Semantic Web Interest Group IRC Chat Logs for 2004-07-14

00:05:13 <CloCkWeRX> heh

00:05:16 <CloCkWeRX> imagine that

00:05:30 <CloCkWeRX> i found another geek i know from work on this network

00:05:52 <CloCkWeRX> now to make him love foaf

00:06:00 <CloCkWeRX> *cackles*

00:08:27 * Emmy Good night all :)

00:08:37 <CloCkWeRX> gnite Emmy

01:00:48 <dude1> dude1 is now known as golbeck

02:27:14 <bitsko|away> bitsko|away is now known as bitsko

02:37:38 * irc waves from DFW

02:37:50 <irc> Flight to SAN delayed. Phtpht

02:38:00 <irc> .t pdt

02:38:02 <phenny> Tue, 13 Jul 2004 19:38:00 PDT

06:11:28 <irc> .t PDT

06:11:30 <phenny> Tue, 13 Jul 2004 23:11:28 PDT

06:12:00 <irc> Landed at SAN

06:12:17 <irc> Hope there's a hotel shuttle

09:20:51 <`flaw> `flaw is now known as flaw

09:26:25 <_jeen_> _jeen_ is now known as jeen

10:18:04 <libby>http://foafcamp.asemantics.org/#registration

10:18:06 <dc_rdfig> A: http://foafcamp.asemantics.org/#registration from libby

10:18:22 <libby> A:|FOAFCamp registration form

10:18:22 <dc_rdfig> Titled item A.

10:19:06 <libby> A:FOAFCamp is 19-20 Aug 2004, near Enschede, the Netherlands

10:19:07 <dc_rdfig> Added comment A1.

10:19:25 <libby> A:"Two days of talking, hacking, socializing and making FOAF better. Held in the parklike surroundings of Twente University, hometown of the Grolsch beer brewery."

10:19:26 <dc_rdfig> Added comment A2.

10:19:37 <libby> A:costs 100 Euros

10:19:38 <dc_rdfig> Added comment A3.

10:24:25 <libby> A:registration form uses (but doesn't require) your foaf file

10:24:25 <dc_rdfig> Added comment A4.

10:29:25 <libby> A: edging towards [http://rdfweb.org/pipermail/rdfweb-dev/2002-December/010658.html|Dan's vegetarian usecase] for FOAF

10:29:25 <dc_rdfig> Added comment A5.

10:31:01 <libby> A:jen Golbeck has an [http://www.mindswap.org/2003/vegetarian.owl|owl vegetarian ontology] btw

10:31:02 <dc_rdfig> Added comment A6.

10:32:19 <libby> not about vegtarian owls though

10:33:16 <Wack> :]

10:36:34 <OliB> so i need to set a http://xmlns.com/2001/08/wordnet/Omnivore on myself :-)~~~

10:37:02 <OliB> oops. cut'n'pasete misstake. i ment http://xmlns.com/2001/08/wordnet/Omnivore

10:37:28 <OliB> argh. what's with this app. different clipboard or what

10:38:06 <OliB> finally: http://xmlns.com/wordnet/1.6/Omnivore _|~|o (me crumbling)

10:51:16 <libby> heh

11:00:24 * dajobe notices jeen

11:00:49 <dajobe> jeen: your ISWC paper "A comparison of RDF query language proposals", is there a working PDF around?

11:08:33 <jeen> there should be, hang on.

11:09:20 <jeen>http://www.aifb.uni-karlsruhe.de/WBS/pha/rdf-query/

11:09:21 <dc_rdfig> B: http://www.aifb.uni-karlsruhe.de/WBS/pha/rdf-query/ from jeen

11:09:39 <jeen> B:|Comparison of RDF query languages

11:09:40 <dc_rdfig> Titled item B.

11:10:28 <jeen> hm

11:10:34 <jeen> the link to the pdf doesn't work

11:14:42 <jeen> dajobe, I'll check with Andreas if there is a public version around and get back to you.

11:20:06 * danja wonders how feasible 'most specific type' might be...

11:20:37 <jeen> danja, what do you mean exactly with 'feasible'?

11:21:31 <danja> useful, I suppose

11:22:11 <danja> I'm looking at a triplestore

11:22:38 <danja> recording the type(s) of resources

11:23:15 <danja> contemplating a resource table:

11:23:52 <danja> id, is_blank, uri, multiple_types (bool), types_id

11:24:27 <jeen> i see

11:24:44 <danja> if only a single type (aside from Resource) has been encountered

11:25:00 <danja> I was thinking about putting the ref to it in types_id

11:25:19 <danja> otherwise it would point to a table containing multiple values

11:26:12 <Emmy> Good morning all :)

11:26:40 <danja> morning! (13:24)

11:27:11 <danja> hmm, not sure though, I'd like to have the db handling as much of the logic as possible

11:27:16 <Emmy> :-) welll... :-))

11:27:43 <jeen> FWIW in sesame we have a seperate table for direct type relations: instance_id, type_id

11:27:44 <danja> maybe I will go for: first_type, rest_of_types

11:29:56 <Jubin> Jubin is now known as jubin|away

11:29:57 <danja> right, thanks

11:30:40 <danja> I may be trying to optimize prematurely

11:31:14 <danja> but most of the time it may not be necessary to know more than the 'primary' type

11:31:29 <danja> dunno...

11:31:33 <jeen> we have some internal report on different db schemes that we tried, but it's in dutch unfortunately. if it's of interest to you nonetheless I'll see if I can look it up and send it to you.

11:31:54 <danja> I'd be grateful, thanks

11:34:43 <danja> the main use for this will be RSS/Atom data

11:34:57 <danja> but I want it capable of handling any RDF

11:35:31 <danja> so I was thing primary_type = "rss:item" sorta thing

11:35:38 <danja> thinking

11:36:46 <jeen> problem with that in general is that an instance may not have one single most specific type

11:37:12 <danja> yep, that's what I was trying to say with "feasible"

11:37:24 <jeen> ok, so I was stating the obvious then :)

11:37:45 <danja> no, confirming it, a darn sight more helpful ;-)

11:38:17 <jeen> this is why we use the seperate table. more than one direct type is simply more rows in the table

11:39:32 <danja> right, but it's also another lookup

11:39:56 <jeen> true. the assumption is that it is information you only need sporadically.

11:41:17 <danja> I want to be able to ask the equivalent of SELECT * FROM items

11:41:28 <danja> WHERE date == today

11:42:03 <jeen> on which DB are you implementing?

11:42:31 <danja> sqlite at the moment, but trying to make it reasonably generic

11:42:42 <jeen> you could consider using an ORDBMS, and using subtables for the class hierarchy. Then class-instance relations are simply encoded by table entries.

11:43:28 <jeen> only works well in a scenario where there are relatively few updates to the data, though.

11:43:33 <danja> right, I think that's how RDFSuite does its schema-based bits

11:43:38 <jeen> yes

11:44:48 <danja> I need this to be low-common-denominator, to show how a triplestore can be used for RSS

11:44:59 <jeen> in a gray past sesame used to do that as well. we stepped away from it because updates were horribly slow

11:45:11 <danja> that's interesting

11:45:11 <jeen> ah i see

11:45:56 <danja> got a *long* write-up to do ;-)

11:46:09 <chaalsCPH> anyone got time to help me with cwm for 15 minutes?

11:46:46 * danja waves to chaals then dashes away

11:47:16 <danja> thanks jeen

11:47:29 <jeen> np. i'll look up that report

11:47:36 <danja> ta

11:48:08 * chaalsCPH sighs :-)

12:13:53 * CloCkWeRX loox for someone who can read japanese at all

12:18:19 <sinjax> hi, has anyone here played with sesame under high load?

12:19:27 <sinjax> im currently running it with a MySQL repositroy with 3million+ triples and getting some very odd results

12:19:47 <sinjax> some querys such as SELECT A FROM {A} B {C} WHERE C = "0.0" are ultra fast...80ms

12:20:05 <sinjax> but when i try to do something like SELECT A FROM {A} B {C}....it gives me an outofmemory exception

12:20:09 <sinjax> anyone? :)

12:20:21 <crschmidt> Maybe you're just getting a huge set of results?

12:20:27 <rreck> is this under MS or unix /linux ?

12:20:32 <crschmidt> I've never played with seaseme.

12:22:12 <jeen> sinjax, do you use Sesame through tomcat?

12:22:19 <sinjax> jeen, yes

12:22:34 <sinjax> i was unaware that there was another way to use it :D

12:23:01 <jeen> you could try increasing the JVM's heap size. By default, tomcat initializes the jvm with a 64MB heap, which can be small. try increasing it to, say, 256M

12:23:16 <jeen> Sesame can also be used standalone or as a java library ;)

12:23:51 <sinjax> Sesame can be run stand alone...i thought it was a servlet that needed a container?

12:24:02 <jeen> er

12:24:03 <jeen> right

12:24:25 <sinjax> what theres a way to run it without tomcat?

12:24:31 <jeen> brainfart :) I meant that you can use it in a standalone application, and not in a server-client setting.

12:24:53 <jeen> it also has an RMI interface and that does not require tomcat.

12:26:16 <sinjax> jeen, ah ok...hmm...but do you think running it stand alone would solve this problem?...i do you know what exactly its doing when i do such a query

12:26:24 <sinjax> is it trying to keep the results in memory for some reason?

12:26:30 <sinjax> cus if it is thats really stupid ;)

12:26:51 <rreck> maybe you could look at the memory usage and see if its running out ?

12:27:33 <sinjax> good call

12:27:39 <jeen> it shouldn't, results are fed back streaming, normally. but on 'big' queries things can sometimes clog up a little. Increasing the heap size should fix the problem though.

12:28:00 <jeen> 3+ million triples isn't that much, it should handle that easily.

12:28:30 <sinjax> jeen, yes this is what im thinking ;)

12:28:45 <sinjax> i mean ive put exactly the same data (in non triple form) into SQL and its fine

12:29:12 <sinjax> heap ok...how can i increase tomcats heap?

12:30:18 <jeen> in TOMCAT_HOME/bin/catalina.bat (or .sh under unix), there's a line that defines JAVA_OPTS

12:30:19 <rreck> JAVA_OPTS=-Xms, i googled and got answers

12:30:37 <jeen> add -Xmx256M to set it to 256M.

12:31:39 <jeen> if you still run into problems, I'd like to hear about it...

12:32:00 <sinjax> jeen, ok :)

12:36:20 <sinjax> jeen oooo now its gobling up the memory....250mb so far ;)

12:36:43 <sinjax> oh no now its dropped back to 30 :D...random

12:38:11 <jeen> java garbage collection, it's almost magic :)

12:40:00 <sinjax> ah nope dead again

12:40:14 <sinjax> java.lang.OutOfMemoryError ....took longer this time

12:40:22 <jeen> this is strange

12:41:31 <jeen> can you send me an e-mail about this, with the details, like what the query is, what your setup is, etc?

12:41:31 <sinjax> i mean i dont think ive done anything wrong...i've added the RDF/XML file (a rather simple although large one i generated) via the web interface to the MySQL database

12:41:40 <sinjax> jeen sure :)...email?

12:41:45 <jeen> jeen@aduna.biz

12:42:24 <jeen> do you do the query through the web interface as well?

12:43:46 <JimH> this is a nice touch, on the welcome note of #rdfig this url - http://logicerror.com/rdfIRCWelcome

12:47:29 <sinjax> jeen, yes im querying through the web interface

12:47:41 <sinjax> email sent dude

12:47:44 <jeen> ok thanx

12:49:54 <jeen> got it. one thing: is there anything else in the stacktrace after the outofmemoryerror?

12:50:44 <sinjax> "Apache Tomcat/4.1.30" thats it

12:50:51 <sinjax> thats on the web interface

12:51:30 <sinjax> can you join #flood ?....the error i get trying to do the query form java is there

12:51:44 <jeen> hm. in the sesame log file perhaps? ([SESAME_DIR]/WEB-INF/logs/repositories/[repository].log)?

12:53:13 * dajobe returns from learning javascript

12:53:20 <sinjax> jeen, in flood again

12:53:23 <dajobe> yeah jeen that was the 404 I got

12:53:24 <jeen> got it

12:54:10 <jeen> dajobe, I just got a mail from Peter Haase. Their server is acting up but he expects it fixed soon. unfortunately I don't have a copy here otherwise I'd send it t you.

12:54:29 <sinjax> jeen, hmm so yeh thats what the log file shows...hehe i didnt know this was an error i had to fix...i thought sesame was just a bit crappy ;)

12:56:16 <jeen> why, thank you... :P

12:56:41 <sinjax> no no im sure its an incredible peice of software...i think i just gave up on it too quickly

12:57:03 <jeen> heh, no worries

12:57:12 <sinjax> ive been reviewing forms of RDF data storage you see...alot of the others couldnt even handle this amount of data ;)

12:57:26 <sinjax> (Jena, Joseki, 4Suite ...etc...etc)

12:57:28 <dajobe> what size data?

12:57:42 <sinjax> 3.5 million triples...they just fell apart as i was tyring to load the file

12:57:43 <jeen> ok. can you send me a (zipped) version of your data file? I'll try to reproduce the error.

12:57:45 <sinjax> (400mb file )

12:58:00 <sinjax> ....400mb?...jeen? :D

12:58:07 <jeen> er

12:58:08 <jeen> right

12:58:16 <jeen> can you put it online somewhere, so I can download?

12:58:25 <jeen> might be a better idea :)

12:58:27 <sinjax> err not from here...damn work computer

12:58:40 <sinjax> i can give you the java that generated it if you like

12:58:49 <jeen> that would be fine

13:00:11 <sinjax> emailed jeen

13:00:23 <jeen> great thanx. I'll get back to you on this ASAP alright?

13:00:31 <sinjax> thanks dude :)

13:08:37 <jsled> dajobe: thanks for you help last night re: redland / turtle...

13:09:00 <jsled> dajobe: I noticed / are you aware that it doesn't handle the integer production correctly?

13:09:01 <dajobe> oh good. I'm awake now, you can ask me harder questions :)

13:09:03 <jsled> :)

13:09:13 <dajobe> no

13:09:26 <jsled> yeah, it doesn't, and it makes me sad.

13:09:38 <dajobe> give me a valid example that fails

13:11:00 <jsled> FWIW, it also needs to be [foo = RDF.Parser("turtle", "application/x-turtle")]

13:11:08 <jsled> Leaving out the mime_type argument does not work.

13:11:34 <jsled> http://www.asynchronous.org/todo.owl+n3

13:12:01 <jsled> Note that the first Restriction will not be present in the parse triples; you may or may not get a warning during parsing.

13:12:15 <jsled> This was with 0.9.15, obviously using the python API-binding.

13:12:50 <dajobe> 0.9.16 is latest

13:13:07 <jsled> yeah ... going to sync portage today.

13:13:26 <dajobe> ah, you are the gentoo devel? they are way behind in my packages mostly

13:14:18 <sinjax> Jeen is there any document which outlines how Sesame should perform and what its limits are in terms of load?

13:14:30 <dajobe> 0.9.15 is surely so old that raptor didn't have turtle integer support

13:14:32 <jsled> no ... I'm no gentoo devil. ;) It does look like they have an ebuild for 0.9.16 ... though both 0.9.1[56] are package-masked.

13:14:35 <jsled> ah.

13:14:43 * jsled headslaps.

13:14:50 <dajobe> I can't see any problem with my copies here

13:14:58 <jsled> Of course... the timeline just sync'ed in my head....

13:15:08 <dajobe> 0.9.15 had raptor 1.2.0

13:15:27 <dajobe> raptor 1.3.0 had turtle integer literals

13:15:33 <dajobe> well, I'm glad it's easily solved ;)

13:15:39 <jsled> sorry to waste your time.

13:23:06 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

13:25:21 <sinjax> ??

13:26:08 <jeen> yay for netsplits

13:28:10 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

13:29:24 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

13:33:13 <sinjax> whats happening to this channel? :D

13:33:23 <crschmidt> It's the network, not just this dchannel

13:33:29 <sinjax> ah

13:33:30 <crschmidt> Netsplits

13:33:33 <danbri_dna> .g netsplit irc

13:33:35 <phenny> netsplit irc: http://irc.netsplit.de/channels/

13:33:41 <crschmidt> heh

13:33:41 <danbri_dna> hm

13:33:50 <crschmidt> not quite what you were looking for ;)

13:33:56 <crschmidt> foiled by a hostname!

15:10:08 <bparsia> bparsia is now known as bijan

16:00:21 <jsled> dajobe: hmm. FWIW, the gentoo redland-0.9.16 package seems like it's fux0r3d.

16:00:43 <jsled> it claims that raptor-1.3.0 is installed, but `rapper --input turtle <file>` fails in the same way.

16:01:02 <jsled> Plus..../usr/lib/libraptor.la still claims 1.2.0

16:01:29 <jsled> [plus, they have the stand-alone raptor library in media-libs, and the redland package doesn't depend on it [!]]

16:01:45 <jsled> In any case, I'll follow up with the gentoo package maintainer, but wanted to let you know...

16:02:11 <chaalsCPH> ?

16:02:18 <chaalsCPH> a:

16:02:20 <chaalsCPH> A:

16:02:21 <dc_rdfig>http://foafcamp.asemantics.org/#registration

16:02:22 <dc_rdfig> FOAFCamp registration form

16:02:23 <dc_rdfig> (1:libby) FOAFCamp is 19-20 Aug 2004, near Enschede, the Netherlands

16:02:24 <dc_rdfig> (2:libby) "Two days of talking, hacking, socializing and making FOAF better. Held in the parklike surroundings of Twente University, hometown of the Grolsch beer brewery."

16:02:25 <dc_rdfig> (3:libby) costs 100 Euros

16:02:26 <dc_rdfig> (4:libby) registration form uses (but doesn't require) your foaf file

16:02:27 <dc_rdfig> (5:libby) edging towards [http://rdfweb.org/pipermail/rdfweb-dev/2002-December/010658.html|Dan's vegetarian usecase] for FOAF

16:02:28 <dc_rdfig> (6:libby) jen Golbeck has an [http://www.mindswap.org/2003/vegetarian.owl|owl vegetarian ontology] btw

16:02:52 <dajobe> jsled: sure, or we can take this to #redland for working out the details

16:05:48 <jsled> ah ... didn't realize there was a #redland [though I should have pre-emptively checked]

16:07:09 <chaalsCPH> danbri_dna?

17:14:11 <Ismael_> Ismael_ is now known as Ismael

17:14:27 <Ismael> (quit

17:20:04 <crschmidt> I'm looking for a structure for something similar to Trackbacks, only for RDF files: When I write about something described in RDF file http://example.com/foo.rdf , I want to be able to open it, look for a URL to send a "ping" to, and send it in a format that's understood. Does such a thing exist?

17:21:37 <crschmidt> Also, is there a better generalized address listing than what ChefMoz (http://chefmoz.org/rdf/elements/1.0/) uses? Does vCard provide access to Address type information?

17:21:49 <libby> yeah re vcard

17:21:54 <libby> .g vcard rdf

17:21:57 <phenny> vcard rdf: http://www.w3.org/TR/2001/NOTE-vcard-rdf-20010222/

17:22:01 <guido> guido is now known as guido-away

17:22:20 <libby> or http://www.w3.org/TR/vcard-rdf

17:22:34 <libby> that's not a great RDF transliteration, imho

17:22:47 <wkearney> vcard, welcome to hell

17:22:47 <libby> notclear where to get address properties from else though

17:23:02 <libby> we'd like some in foaf, but it's soo complex :(

17:23:06 <wkearney> and people think rdf is fun to parse...

17:23:15 <libby> yeah

17:23:20 <crschmidt> Chefmoz providdes real easy access to them, but it's a schema that's got everything in it, including a complete redefinition of Geo:

17:23:35 <libby> yeah.

17:24:04 <libby> I did a sample conversion of their opening hours format to rdf somewhere

17:24:11 <libby> i.e. rdf ical

17:24:29 <crschmidt> ooh, that's a real nice vcard page, i didn't find that the last time i was looking

17:25:07 <libby> yeah. it was early in rdf's history: they might have done some things differently now

17:25:22 * libby heads home

17:32:01 * mortenf leaves for beer with chaals

18:09:53 <crschmidt> a geo:Point can contain more information about that point, right?

18:10:09 <crschmidt> For example, i can have lat, long, dc:title, dc:description?

18:13:29 <danbri_dna> "contain"?

18:13:29 <danbri_dna> points are spatial things, they can't have infomration _inside_ them

18:13:39 <danbri_dna> geo:Point XML elements in RDF/XML can have stuff inside them, of course

18:13:51 <danbri_dna> ...anything that makes sense of a Point

18:14:25 <crschmidt> Yeah, that's what I meant, sorry

18:14:25 <danbri_dna> sorry I sound like a pedantic arse! I just mean, that RDF vocabs don't get to choose whether others can add more info about the things they're designed to describe

18:14:33 <danbri_dna> RDF itself guarantees that Freedom

18:14:50 <crschmidt> Well, can't range be defined such that only specific types of elemennts could be included?

18:14:52 <danbri_dna> it's a freedom we should be trumpetting in a GNU-ish evangelical sort of way, I sometimes think

18:15:31 <danbri_dna> No. Range/Domain are claims about the world. They btw apply to properties rather than classes, ie. things like geo:lat rather than geo:Point

18:15:39 <crschmidt> Ah.

18:15:48 <crschmidt> Well then. Learn something new about RDF every day.

18:15:50 <danbri_dna> canonical case is the one in the MCF tutorial

18:15:56 * danbri_dna googles

18:16:18 <danbri_dna> I wrote a few things on this sort of stuff in foaf weblog last summer, might not be so easy to find tho

18:17:03 <danbri_dna> check out the 1st diagram in http://www.w3.org/TR/NOTE-MCF-XML/

18:17:21 <danbri_dna> (MCF is an RDF and RSS precursor tech, from Guha when at Apple and Nescrape)

18:17:35 <danbri_dna> ie. http://www.w3.org/TR/NOTE-MCF-XML/simple.gif

18:18:10 <danbri_dna> range of 'author' is 'Person' says, if you see x--author-->y, and its a true description of the world, you've also just learned that ?y is a Person

18:18:22 <danbri_dna> it doesn't say anything about what actual instance data must contain

18:18:33 * crschmidt nods

18:18:52 <danbri_dna> which is why RDF feels somewhat slippery sometimes, and RDF vocabs don't seem to say enough to satisfy coders working in the XML tradition

18:19:27 <danbri_dna> they don't demand that a "Document" element MUST contain an "author" subelement, nor a "Person" below that. It just tells you a little of what it means if you do choose to write these things.

18:19:58 * crschmidt nods. Cool. :)

18:20:03 <crschmidt> Like I said, learn something new about RDF every day

18:20:09 <danbri_dna> The reason is that RDF is designed to play in the spaces between applications, and apps can rarely agree on which information items are critical vs optional, especially given such widely relevant topics as 'documents', 'persons', etc.

18:20:30 * crschmidt nods

18:20:32 <danbri_dna> within more constrained environments, eg. load/save amongst clones-of-each-other Social Network sites, there may be scope for tighter consensus

18:20:38 <danbri_dna> hence things like foafnet have a role

18:20:58 <danbri_dna> so long as they don't undermine the core freedoms that allow that data to play with a wider world of person/doc etc descriptions...

18:21:05 <danbri_dna> (btw FOAF also owes a lot to MCF)

18:21:22 <danbri_dna>http://rdfweb.org/mt/foaflog/archives/000047.html

18:21:23 <dc_rdfig> C: http://rdfweb.org/mt/foaflog/archives/000047.html from danbri_dna

18:21:38 <danbri_dna> C:|Missing isn't broken: data validation and freedom on the Semantic Web, by danbri July 2003

18:21:38 <dc_rdfig> Titled item C.

18:21:48 <danbri_dna> logger, chump bookmark

18:21:48 <danbri_dna> BOOKMARK:See [http://www.ilrt.bris.ac.uk/discovery/chatlogs/rdfig/2004-07-14#T18-21-48|discussion]

18:21:49 <dc_rdfig> Label BOOKMARK not found.

18:21:55 <danbri_dna> logger, chump C

18:21:55 <danbri_dna> C:See [http://www.ilrt.bris.ac.uk/discovery/chatlogs/rdfig/2004-07-14#T18-21-55|discussion]

18:21:56 <dc_rdfig> Added comment C1.

18:22:02 <danbri_dna> good bot!

18:22:34 <danbri_dna> hope all that made sense...

18:23:01 <crschmidt> Yep

18:24:29 <crschmidt> "With freedom comes responsibility."

18:24:30 <crschmidt> :)

18:31:06 <danbri_dna> was that from Spiderman, I forget ;)

18:32:26 <crschmidt> I think so, yeah

18:32:33 * crschmidt is excited, he now has business cards

18:35:57 <crschmidt> Would it be inappropriate to include vCard information - country, locale, address -- in a geo:Point structure?

18:38:27 <crschmidt> For example, I live at <Point><lat>foo</lat><long>bar</long><vcard:country>United States</vcard:country><vcard:Street>1032 Valley</vcard:Street><vcard:Locality>Manchester</vcard:Locality></Point>

18:42:09 <OliB> what's the domain of these vcard properties ?

18:44:23 <crschmidt> Doesn't look like there is one

18:44:32 <crschmidt> The schema doesn't contain the word "domain" anywhere

18:45:03 <OliB> how about the original vCard RFC ? what's supposed to be described by a vcard ?

18:46:37 <OliB> even if there's no domain specified in the schema/ontology, if vcards are for people (assumption), it's probably not a good idea to use vcard properties for geo:Points

18:48:25 <crschmidt> So if I want to say, <thingy id="myhouse" />, I should have a geo:Point and a seperate vcard:ADR, you think?

18:48:40 * crschmidt is honestly just trying to come up with the best way to do these things, doesn't really have any idea what is good/bad

18:49:02 <crschmidt> Yeah. I think that looks better.

18:49:16 <crschmidt> <vCard:ADR rdf:parseType="Resource">

18:49:17 <crschmidt> <vCard:Street> 111 Lake Drive </vCard:Street>

18:49:22 <crschmidt> </vcard:ADR>

18:50:29 <OliB> addresses and geo:Points are certainly connected (for a given moment in time), but that seems cleaner, IMHO

18:50:46 * crschmidt nods.

18:51:08 <OliB> tough i personally despise rdf:parseType="Resource"

18:51:34 <jsled> what does parseType=Resource indicate, again?

18:51:34 <crschmidt> I'm looking to develop a combination of schemas to allow people to accurately describe shops, stores, businesses in a way that doesn't completely ignore the existing schemas like ChefMoz does

18:51:46 <crschmidt> I don't really know, I just copy pasted from the vcard example :)

18:52:06 <jsled> :)

18:56:18 <danbri_dna> vcard... does it even have a schema?

18:56:34 <danbri_dna> rdf vcard is ancient history, never really went anywhere.

18:56:36 <danbri_dna> what's the url?

18:57:31 <OliB> probably http://www.w3.org/TR/2001/NOTE-vcard-rdf-20010222/

18:57:44 <mattmcc> So in the absence of rdf vcard, is there a good alternative for mailing addresses and the like?

19:01:16 <crschmidt> RDF vcard has a schema

19:01:22 <crschmidt> Linked from http://www.w3.org/TR/vcard-rdf

19:01:37 <crschmidt> as http://www.w3.org/2001/vcard-rdf/3.0#

19:02:42 <crschmidt> Hm. ADR is labelled as a property... can it be used like it was in my example above?

19:02:49 <crschmidt> mattmcc: not that i've found.

19:06:41 <danbri_dna> anywone wanna write up a proposal for a foaf extension? (I think there were suggestions from Graham Klyne some time back...)

19:07:03 <danbri_dna> the geopriv WG at ietf have a (not quite RDF?) namespace for addresses too, I think

19:07:57 <crschmidt> FOAF extension to describe addresses? To replace the current vcard stuff?

19:08:56 <danbri_dna> not 100% subsume all of vcard, but make us be a bit more self-standing in that regard...

19:09:21 <danbri_dna> I don't forsee http://www.w3.org/TR/2001/NOTE-vcard-rdf-20010222/ getting updated in a hurry

19:09:28 * crschmidt nods.

19:09:35 <danbri_dna> not that foaf is getting hurried updates either, but it is alive and moving

19:09:43 * crschmidt isn't willing to commit, but may poke at it at some point

19:09:50 <danbri_dna> ta

19:10:51 <OliB> vcard RDF feels too much like a hack. something new and modern would be great

19:11:40 <crschmidt> OliB: one of the benefits of it is that the information can be converted to and from many popular address book programs without data loss

19:12:09 <danbri_dna> yup. i think (esp after chat w/ mortenf in spain) that foaf's horky firstname/lastname fields may have value in that regard too

19:12:21 <OliB> the new thing could describe a conversion rule

19:12:34 <danbri_dna> too many (bad) desktop and palmtop apps use those concepts instead of family/given, which themselves don't have 100% applicability but are much better

19:12:50 <danbri_dna> but I think they can be left in the spec 'for representing legacy data'

19:13:55 <MacIntire> I'm going to login to a bunch of dating sites over the next few weeks and see if I can come up with some sort of FOAF traits or "dating" extension.

19:14:39 <MacIntire> i.e. politics/turn offs/religion - kind of an extension onto the Myers-Briggs property

19:15:13 <OliB> if I remember correctly, orcut had a bunch of such properties

19:15:35 <MacIntire> FOAF properties?

19:15:39 <MacIntire> or just Orkut?

19:16:21 <OliB> i think i remember answering questions about polical/sexual preference, when i created an orkut accout. had nothing to do with FOAF

19:16:37 <MacIntire> gotcha - yeah, I'll hit YASNs too.

19:16:53 <MacIntire> It will be an interesting survey

19:16:56 <danbri_dna> good plan!

19:17:16 <danbri_dna> i have logins on lots, similar motivation, but never wrote it all down. also didn't want to be accused of stealing ideas.

19:17:27 <danbri_dna> but if you can summarise common trends at least that'd be cool

19:20:15 <MacIntire> right...and common values

19:32:28 <CloCkWeRX> i think

19:32:42 <CloCkWeRX> i might just finally get how to work this public key encryption gambit

19:54:23 <bengee> danja, is your project vocab still active?

19:55:41 <bengee> s/active/maintained

19:58:51 <bengee> I'd like to use it, but the props have multiple domains assigned instead of an owl:unionOf..

20:05:04 <bengee> edd's doap, too. but only on a property I won't need..

20:06:23 * bengee continues searching...

20:10:23 <danbri_dna>http://www.betaversion.org/~stefano/linotype/news/57/

20:10:23 <dc_rdfig> D: http://www.betaversion.org/~stefano/linotype/news/57/ from danbri_dna

20:10:40 <danbri_dna> D:|Stefano's RDF/SW intro (for "XML people")

20:10:41 <dc_rdfig> Titled item D.

20:57:23 * mortenf burps

20:59:30 <crschmidt> heh

20:59:33 <crschmidt> good beer? :)

20:59:58 <mortenf> well, at least it was technically beer, this being .dk... :/

21:00:19 <mortenf> didn't have time to go to the right places

21:00:46 <mortenf> allthough we actually did, since a train was cancelled

21:01:03 <mortenf> still no luck getting someone to sponsor your trip to galway?

21:01:12 <crschmidt> Eh, probably not

21:01:20 <crschmidt> right now I'm more concerned about getting submissions in then going

21:01:34 <mortenf> heh, looking forward to that

21:01:34 <crschmidt> Will worry about possibly going after this weekend, visiting parents, which is also on my mind

21:01:46 <mortenf> hmm, far away?

21:01:48 <crschmidt> I've got one paper done, 2 started, and 7 ideas

21:01:55 <crschmidt> yeah, 1000 miles or so, going to visist for mom's birthday

21:27:47 <MacIntire> MacIntire is now known as bkdelong

21:28:05 * bkdelong waves to mortenf

21:28:14 <bkdelong> I installed your Wordpress plugins today

21:28:15 <mortenf> hola

21:28:22 <mortenf> ah, did they work?

21:28:27 <bkdelong> *rubs hands maniacally*

21:28:34 <mortenf> :)

21:28:46 <bkdelong> Still need to add in the sha1 plugin and try out some other things

21:29:17 <mortenf> i'm sorry, but you'll have to upgrade in a few hours - i found a bug...

21:29:35 <bkdelong> hehehe

21:29:35 <mortenf> erhm, actually ndw found it for me

21:29:40 <bkdelong> Where?

21:30:03 <mortenf> it's the uris for us states, points to the wrong ontology

21:30:06 <bkdelong> Do you know if it's possible to expand the Wordpress profile page to add in new FOAF-friendly features? or to add a second page?

21:30:13 <bkdelong> ah..gotcha

21:30:41 <mortenf> see http://b4mad.net/datenbrei/index.php?p=18 for an example

21:31:13 <mortenf> also, danja asked for seealso's - not sure how to handle that yet, still thinking...

21:32:03 <bkdelong> It sounds like Wordpress may work as a means for creating a WYSIWYG FOAF editor...since its so hackable

21:32:22 <mortenf> yep, that's the direction i've been thinking as well

21:32:43 <mortenf> plus, quite a lot of the info is already there

21:33:06 <bkdelong> exactally. I'm thinking it will be a good platform for the MeNow extension as well

21:33:43 <mortenf> methinks menow is a different beast

21:34:16 <mortenf> since rdf is monotonic, it doesn't deal well (yet?) with retractions

21:34:26 <bkdelong> ah...right.

21:34:54 <mortenf> of course, it'd be kinda silly to do it completely detached from foaf

21:35:30 <mortenf> so, while i think the concept is sound, i'm not sure how to tie it to foaf

21:35:39 <bkdelong> Well, for MeNow we were toying with PersonalProfileStatus

21:35:50 <bkdelong> or PersonalProfilePresence

21:36:27 * mortenf hopes there'll be some presence stuff going on at foafcamp

21:36:52 * bkdelong shrugs.

21:37:08 <bkdelong> crschmidt is doing a lot with presence + FOAF. Is anyone else?

21:37:25 <mortenf> edd and jibberjim

21:40:46 <bkdelong> Edd and his Bluetooth stuff? What's Jibberjim doing?

21:41:04 <mortenf> he was doing some bt as well, with maps

21:41:16 <mortenf> not sure where it went...

21:41:50 <bkdelong> ah gotcha.

21:43:06 <mortenf> hmm

21:43:49 <bkdelong> Speak of the Baatezu

21:45:35 <mortenf> it seems a property named http://www.daml.ri.cmu.edu/ont/CanadianState.daml#name is used in http://www.daml.ri.cmu.edu/ont/CanadianState.daml, but it's not defined, but http://www.daml.ri.cmu.edu/ont/State.daml#name is defined and http://www.daml.ri.cmu.edu/ont/State.daml is imported from the former - something i'm missing (perhaps the semantics of owl:imports)?

21:50:01 * mortenf checks the owl specs and doesn't get any wiser :(

21:52:21 <mortenf> hmm, i find it hard to believe that an import would remap all classes and properties to the importing vocabulary's namespace

21:53:56 <dajobe> huh?

21:54:07 <dajobe> classes and properties (model) namespace (syntax)

21:54:24 <dajobe> not that I can help with what owl:imports means

21:54:33 <mortenf> yeah, but see graphs at uris mentioned above

21:55:00 <mortenf> it seems like the authors interpret import that way - or they made a mistake

21:55:14 <dajobe> it's a document import, more triples

21:55:24 <mortenf> yeah, that's what i though

21:55:28 <mortenf> +t

21:55:33 <dajobe> but it might not be, it's unclearl

21:55:57 <mortenf> yeah, i don't exactly know how to interpret the spec text(s)

21:55:57 <dajobe> I once asked if I should implement owl:imports in a parser

21:56:01 <dajobe> the answer was "maybe"

21:56:05 <mortenf> heh

21:56:52 <bkdelong> are those schema URLs at daml.org or only daml.ri.cmu.edu

21:57:04 <mortenf> don't know...

21:57:22 <bkdelong> might want to check....I think daml.org is more official/permanent. ;)

21:57:31 <mortenf> yeah...

21:58:31 <bkdelong> BTW, does USNOAAD superceed the twilight vocab?

21:58:49 <mortenf> hmm

21:59:00 <mortenf> aren't they the same?

21:59:11 <bkdelong>http://www.wasab.dk/morten/2004/04/usnoaad

21:59:12 <dc_rdfig> E: http://www.wasab.dk/morten/2004/04/usnoaad from bkdelong

21:59:15 <bkdelong>http://www.wasab.dk/morten/2004/04/twilight/

21:59:16 <dc_rdfig> F: http://www.wasab.dk/morten/2004/04/twilight/ from bkdelong

21:59:32 <bkdelong> doh....sorry about that.

21:59:56 * bkdelong needs to remember how to label chump entries.....

22:00:27 <mortenf> E:|usnoaad

22:00:27 <dc_rdfig> Titled item E.

22:00:45 <mortenf> E:|usnoaad vocabulary

22:00:45 <dc_rdfig> Titled item E.

22:00:57 <mortenf> F:|twilight service

22:00:58 <dc_rdfig> Titled item F.

22:01:15 <bkdelong> So....what do you think?

22:01:23 <mortenf> that's just one vocab or am i not looking straigt?

22:03:10 <bkdelong> oh blast.

22:03:24 <bkdelong> I should actually look at the URLs :/

22:03:34 <mortenf> :)

22:03:50 <bkdelong> Updating http://rdfweb.org/topic/FoafExtensions

22:07:25 * mortenf doesn't find the state stuff at other uris on daml.org

22:07:56 <bkdelong> weird

22:08:16 <bkdelong> I hate unorganized decentralization

22:08:27 <mortenf> however, i'm still confused about the property uris vs. the use of owl:imports, hints would be appreciated...

22:08:48 <mortenf> hmm, that'd maker a nice streamer :)

22:08:49 * bkdelong stares at em

22:09:40 <bkdelong> I'm owl-challenged

22:09:51 <CloCkWeRX> oh oh

22:09:52 <CloCkWeRX> i think

22:10:06 <CloCkWeRX> i just stuck my hand up for documenting bits of the mozilla api

22:10:12 * CloCkWeRX hides under a rock

22:15:13 <mortenf> dajobe, on public-rdf-dawg alberto states that "using '/' instead would also make those URIs cacheable while '?' would generally not (most proxies/firewalls grep URL queries parts off)" -- any chance you could ask for supporting evidence, i think it's a myth?

22:15:42 <dajobe> I know squid.conf typically junks urls with ?

22:15:52 <mortenf> hmm, ok then

22:16:09 <dajobe> but ask him yourself, lazy :)

22:16:24 <bkdelong> heh

22:16:31 <mortenf> i can't write there (heck, i'm not even supposed to be reading)

22:16:48 <mortenf> but ok, will do

22:16:50 <dajobe> reading's fine

22:17:49 <mortenf> via web, yeah...

22:18:11 <dajobe> there's a DAWG f2f going on right now

22:18:16 <dajobe> query requirements being considered

22:18:31 <mortenf> right, and you're not there?

22:18:37 <bkdelong> DA?

22:18:40 <dajobe> virtually

22:18:43 <mortenf> ah

22:20:05 <dajobe> RDF Data Access WG

22:26:04 <bkdelong> mortenf - what's the difference between the photo vocab and exif? Is it an exif extension?

22:26:21 <mortenf> yeah, it's stuff that isn't in exif

22:26:27 <mortenf> ... and that i had a need for

22:27:20 <bkdelong> gotcha - cool.

22:27:47 <bkdelong> I wish Flickr was FOAF-friendly. I'd love to be able to annotate using FOAFs

22:28:45 <CloCkWeRX> flickr?

22:28:50 * CloCkWeRX rubs tired tired eyes

22:29:16 <bkdelong> A pretty cool photo YASN

22:29:24 <CloCkWeRX> ah

22:29:30 <bkdelong> www.flickr.com

22:30:09 <eaon> isn't flickr something from the fotonotes guys?

22:30:15 <eaon> (which is related to w3photo.org)

22:30:24 <eaon> (iirc)

22:31:10 <bkdelong> I'm not sure of the six degrees there.

22:31:28 <bkdelong> But it's all interrelated - they're definitely making use of fotonotes

22:39:52 <crschmidt> Anyone here know if there is there any limit as to who can come to FOAF-Galway? Someone in another channel was asking.