ns# in the query box

15:57:47 <GNUPredator> A:

15:57:47 <dc_rdfig>http://unc.dl.sourceforge.net/sourceforge/introspector/extract-structure.n3

15:57:48 <dc_rdfig> A link from mdupont whose title I accidentally wiped out

15:57:49 <dc_rdfig> (1:jmdupont) Primitive and inkorrekt attempt at writing rules from the data

15:57:50 <dc_rdfig> (2:jmdupont) might not even be korrekt owl

15:57:51 <dc_rdfig> (3:DanC) ah... seems similar to [http://www.w3.org/2000/10/swap/util/sniffSchema.n3|sniffSchema.n3]

15:57:52 <dc_rdfig> (4:jmdupont) thanks danc!

15:57:54 <GNUPredator> GNUPredator is now known as mdupont

15:58:06 <mdupont> ls

15:59:11 <libby> .time

15:59:12 <JibberJim> Hi

15:59:15 <libby> heh

15:59:20 <libby> is it time?

15:59:45 <JibberJim> 15:59:20 <libby> is it time?

15:59:59 <libby> :)

16:00:27 <libby> C:attending - Libby Miller

16:00:28 <dc_rdfig> Added comment C2.

16:01:04 * bryce regrets missing the meeting yesterday... he has been very ill. but he has read the mailing list and logs to catch up.

16:01:13 <libby> aw, poor bryce

16:01:26 <bryce> C:attending - Bryce Benton

16:01:27 <dc_rdfig> Added comment C3.

16:01:32 <JibberJim> C:attending Jim Ley

16:01:32 <dc_rdfig> Added comment C4.

16:01:36 <libby> anyone else here for an image meeting?

16:01:45 <bengee> C:attending Benjamin Nowack

16:01:45 <dc_rdfig> Added comment C5.

16:01:53 <libby> I don;t think mortenf can make it

16:02:09 <libby> C:regrets mortenf

16:02:09 <dc_rdfig> Added comment C6.

16:02:25 <masaka> C:[http://kanzaki.com/|masaka] attending for the first half

16:02:26 <dc_rdfig> Added comment C7.

16:02:31 <libby> hurrah!

16:02:43 <libby> Greg should be about somewhere

16:02:45 <kota> C:[http://iwai.alib.jp/irc/sw-ja/|kota] attending

16:02:45 <dc_rdfig> Added comment C8.

16:03:33 <libby> should probably wait a minute for greg

16:03:45 <JibberJim> Any Agenda/

16:04:00 <libby> sorry, no, not really

16:04:10 <libby> (greg's not on irc, so we should start)

16:04:24 <JibberJim> I'd like to agree a location for the namespace, and hear if there's any issues or agreement with yesterdays.

16:05:04 <masaka> I want to have space for coordinate

16:05:08 <libby> just the broad scope of writing down an RDF document that represnts what we want to say about photo anotations in this project

16:05:09 <bengee> me too

16:05:35 <libby> ok. did you want danbri here for the namespace discsussion jibberjim?

16:05:48 <JibberJim> Is space compatible with HTML image maps?

16:05:56 <bengee> don#t think so..

16:06:18 <masaka> no, but it's very easy to translate from space to commas

16:06:32 <JibberJim> Yep, I'd like the namespace to be on a respectable long term domain, so w3.org, daml.org or xmlns.com - and any others which I don't know about.

16:06:35 <danbri> I'm in teleocn still. One thing to say, we can if useful/wanted/etc have a namespace at w3c, alongside the RDF-exif thing...

16:06:37 <masaka> but opposite is not easy

16:06:59 <danbri> I think w3.org more respectable than xmlns.com ;) (but xmlns.com uris should last decades too, I hope)

16:07:14 <JibberJim> I'm happy with space.

16:07:17 <libby> hi greg

16:07:22 <GregElin> Hi libby!

16:07:29 <masaka> hello Greg

16:07:33 <JibberJim> Yes danbri, but w3.org might be too respectable :-)

16:07:35 <GregElin> Hi Masaka!

16:07:41 <libby> how quickly could something be placed at w3c do youthink danbri?

16:07:59 <libby> also - if it changed, would that matter?

16:08:00 <GregElin> I apologize for disappearing abruptly yesterday, Jim. I lost connectivity and could not get it back.

16:08:17 <JibberJim> no worries Greg, did you see the logs and conclusions, any comments?

16:08:55 <GregElin> Saw mortenf's great write up. Haven't read all logs yet.

16:09:02 <danbri> libby, pretty quick

16:09:38 <danbri> (noting that this isn't standards-track, etc. but development collaboration)

16:09:56 <libby> so if it changed that'd be ok (rdf calendar one does)

16:10:01 <GregElin> I think it is pretty good. I think the boundingBox as strong should will have advantages in speeding adoption...but you know how I feel. I think we are looking good.

16:10:06 <danbri> yes, would be following that same model

16:10:14 <libby> thanks danbri

16:10:28 <libby> any other opinions about a namespace place?

16:10:46 <bryce> +1 bounding box... has both foresight and easier adoption.

16:10:47 <GregElin> I did register w3photo.org

16:10:55 <JibberJim> I think many tools will include the bounding box, but I don't think we should look down on those that don't.

16:10:56 <GregElin> I have it up on a server.

16:11:04 <libby> nice greg

16:11:16 <GregElin> I agree, Jim.

16:11:32 <JibberJim> I have one problem which is the integer aspect. I'm not currently integer based in my tools.

16:12:26 <JibberJim> bryce, Masaka has shown that XSLT can do bounding box from polygon, I don't think adoption is a problem.

16:12:38 <bengee> how to deal with owl ontology vs rdfs doc? one namespace, two spec urls?

16:13:02 <GregElin> C:attending GregElin

16:13:03 <dc_rdfig> Added comment C9.

16:13:14 <GregElin> Excellent, Jim.

16:13:17 <libby> I think we can do it in one bryce, e.g. foaf is an owl ontlogy...

16:13:30 <bengee> yes. but with rdfs:Class...

16:13:37 * bwm sends regrets - in a meeting

16:14:12 <libby> I don;t know. it seeems like it would be a very bad thing to have two

16:14:20 <libby> perhaps it can be avoided

16:14:24 <danbri> rdfs:Class is OK by the OWL docs.

16:14:54 <bengee> w3photo spec could be OWL DL, I think.

16:15:04 <masaka> boundingBox would make creating annotation a bit annoying, i think

16:15:29 * GregElin Raises hand.

16:15:31 <bengee> OWLers will have a separate ontology somewhere. but maybe we could put in on an "official" server..

16:15:34 <danbri> Dublin Core is not OWL DL. I think we should think carefully about not using DC...

16:15:38 * libby wonders if we could hive off a subgroup to looks at the owl/dl/rdfs questions

16:15:55 <GregElin> I think a subgroup to look at these issues is a good idea.

16:15:59 <libby> I say this because it's impritive that we look at how to combine vocabs today

16:16:07 <GregElin> What is EASIEST and FASTEST to do a first version? One URL?

16:16:11 <libby> although these are imprtant issues

16:16:13 <bengee> agreed. sorry,

16:16:25 <libby> no, it's ok, it's tricky question

16:16:36 <GregElin> (I think Dublin Core will be important in long run...want to include digital library)

16:16:54 <bengee> dc can be used in an OWL DL ontology.

16:17:05 <libby> bengee, do you know a lot about owl dl?

16:17:20 <bengee> well.

16:17:38 <bengee> a little bit. I'm working on an OLW editor at the moment

16:17:55 <GregElin> I think we are, at this point in time, always discussing two things: (1) what do we do as default to get something up immediately and (2) what do we want to do in time for www2004 in May and thereafter?

16:18:33 <libby> perhaps if we could get jim and mortenf to devise an rdfs schema, then we could get togther to see what an owl dl version woudl look like

16:18:39 <libby> maybe they already did that

16:18:47 <bengee> I could try to build an integrating (relevant foaf, dc, img annot) OWL ontology and publish it somewhere..

16:18:59 <bengee> the current schema is owl dl.

16:19:08 <GregElin> We will do more, IMHO, for the cause of image annotation & semantic web if start with things simple enough that the can be modified and avoid people having to get religious to either get involved, or create own variation.

16:19:13 <libby> ok, if you like bengee, that souns like a way forward

16:19:30 * libby would like to decide on some ways of combining those vocabs here today

16:19:35 * GregElin climbs down from soapbox.

16:19:53 <libby> :)

16:20:13 <JibberJim> We really need Jen here to see how the proposal from yesterday matchs up with their stuff.

16:20:19 <libby> I like to have example documents ot look at

16:20:20 <GregElin> If possible, Libby, great!

16:20:52 <libby> sorry, I've not handled this well - I assumed we would be talking about different vocabs today

16:21:11 <libby> hm, golbeck is logged in

16:22:14 * libby wonders about looing at http://esw.w3.org/topic/ImageDescriptionUserInterface section 1.2

16:22:24 <libby> - a lst of things we wanted to talk about

16:22:25 <masaka> I have some examples, but what kind of example do you want ?

16:23:22 <libby> I think we need to decide what it is we would like in the document e.g. creative commons license, date annotated, date taken, creator etc

16:23:38 <libby> i.e. having a schema for annotating parts of an image only gets us part of the way there

16:24:22 <GregElin> Define "there"? Ontology for whole image (whole + parts)?

16:25:03 <libby> no, I meant: once we have an ontology for annotating parts of an image, that's great, and we've got a long way

16:25:26 <libby> but we need to add some other vocabulary tersm in the way we discussed for the UI

16:25:35 <libby> which probably won;t be in out parts of images ontology

16:25:40 <libby> s/out/our/

16:26:10 <GregElin> Ah. Right, libby.

16:26:27 <libby> I want to have something like this: http://swordfish.rdfweb.org/discovery/2004/01/www2004/path-eg.rdf as a sample document to show people

16:26:41 <libby> sorry if I'm being obtuse!

16:26:49 * libby fears she has bored everyone to death

16:27:50 <masaka>http://kanzaki.com/works/2004/imgdsc/annot-gallery-e.rdf

16:27:51 <dc_rdfig> E: http://kanzaki.com/works/2004/imgdsc/annot-gallery-e.rdf from masaka

16:28:12 <JibberJim> The annotea "annotator" properties and a foaf:Person seem good to buy for the person doing the annotating.

16:28:15 <GregElin> That looks pretty good to my rdf newbie eyes. Doesn't that example just need to be vetted against some of the other discussions and notes for descrepancies to be resolved?

16:28:15 <masaka> most examples are simple rectangle and dc: vocab

16:28:33 * libby checks out masaka's example

16:28:42 <libby> yes I think so GRegelin - it may not e a big deal

16:29:59 <libby> so masahide uses dc title and description and date

16:30:04 <libby> dc:creator as a string

16:30:05 <GregElin> I think we really have all the pieces...and most of the issues identified.

16:30:37 <libby> that's great! so which properties shall we use?

16:30:54 <masaka> those strings are easy for annotators to input

16:31:09 <libby> for creator masahide, or all of them?

16:31:29 <JibberJim> foaf:Person is more useful if we can do it.

16:31:38 <masaka> well, some basic annotation

16:31:39 * JibberJim really needs to talk in the right place..

16:32:17 <masaka> of couse, wordnet or some other classes are useful

16:32:22 <libby> yes re Person. Also creative commons is a bit yucky,requiring date of birth and name

16:32:48 <GregElin> date of birth and name? Can you elab, libby?

16:33:08 <libby> I can update this file http://swordfish.rdfweb.org/discovery/2004/01/www2004/path-eg.rdf according to the decisions we've alreday made and send to the list if people prefer

16:33:44 <libby> well, it just seems like (a) that's not a unique key for people, and (b) people might not want to make their date of bitrth public

16:33:53 <libby> but - it's how they specify to do it

16:33:55 <JibberJim> CC requires d-o-b - presumably for the N yeard after death aspect.

16:34:04 <libby> ah, that's way then

16:34:09 <libby> s/way/why/

16:34:34 <GregElin> Ah...unique key for people. A whole issue in itself.

16:34:37 <libby> I guess we could stick a mbox_sha1sum in there

16:34:40 <libby> yes indeed

16:34:56 <GregElin> Which cc, for the Creator?

16:35:24 * libby also wanted to ask if we would still accept photos taht only say who is in them, and not where they are in the picture?

16:35:49 <GregElin> Yes, absolutely, libby.

16:35:52 * bengee thinks we should

16:35:57 <libby> for both I guess - creator of metadata and creator of the photo

16:36:14 <libby> perhaps we should make it mboxsha1sum optional

16:36:18 <bryce> yes, someone else could finish the annotation.

16:36:20 <libby> cool

16:36:47 <GregElin> right, bryce.

16:36:54 * libby wonders about sayoing the photo depicts an event

16:37:08 <libby> i.e. how to identify the event, basically

16:37:08 <bryce> perhaps we offer the chance to finish/correct annotations?

16:37:45 <bryce> at the conference, by time and room number to ID the event.

16:37:47 <libby> I think we certainly need to be able to have multiple annotaions on the same photo bryce

16:38:17 <libby> what about the conference itelf? is homepage suffcient?

16:38:35 <libby> so we perhaps need room number in the ontology?

16:38:41 * libby checks if it's there

16:39:42 <masaka> sorry, i should go away now

16:40:38 <bryce> what do we do when annotations conflict? also, does each annotator draw their own region shape, or can they use someone else's?

16:40:51 <libby> bye masaka, thanks for coming

16:41:02 <masaka> bye all :)

16:41:08 <bengee> bye masaka

16:41:21 <GregElin> Good questions, all, Bryce.

16:41:26 <libby> I don't know bryce :)

16:41:29 <GregElin> bye Masaka, thanks for making it.

16:41:56 <libby> in the wild world, there will be many annotations for the same picture

16:42:12 <libby> within the main UI, I guess you could restrict it

16:42:18 <JibberJim> I would suggest annotating other regions is a good thing, and don't see anything wrong with using rdf:ID to allow it.

16:42:20 <GregElin> I'd prefer to return to that question after first version is up.

16:42:32 <bryce> right, greg.

16:42:36 <JibberJim> UI's are of course a seperate issue.

16:42:46 <GregElin> I think baskically it should be easy to turn on/off annotations belonging to different people.

16:42:55 <GregElin> +1 Jim.

16:43:20 <JibberJim> It matters on the RDF level Greg, as if the Region is just a bnode, other people can't deal with it.

16:43:36 <libby> in the mindswap conferece ontology, there is hasLocation - but I can;t tel if that's a string or a resource or what ... my owl sucks

16:44:15 <libby> I think all the vocabs we've seen give the region a uri right?

16:44:45 <GregElin> A related question is if and how a person excercises any control or special privileges regarding one's photo or one's appearance in an image.

16:45:56 <libby> I think we could sidesttep that issue a bit - get them to email one of us and promise todeal with it swiftly: see how bad the problem is

16:46:29 <JibberJim> Masaka's example earlier didn't.

16:46:31 <GregElin> Jim, you'll have to explain that issue a bit further. Wouldn't person A as the creator of an annotation excercise some control if they want other people to annotate their annotation (or have to create own tool to point at person A's annotation)?

16:46:54 <libby> ah, ok jim, didn;t notice that

16:47:51 * GregElin is buying online time as he travels. His time expires at 5 minutes after the hour.

16:48:18 <libby> I think jim just means that a tool cannot find all the regiosn in an image and add to them if they do not have identifiers

16:48:24 <libby> - this is an RDF thing

16:48:47 <JibberJim> Well yes they do, if they don't give it a URI, no-one else can say anything about it Greg, if they do, people can.

16:49:02 <libby> he suggests using rdf:ID which would enable tools to point to an annotation from another file

16:49:24 <libby> - very much like a fragment id as we were discussing a few days ago gregelin

16:50:07 <GregElin> Cool. Doesn't that answer the question, then at some level? Frankly, I think the fragments should be addressable via URL.

16:50:21 <GregElin> In fact...i have a proof of concept...Let me see if it working.

16:50:59 <libby> it works for me - perhspa needs a followup mail to morten'd though

16:51:46 <GregElin> umm...want to point to a URL...but don't want to include in log. How do I do that?

16:51:57 <libby> at the strt of a line

16:52:09 <GregElin> The photos are a set of photos that while aren't private...aren't for broad consumption.

16:52:16 * libby volunteers to said mail to list

16:53:58 <libby> I get binary stuff in the last one but the rest looks great

16:54:06 <GregElin> Kind of like the URL of partname@imagefilename.jpg, but am not yet wedded to any syntax.

16:54:24 <GregElin> You are suppose to get the binary, libby.

16:54:29 <libby> cool :)

16:54:31 <GregElin> (At least right now!)

16:54:52 <libby> ok well I think we are agreed that the fragments should have unique ids then, yes

16:54:55 <libby> ?

16:55:17 <bengee> agreed.

16:55:28 <libby> hurrah!

16:55:30 <GregElin> I think so. I think that is what could really open up possibilities.

16:56:04 <libby> ok, well greg has to go in 10 mins. yes?

16:56:15 <bryce> also, makes for easier "proofreading" by placing fragments of pix side by side... can spot mistakes.

16:56:25 <JibberJim> MUST/SHOULD? or are encouraged too?

16:56:45 <libby> shall I try and create a few documents that we would accept including descripton, title etc?

16:56:48 <GregElin> I do, but feel free to continue. I expect to be back online in about 30 minutes after that.

16:56:55 <libby> MUST?

16:56:57 <GregElin> Yes.

16:57:04 <bengee> no. ;-)

16:57:13 <libby> heh

16:57:18 <GregElin> Sorry...my yes was to libby's doc question.

16:57:30 <GregElin> must have a URL? No.

16:57:32 <libby> SHOULD then?

16:57:34 <bengee> I would prefer "encouraged to"

16:57:41 <GregElin> encouraged, yes.

16:57:55 <libby> cool

16:58:46 <libby> C:action libby create some sample RDF documents summarising what's been said here today and send to the list

16:58:46 <dc_rdfig> Added comment C10.

16:59:32 <libby> C10:action libby create some sample RDF documents summarising what's been said here today and send to the list for discusssion there

16:59:32 <dc_rdfig> Replaced comment C10.

17:00:10 <libby> C:action bengee try to build an integrating (relevant foaf, dc, img annot) OWL ontology and publish it somewhere.

17:00:10 <dc_rdfig> Added comment C11.

17:00:41 <libby> actually, bengee, maybe try with just the image annotation stuff? might be easier to discuss to start with?

17:00:55 <bengee> yes.

17:01:03 <JibberJim> www.w3.org/2004/01/image# or xmlns.com/images/0.0.1/# or ....

17:01:26 <bengee> libby: I can create accounts so that we can all hack the ont..

17:01:29 <libby> then we can do a nice comparison of an rdfs and owl dl version and see where the differences lie and what we can do about it

17:02:11 <libby> C:action bengee try to build OWL DL verson of the parts of image annotation ontology and publish it somewhere.

17:02:11 <dc_rdfig> Added comment C12.

17:02:16 <libby> arse

17:02:20 <libby> C11:action bengee try to build OWL DL verson of the parts of image annotation ontology and publish it somewhere.

17:02:20 <dc_rdfig> Replaced comment C11.

17:02:26 <libby> C12:""

17:02:26 <dc_rdfig> Deleted comment C12.

17:02:29 <GregElin> Who can help me adopt the Fotonotes interface to right RDF instead of the minimal XML it does now?

17:02:52 <GregElin> I have no code for writing RDF or Foaf.

17:03:01 <libby> cool bengee

17:03:15 <libby> you don;t really need special code greg

17:03:32 <libby> once we know what the rdf should look like, should be easy

17:03:58 <GregElin> I know I don't need special code...but I'd rather not write from scratch. just looking for some php code or a little aid.

17:04:17 <libby> how do you write the xml at the moment greg?

17:04:23 <libby> have you got an example?

17:04:23 <JibberJim> You can treat it as simple XML when generating Greg.

17:04:43 <libby> yes uindeed and that's often betweer than using a serializer

17:04:50 <libby> - prettier results

17:04:53 <GregElin> create an array...then have a function that converts the array to xml. Another function writes that result to the jpeg header (or elsewhere)

17:04:56 <libby> better

17:05:14 <GregElin> function...I mean method. That part of the PHP code is classes and methods.

17:05:54 <GregElin> So I'm just missing the "turn into RDF" and "parse from RDF" methods.

17:06:46 <libby> I suggest treating it as xml in both cases

17:08:47 <libby> there is some RDF php code about though I think

17:09:57 <libby> D[a]vey has been working on some, but I don;t know if it is completely generic

17:10:00 <libby> darn

17:11:32 <bryce> maybe i'm missing something, but i really don't think it's that hard to output rdf. you just take some example rdf and determine what loops you need to spit it out...

17:11:40 <libby> yes indeed

17:11:45 <libby> harder to read it though

17:11:58 <libby> unless you're controlling the format

17:12:33 <libby> although masaka does a very good job with xslt. though that's not usually advisable

17:12:59 <bryce> i think it will be easy for greg, once the example rdf is available. the pieces will fall into place.

17:19:53 <danbri>http://www.w3.org/MarkUp/html-spec/html-spec_7.html#SEC7.6

17:19:53 <dc_rdfig> F: http://www.w3.org/MarkUp/html-spec/html-spec_7.html#SEC7.6 from danbri

17:20:00 <danbri> F:|Imagemaps in HTML2

17:20:00 <dc_rdfig> Titled item F.

17:20:13 <danbri> F:"If the ISMAP attribute is present on an IMG element, the IMG element must be contained in an A element with an HREF present. This construct represents a set of hyperlinks. The user can choose from the set by choosing a pixel of the image."

17:20:13 <dc_rdfig> Added comment F1.

17:22:32 <danbri> F:See also [http://www.w3.org/TR/html401/struct/objects.html#h-13.6|image maps] in HTML4.01 spec., which provides client-side image maps too.

17:22:32 <dc_rdfig> Added comment F2.

17:23:08 <danbri> F:"Client-side: When a user activates a region of a client-side image map with a mouse, the pixel coordinates are interpreted by the user agent. The user agent selects a link that was specified for the activated region and follows it." (hmm re 'mouse')

17:23:09 <dc_rdfig> Added comment F3.

17:23:20 <D[a]vey> libby: hmm?

17:24:09 <danbri> F:HTML clientside imagemaps allowed shape as one of: default (entire region), rect, circle, poly (polygonal region).

17:24:09 <dc_rdfig> Added comment F4.

17:24:18 <libby> hi davey. just wondered if your php tools wee generic for RDF or specific to foaf

17:25:00 <D[a]vey> *my* tools are specific to FOAF, but they're using a generic RDF API, google for "RDF API for PHP" and you'll come across "RAP"

17:25:23 <libby> ah, cool. thanks :)

17:25:29 <danbri> F:Other background: see [http://www.faqs.org/rfcs/rfc2396.html|rfc2396] section 4.1 for b/g on URI fragment Ids.

17:25:29 <dc_rdfig> Added comment F5.

17:25:32 <libby> did you want to ask me something earlier?

17:33:06 <D[a]vey> libby: yes, can I PM?

17:33:58 <libby> help yourslef

17:33:58 <workbench> There's no help for topic 'yourslef'.

17:34:03 <libby> heh

17:34:04 <D[a]vey> lol

17:34:12 <D[a]vey> D[a]vey is now known as Davey

17:45:34 <wkearney> for those that might be interested, http://www.rsswinterfest.com, an IRC channel here #rss-winterfest and a realaudio stream: http://webcast.on24.com/media/news/corporatevideo/events/4005/1_rmaudio.ram

17:46:40 * libby guesses image meeting has ebnded, sorry for not saying so expicitly

17:46:44 <libby> ended even

17:48:24 <danbri> adjourned.

17:49:04 * danbri feeds 'adjourned' to Ihttp://honyakuinfoseek.infoseek.co.jp/amitext/indexUTF8.jsp and then back-translates the Japanese into English: "It was postponed." Hmm, almost.

18:33:42 <bryce> bryce is now known as bryce_

18:55:01 <Pattty> is anybody here?

18:56:53 <jeen> sort of

18:58:07 * swh is away: home

19:50:13 * bengee wonders if he could ask danbri a question... (re OWL/DC issue)

19:50:36 <JibberJim> Just chuck it out bengee - what's the worst that can happen?

19:50:43 <bengee> ;-)

19:51:22 <bengee> I just think that DC could be "ported" to OWL DC space. but mybe I missed something.

19:51:47 <bengee> perfect typo. OLW DL.

19:52:30 <danbri> OWL DL doesn't like it when a vocab allows a property to point at either non-literal resources as well as at literals.

19:52:35 <danbri> DC allows that (dc:creator etc).

19:53:06 <danbri> you could impose a DL convention on DC in a particular dataset (or try to) but it wouldn't make DC in general OWL DL-happy.

19:53:40 <deltab> "either ... as well"?

19:53:45 <bengee> ok. we would have to restrict them in the image vocab to DatatypeProps.

19:54:11 <bengee> I see the issue.

19:54:54 <danbri> that was a messy sentence, sorry

19:54:57 <danbri> i mean if you do

19:55:31 <danbri> <dc:creator><Person/></dc:creator> and <dc:creator>foo</dc:creator> in the same dataset, you're not in OWL DL (I think).

19:55:44 <bengee> that's right.

19:55:58 <danbri> There _is_ some business with 'annotation properties' that might be used to hack out of this, but I'm skeptical it's really applicable.

19:56:04 <JibberJim> This is the creator of the image? why not just create fred:photographer?

19:56:31 <danbri> just a general problem with trying to make DC into an OWL DL vocab.

19:56:48 * JibberJim has lots of problems with DC :-)

19:57:12 <bengee> annotation properties don't work either. they may be ignored by reasoners.

19:57:38 <bengee> tried to get an answer on that from the webont wg weeks ago.

19:57:49 <bengee> no reply so far ;-(

19:58:14 <danbri> its somewhat; a pragmatic compromise, result of OWL really being two languages.

19:58:18 <bengee> you must not define domain and range for AnnoProps..

20:00:32 <bengee> so you'd say if a defined an <owl:DatatypeProperty rdf:about="&dc;creator" /> that could lead to big trouble?

20:00:52 <arnarl> arnarl is now known as arnarl|away

20:01:06 <danbri> it'd be untrue; dc:creator just _isn't_ an owl:DatatypeProperty.

20:01:23 <bengee> well, it's not defined in OWL space so far..

20:01:52 <danbri> If you fed that untruth to a system that only saw data consistent with that assumption, and took care not to feed it anything else (eg. angry denials from DC folk encoded in OWL ;) you'd maybe be able to make some tools do stuff.

20:02:11 <danbri> you shouldn't have to redefine all RDF vocabs to be OWL vocabs. They're OWL-ready already.

20:02:14 <danbri> That's the intent anyhow.

20:02:38 <bengee> hm, OWL full only..

20:03:03 <danbri> I'm not going to go through all my schemas and change rdfs:Class to owl:Class; DL toolkit creators are quite capable of figuring out that mapping...

20:03:15 <danbri> the Lite-like subset of Full, yes.

20:10:50 * bengee 's toolkit can't read rdf schema docs. :-( he apparently should make it to be able to..

20:10:59 <danbri> yes please!

20:13:08 <bengee> danbri: do you know of any pointers concerning this? there is no best practice wg yet, and any owl onto I know so far redefines rdfs terms.

20:13:37 <danbri> well, all the RDFS vocabs you know which don't... they're also OWL ontologies.

20:13:50 <bengee> owl ref: "for any term used, there has to be a typing triple in the doc"

20:13:52 <danbri> That's why we spent so much time having those two groups talk to each other re architecture, layering etc.

20:14:14 <bengee> which means I'd have to do an owl:imports

20:15:11 <yonderboy> is there a info page for the logger bot?

20:15:54 * bengee may ask holger knublauch, how protege handles rdfs ontologies.

20:21:12 <Davey> cool, I've successfully extended my XML_FOAF_Parser to parse the custom properties in the PaWS FOAFs :)

20:21:30 <bengee> what would dc:creator than be in OWL Full space? an owl:ObjectProperty, I guess.

20:23:11 <bengee> thinking loud: what if I defined an owl:ObjectProperty w3photo:creator, being an rdfs:subPropertyOf dc:creator

20:23:36 <bengee> with w3photo:creator rdfs:range foaf:Agent..

20:24:06 <bengee> wouldn't that be consistent?

20:25:00 * bengee has to think about those hybrid dc properties..

20:33:23 <danbri> each w3photo:creator triple would also imply a dc:creator triple, so you'd be back where you started

20:34:15 <bengee> yes. but this time with objProps. owl tools could read them rdfs tools, too

20:34:37 * DanC waves

20:35:20 <bengee> if the owl photo tools ignore dc literals...

20:36:39 <bengee> but I think we would like to use literals for creator etc. so it doesn't really solve the practical problem..

20:37:22 <JibberJim> No, we don't want literals for creator!

20:37:23 * bengee wonders if there will _ever_ be widespread OWL DL ontologies..

20:37:28 <bengee> no?

20:37:35 <JibberJim> A FOAF person with just a name is better than a literal surely?

20:37:54 <bengee> true

20:38:13 <JibberJim> consuming code will have to deal with FOAF people with most of us I'd've thought - we want those IFPs!

20:38:43 <JibberJim> so it will be easier if they are guaranteed such a Person (image:Photographer or whatever has range foaf:Agent)

20:39:19 <bengee> what about dc:title. that's a DTProp, no?

20:39:33 <bengee> dc:description as well..

20:39:55 <JibberJim> Yeah, as I said before I had the problems with dc :-)

20:40:04 <Davey> Davey is now known as D[a]vey

20:40:08 <bengee> :-)

20:40:53 <bengee> but we could re-use those, ignore dc:creator and create w3p:Photographer, w3p:Annotator, ...

20:41:01 <JibberJim> not that I'm in the OWL world enough to care.

20:41:59 <bengee> it's a pity with dc. it's so widespread. foaf2owlDL is easier.

20:42:21 <bengee> main problem is the invFunProp mbox_sha1_sum

20:42:22 <danbri> do you have a summary of what's needed there?

20:42:33 <danbri> 'cos it is literal valued?

20:42:38 <bengee> yep.

20:42:48 <danbri> ah. well I'm gonna be stubborn ;/

20:43:08 <danbri> that's a bug in OWL DL. Literal valued properties that are uniquely identifying are really really useful.

20:43:16 <danbri> eg. company stocktickers

20:43:27 <bengee> oh yes they are. any DB uses them.

20:43:30 <JibberJim> Absolutely! that's a big bug then IMO!

20:43:41 * JibberJim didn't realise that.

20:43:45 <bengee> well, it's because of DL reasoning.

20:44:10 <danbri> I don't expect web crawlers to do much DL reasoning

20:44:12 <bengee> it's not a bug, it's a feature ;-)

20:44:35 <danbri> I expect DL reasoners to do their own input-filtering cleanup, rather than restrict the rest of us from handy scruffyness...

20:44:38 <danbri> is that unreasonable?

20:45:19 <bengee> I guess DL is more for intranet projects. not the www..

20:46:08 <bengee> problem is, that most people think, OWL is somehow "better", which was never stated anywhere..

20:46:16 <bengee> oops OWL DL

20:47:10 <bengee> I'm using OWL Full, but try to keep it as close to DL as possible..

20:48:11 <bengee> don't see DL problems concerning the other FOAF terms.

20:50:40 <danbri> dnaChecksum value space might also be literals, I forget if that got decided. It's a joke one anyhow.

20:51:05 <danbri> although with biometric passports coming in, ...

20:51:17 <bengee> I once saw an OWL DL version of the foaf spec (at mindswap?). all they did was replacing rdfs:Class with owl:Class an commenting out datatype's InvFunProp typing

20:51:51 <danbri> interesting

20:51:55 <bengee> yes. and not to forget the soon to be included foaf:rfid?

20:52:38 <danbri> I didn't think I'd mentioned that yet :)

20:52:53 <bengee> bengee: the reasoner says you are on sale today.

20:52:56 <danbri> isn't clear what values it should take...

20:53:53 <bengee> ...underspecified domain ;-)

21:00:03 <bengee> Jim: we are going to annotate the annotation somehow, right? (foaf:maker of annotation, annotation = instance of foaf:Document?)

21:02:18 <JibberJim> I'd've thought foaf:maker of the document, or there's an annotator property in annotea IIRC.

21:03:01 <JibberJim> I don't know much about it though - the image description part I've thought a lot more on.

21:04:25 <bengee> the example at http://swordfish.rdfweb.org/discovery/2004/01/www2004/long-eg.rdf uses maker for both img and doc. I'll adopt that for a first version, ok?

21:06:39 <JibberJim> seems fine to me.

21:18:33 * bengee discovers that protege mentions rdfs (import?) support for the OWL plugin. feature request with second but lowest priority though.

21:28:12 <bengee> bye. danbri: thanx for the conversation/help.

21:52:02 <JibberJim> mortenf, I was wondering if perhaps the space rather than the comma syntax would be better, masaka could do xslt on the space, but more complicated with the comma?

21:52:22 <mortenf> i just now sent a message to the list on that

21:52:39 <mortenf> .. supporting our decision because of image maps

21:53:01 <mortenf> it's not that i don't agree, but...

21:53:19 <JibberJim> Which is what I was commenting on!

21:53:35 <mortenf> hmm

21:53:46 <JibberJim> I guess the question is, are the people generating image maps simply, more important than the people processing the data simply?

21:54:17 <mortenf> i was kinda figuring that image maps was the simple solution, thus we shouldn't make the simple solution more complicated...

21:54:54 <mortenf> but if nobody (that is currently on the list) is going to do image maps anyway...

21:55:17 <JibberJim> Well the audience at the moment doesn't appear to use imagemaps apart from danbri.

21:55:20 <mortenf> i like spaces as well, keeps it in line with the xml approach

21:55:26 <mortenf> danbri, where?

21:55:39 <JibberJim> the original one!

21:55:44 <mortenf> ah, yeah

21:56:24 <JibberJim> hmm, trying to find it made the mistake of thinking "danbri chaals pub" would turn it up quickly.

21:56:38 <mortenf> heh

21:57:26 <JibberJim> seems to be http://www.w3.org/2001/08/rdfweb/foafwho/imagemap/

21:57:38 <JibberJim> - danbri charles pub did get it!

21:58:53 <JibberJim> I'm SVG/javascript Greg is PHP I believe, masaka likes XSLT, the mindswap folk are java I think, so if XSLT can do the conversion I can't see what language anyone will use that can't do s/ /,/gm for the conversion.

21:59:11 <mortenf> hmm, ok, and since that's done by xslt (no other image map usage examples, right?), i'd be okay with spaces anyway

22:00:13 <mortenf> could reply to me on the list (so i don't have to talk to myself)?

22:00:27 <JibberJim> sure as I have proper bandwidth...

22:00:31 <JibberJim> logger, pointer

22:00:31 <JibberJim> See http://www.ilrt.bris.ac.uk/discovery/chatlogs/rdfig/2004-01-21#T22-00-31

22:01:23 * mortenf would like for "logger, pointer" to take an optional numeric argument, N, so that the pointer could be to N minutes earlier...

22:03:34 <mortenf> be sure to specify in the mail where the spaces should now be placed, re circle etc.

22:28:15 <libby_jp> libby_jp is now known as libby

22:32:52 <DanC> libby? I'm having trouble with the new dependencies in ical2rdf.pl, so I'm looking at the code... gottaminute to help me clean it up a bit?

22:33:12 <libby> sure

22:34:31 <libby> which ones?

22:35:02 <DanC> I'd rather take a "shared ownership" approach, leavning names out of the inline comments.

22:35:11 <libby> ok

22:35:37 <libby> I didn;t want to take too many of masahode's comments out before checking it all worked

22:35:46 <libby> it's a bit messy though

22:35:59 <libby> are you going to update it with olivier's code?

22:36:07 <DanC> wasn't planning to.

22:36:22 <DanC> right now, I'm trying to find a way to make the jcode and SHA modules optional

22:36:30 <libby> ok, it's on my list to do

22:36:39 <libby> right, make sense

22:36:56 <danbri> why is sha1 needed?

22:37:04 <libby> (but i'm afraid my perl's not greatto put it mildly)

22:37:14 <DanC> sha1 is used for turning prodids into URIs... I'm having 2nd thoughts about that.

22:37:17 <libby> for x-property namespace generation

22:37:34 <DanC> what do you use to edit perl, libby?

22:37:42 <DanC> your code doesn't seem to be indented.

22:37:47 <libby> um, nano

22:37:53 * libby looks embarrassed

22:38:13 <DanC> .google nano editor

22:38:14 <datum> nano editor: http://www.nano-editor.org/

22:38:26 <deltab> it's a free version of pico

22:38:32 <libby> it's the default pico-eque editor on debian

22:38:40 <libby> -esque

22:38:47 <DanC> does it support indentation?

22:38:53 <mortenf> i'd like to use nano, but it doesn't work as reliably as pico...

22:39:02 <deltab> it has very few features

22:39:06 <mortenf> (yet, perhaps)

22:39:12 <libby> I don;t think it does no

22:39:22 <deltab> there's a list at the bottom of the screen with most of the commands

22:39:27 <danbri> Norm's http://www.apps.ietf.org/rfc/rfc3151.html does XML public identifiers -> URNs, wonder if something similar could be done for prodids

22:39:30 <libby> also I sometimes use pico

22:39:58 * libby looks embarrassed

22:40:08 <deltab> it's a simplest-possible kind of editor

22:40:57 * JibberJim uses nano

22:41:18 <DanC> hmm... there's not much of a reason to be embarrassed, but can I take that as a sign that you agree that indentation of code is handy/good/worthwhile?

22:41:46 <libby> people have bugged me about it before but somehow I just didn;t get around to learning emacs etc

22:41:48 <libby> so yeah

22:42:00 * DanC hunts for an outboard indentation tool...

22:42:12 <JibberJim> astyle?

22:43:02 * mortenf does his indentation manually (mostly)...

22:43:24 <deltab> libby: heh, those are opposite ends of the specturm

22:43:39 <danbri> anyone here use Eclipse?

22:43:48 <DanC> well, are you willing to do manual indentation to match emacs's conventions, libby?

22:43:49 <deltab> as is Eclipse :-)

22:44:03 * DanC wonders if there's documentation of the perl community's consensus on indentation

22:44:04 <libby> sure, I thought I had actually, so maybe I did it wrong

22:44:24 <DanC> well, sub sha1 isn't indented at all.

22:44:40 <DanC> nor creatxns

22:44:47 * DanC gets bad FORTRAN flashbacks looking at it

22:45:23 <libby> I probably just forgot

22:45:29 * D[a]vey gets lost in code for his FOAF stuff

22:45:36 * DanC finds discussion of perl indentation... http://dbforums.com/arch/95/2002/9/491008

22:45:36 <libby> want me to update it DanC?

22:45:38 <danbri> what'd be best tab/indent setting?

22:45:48 * danbri investigating Eclipse's built in editor

22:46:23 <danbri> 'set tabs width = <wisdom_of#rdfig> chars' ...

22:46:27 <DanC> I have the code open in emacs; dunno if it's easier for you or me to do it, libby.

22:46:51 <DanC> emacs will do it for me, so I guess it's easier for me to do it.

22:47:28 * DanC proposes a ~20 minute pass over the code with libby

22:47:29 * mortenf likes tab stops of 3 spaces, but is getting converted into 2 spaces by dajobe...

22:47:31 <DanC> got time for that?

22:47:35 <dajobe> mortenf: logger is practically stateless, it just knows the last timestamp

22:47:56 <JibberJim> 2spaces !

22:47:58 <mortenf> yeah, i kinda figured, it'd need an algorithm for finding a stamp

22:48:06 <dajobe> oh yes, like picking an editor, a personal choice

22:48:12 <DanC> first thing I'd like to discuss is "CGI version by..."; that comment is sorta goofy; the CGI version is elsewhere.

22:48:56 * DanC wishes for SubEthaEdit session with libby... oh well...

22:49:19 <danbri> wonder if that'd work thru NAT/firewall etc

22:49:34 <libby> yeah I do have it danc, but perhaps wouldn;t work

22:49:39 <DanC> I don't have a mac anyway, danbri.

22:49:43 <libby> heh

22:49:45 <danbri> ah

22:49:58 <DanC> but can we do a ~20 minute pass together, libby?

22:49:58 <mortenf> DanC, what were your 2nd thoughts on prodid's?

22:50:12 <libby> i'm about for a while but I have something I need to do at same time...

22:50:14 <mortenf> sorry, any time

22:50:36 <libby> I agree re cgi version danc. are you going to do the edits or should I?

22:50:59 <eikeon> untabify :)

22:51:03 <libby> (I havn;t put the cgi version into cvs yet but I intend to.)

22:51:05 <DanC> I guess I'll edit...

22:51:08 <libby> k

22:51:48 <DanC> since you and I have write access, but MK doesn't, I propose to put us down as authors and put MK in an acks section, ok?

22:52:06 <libby> ok

22:52:19 <libby> I think would be good to link to his online version

22:55:18 <DanC> yes, there is a link to his online version; I'm moving that from the copyright blurb to the REFERENCES section

22:55:26 <libby> ok

22:55:43 <DanC> and moving his list of changes to the changelog, and moving the code appended after the changelog to before the changelog.

22:56:19 <libby> right, makes sense

22:57:33 <_andrew_> i'm trying to remember the name of something that was a subset of owl lite...something like owl tiny...

22:57:47 <dajobe> that's it

22:57:58 <_andrew_> cool...do you have a reference?

22:58:04 <dajobe> google has several

22:58:13 <_andrew_> but i tried google

22:58:14 <dajobe> let be more helpful though... hold on

22:58:46 <dajobe>http://www.w3.org/2001/sw/Europe/reports/dev_workshop_report_4/#owl-tiny

22:58:47 <dc_rdfig> G: http://www.w3.org/2001/sw/Europe/reports/dev_workshop_report_4/#owl-tiny from dajobe

22:59:01 <dajobe> G:|"OWL Tiny" discussion at semweb storage & retrieval workshop

22:59:02 <dc_rdfig> Titled item G.

22:59:03 <_andrew_> it was that workshop report...now i remember...

22:59:11 <_andrew_> thank you very much

22:59:29 <DanC> hmm... libby, sub include looks like dead code; I'm taking it out, unless you tell me why it's there.

22:59:37 <dajobe> G:errors in the transcription here are mine, let me know

22:59:37 <dc_rdfig> Added comment G1.

23:00:19 <_andrew_> brb

23:01:42 <DanC> hmm... sure would be nice to include unit tests for creatxns, ala python doctest module.

23:01:42 <dajobe> G:I think the phrase "owl feather" has also been used

23:01:43 <dc_rdfig> Added comment G2.

23:02:21 * libby looking

23:02:53 <libby> yep I think is dead danc

23:03:09 <libby> -the icalwarning is used (but could be removed)

23:03:10 <dajobe> G:no. looks like it was primarily me referring to [http://lists.w3.org/Archives/Public/www-webont-wg/2003Jul/att-0182/01-part|webont impl report draft]

23:03:10 <dc_rdfig> Added comment G3.

23:04:27 <DanC> I'm leaving icalwarning in, but adding a comment wondering if there's a standard perl way to redirect the output of warnings.

23:04:37 <libby> ok

23:05:11 * DanC does s/libby// in comment before printstr

23:06:38 <libby> ok

23:06:45 * DanC strikes lots of "ADDED" thingies, by george, I think...

23:08:20 * DanC adds George's name by MK in acks...

23:09:13 * DanC takes lots of "by MK"s out

23:13:09 <DanC> hmm... MK's "not quite meaningful" comment... I don't grok. any clues, libby?

23:13:48 <libby> hm I saw that but didn;t investigate

23:14:17 <libby> nope no idea

23:14:22 <DanC> hmm... he's getting the encoding from a different place in the ical syntax...

23:14:38 <libby> oh, interesting

23:14:51 <libby> want me to mail and ask him about it?

23:16:20 * DanC reviews http://www.ietf.org/rfc/rfc2445.txt

23:16:29 <DanC> no, I think his code is right, and my $enc code was bogus.

23:16:48 <DanC> remind me how to run the tests?

23:17:35 <libby> I'm afriad I was jjust running them one by one

23:17:45 * danbri tries 'make test', gets [...] NameError: global name 'MATHS_NS_URI' is not defined

23:17:53 * danbri updates cwm filetree

23:18:05 <libby> I added a couple more tests

23:18:09 <DanC> danbri, the way to run the cwm tests is to cd test; make

23:18:17 <danbri> thanks

23:18:44 <DanC> last time I investigated that, danbri, I did The Right Thing and documented it on the cwm home page.

23:19:00 <danbri> this was 'make test' in 2002/12/cal

23:19:06 <DanC> one by one... ew. I forgot how primitive this ical2rdf.pl stuff was

23:19:08 <libby> - MovieInfo.ics, MozMulipleVcalendars.ics, ISWC-related.ics

23:20:34 <danbri> cwm failed on test 14 [[

23:20:35 <danbri> File "/usr/lib/python2.3/urllib.py", line 420, in open_local_file

23:20:35 <danbri> raise IOError(e.errno, e.strerror, e.filename)

23:20:35 <danbri> IOError: [Errno 2] No such file or directory: '/mnt/big130/home/danbri/cvs/w3.org/WWW/2000/10/swap/test/list/last.n3'

23:20:35 <danbri> diff: ref/list-last.n3: No such file or directory

23:20:36 <danbri> Files differ, result= 512

23:20:36 <DanC> now I can see why you were treading lightly on the code, libby; here I am assuming I'll be able to easily test my changes, and hence not being very careful.

23:20:38 <danbri> ]]

23:21:14 <DanC> did you update -d, danbri?

23:21:17 <danbri> yes

23:21:45 <DanC> hard to say what's broken quickly, danbri.

23:21:46 * danbri moves swap to swap.old and starts afresh, to be sure.

23:23:34 <libby> I've never tested it like that. for a long time I couldn't use cwm on my laptop which is where the other files are

23:23:48 <libby> haven't tried recently mind you

23:24:33 <DanC> fyi, timbl/sandro/danc made some plans yesterday for a cwm release.

23:25:54 * DanC gets ical2rdf.pl to pass perl -wc

23:26:05 <danbri> 'make test' afresh barfed trivially: danbri@fireball:~/cvs/w3.org/WWW/2000/10/swap/test$ make

23:26:05 <danbri> rm ../*.pyc

23:26:05 <danbri> rm: cannot remove `../*.pyc': No such file or directory

23:26:10 * DanC reviews diffs

23:26:19 <danbri> (needs a -f or something?)

23:26:32 * danbri made a bogus ../foo.pyc instead

23:26:35 <DanC> rm *.pyc? wierd.

23:26:45 <danbri> OK confirmed hit the same error as before, with a fresh checkout

23:26:48 <DanC> the .pyc problem looks like a bug.

23:27:17 * DanC reviews diffs to ical2rdf.pl ...

23:28:41 <danbri> 2002/12/cal/ make test -> barf, "Some storage module must register with thing.py before you can use it"

23:28:53 * danbri gives up for tonight

23:31:08 <DanC> libby, I just checked in v1.22; could you smoke-test it for me?

23:32:36 <DanC> ah... indeed, it seems to handle Vtodo

23:32:49 <DanC> (which is the hypothesis that started me on this quest)

23:34:13 <DanC> [[[

23:34:14 <DanC> <component>

23:34:14 <DanC> <Vtodo>

23:34:14 <DanC> <dtstart rdf:parseType='Resource'>

23:34:14 <DanC> <dateTime>2003-02-14T00:00:00</dateTime>

23:34:14 <DanC> <tzid>US/Eastern</tzid>

23:34:16 <DanC> </dtstart>

23:34:18 <DanC> <summary>RDF icons all over swap stuuff</summary>

23:34:20 <DanC> <uid>2E1D6D0A-3493-11D7-A8B9-000393914268-RID</uid>

23:34:22 <DanC> <dtstamp rdf:parseType='Resource'>

23:34:24 <DanC> <dateTime>2004-01-21T16:04:24Z</dateTime>

23:34:26 <DanC> </dtstamp>

23:34:28 <DanC> </Vtodo>

23:34:30 <DanC> </component>

23:34:32 <DanC> ]]]

23:36:11 <DanC> ok, well, I dunno if I broke anything in v1.22, but since we don't really have a test harness in place, nobody should have expected too much of the code anyway.

23:37:20 <libby> yes we noticed that when running that mozlla export through it

23:37:29 <libby> sure

23:37:44 <DanC> noticed which? ah... that it groks Vtodo

23:40:09 <libby> yep

23:42:31 <libby> I'm gettign a lot of 'clobbering' messages for http://www.w3.org/2002/12/cal/test/MozMulipleVcalendars.ics

23:42:55 <libby> which is odd since they are all the same thing being clobbered I think

23:42:57 <libby> minor though

23:44:33 <libby> that's the only difference I'm getting. looks fine

23:46:07 <libby> thinking about it I did have a sort of test harness for the java version. That code will need updating now.

23:51:40 <DanC> thanks for reviewing 1.22, libby; we don't need to announce it, do we? i.e. it doesn't represent any changes to the schema, does it?

23:52:25 <libby> no because the files that generate the schema have stayed the smae

23:52:26 <libby> same

23:52:47 <libby> I think we don't need to announce

23:53:20 <libby> oh, there is one definite bugfix - multiple vcalendars in one file are now supported. but that shouldn't affect the schema though

23:53:30 <DanC> and that was announced.

23:53:39 <DanC> well, mentioned, anyway.

23:53:59 <libby> ok yeah

23:54:24 <DanC> tim, you were thinking about turining uid's into fragids?

23:54:43 * DanC tries to remember why we didn't do that before... something about round-tripping, maybe?

23:54:55 <timbl> This is like the problem Marja was dealing with with bookmarks.

23:55:26 <timbl> It is only useful to make a fragid for one when you intend to publish it.

23:56:14 <timbl> In our case, we can just import stuff into a bif planning file, and generate derefable URIs there.

23:56:37 <GregElin> regions are not directly displayable"

23:56:40 <GregElin> oops...

23:57:10 <timbl> Are the UIDs in iCalendar in any way related to any URI scheme?

23:57:35 <libby> there's a recommended way of generating them I think

23:58:11 <DanC> for a while I thought UIDs in iCalendar required to use the syntax of message ids, but I guess not.

23:58:12 <timbl> You would think that would have been chosen to be like Message-Id or uuid but they don't look like either.

23:59:28 <DanC> I think cid:$uid is perhaps not too much of a stretch for embedding them in URI space. But making them fragids seems perhaps more useful.

23:59:41 <timbl> So if we have a concept of an "acheivement" or something in the plan, is that the same as a TODO or is there a relationship between the TODO task and the how-why chart bubble iit is trying to achieve?