Semantic Web Interest Group IRC Chat Logs for 2003-05-15

00:25:34 <jordan> jordan is now known as jjjjjj

00:25:38 <jjjjjj> jjjjjj is now known as jordan

03:00:05 <tav> tav is now known as tav|passedaway

08:45:53 <arnarl> hi

09:27:17 <dajobe>http://www.ucc.ie/opa/pr/PressReleaseSFI.html

09:27:18 <dc_rdfig> A: http://www.ucc.ie/opa/pr/PressReleaseSFI.html from dajobe

09:28:01 <dajobe> A:|12 million euros for Dieter Fensel on "developing "semantic web technology" at NUI Galway

09:28:01 <dc_rdfig> Titled item A.

09:29:09 <dajobe> A:which I guess explains why his vu.nl homepage is a 404 :)

09:29:09 <dc_rdfig> Added comment A1.

09:29:14 <jeen> heh

09:29:50 <jeen> actually, dieter moved to university of innsbruck. apparantly he is also commuting to Galway...

09:30:23 <dajobe> google doesn't seem to know that

09:30:42 <jeen> that's odd. he moved almost a year ago

09:30:56 <dajobe> .google "dieter fensel"

09:30:57 <datum> "dieter fensel": http://www.aifb.uni-karlsruhe.de/WBS/dfe/

09:31:00 <dajobe> 404

09:31:04 <dajobe> or actually, 500

09:31:19 <jeen> ah. that one is even older, that's his Karsruhe homepage :)

09:31:23 <dajobe> maybe somebody needs to discuss with him about key concepts of semweb, uri persistence etc.

09:36:38 <jeen> :)

09:41:12 <sandro_> I wonder who'll end up moving to Ireland, or if they'll develop the expertise there.....

09:41:22 <sandro_> sandro_ is now known as sandro

09:41:38 <dajobe> I'm betting the money comes with strings of *actually working in galway*

09:41:53 <sandro> yeah

09:41:53 <dajobe> it's a nice area

09:42:05 <dajobe> near shannon airport, so not so hard to get to

09:42:26 <JibberJim> you could commute from stanstead pretty easily...

09:42:30 * dajobe checks the sandro time..

09:42:49 <dajobe> it must be just dawn for you

09:49:13 <sandro> nah, I think I saw the sun come up at least an hour ago

09:50:12 <sandro> i'd like to say I've started adjusting my body clock for Budapest, but in fact my family tends to dictate when I sleep

09:53:41 <libby> morning AndyS

09:54:47 <AndyS> Morning Libby

09:54:59 <AndyS> "Morning" Sandro

10:03:28 <bijan> Morning edd

10:03:35 <bijan> long time no see here

10:03:52 <edd> hey bij. yeah, popped in last night. gearing up for www2003 :)

10:04:33 <bijan> Woohoo!

12:44:49 <hooch> hooch is now known as nothooch

12:45:00 <nothooch> nothooch is now known as hooch

12:45:35 <dajobe>http://www.hpl.hp.com/semweb/jena

12:45:35 <dc_rdfig> B: http://www.hpl.hp.com/semweb/jena from dajobe

12:45:43 <dajobe> B:|Jena2 Preview3 Released

12:45:43 <dc_rdfig> Titled item B.

12:46:31 <dajobe> B:see the [http://www.hpl.hp.com/semweb/jena2.htm|Jena2 page] (although the download isn't there right now)

12:46:31 <dc_rdfig> Added comment B1.

12:47:00 <dajobe> B:so get it from the [http://sourceforge.net/projects/jena/|Jena sourceforge site]

12:47:00 <dc_rdfig> Added comment B2.

13:00:29 <AndyS> B: Download: [http://prdownloads.sourceforge.net/jena/Jena-2p3.zip?download|Jena2 preview 3]

13:00:29 <dc_rdfig> Added comment B3.

13:01:10 <AndyS> dajobe takes about 8 mins from the release message to chump it! What took so long? :-)

13:02:20 <edd> coffee break

13:02:26 <dajobe> mmm

13:03:19 <dajobe> boston latte, very nice

13:07:29 <dajobe> hey, it is in chefmoz http://chefmoz.org/United_Kingdom/England/Bristol/Boston_Tea_Party960907218.html

13:14:57 <dajobe> B:sourceforge is currently sucking, so be prepared to wait

13:14:57 <dc_rdfig> Added comment B4.

13:15:05 <dajobe> s/currently//

13:30:14 <AndyS> We have all sorts of SourceForge experiences. upload was incredibly fast today. Quick release system couldn't cope.

13:30:37 <dajobe> the upload server is just one piece of it

13:30:44 <dajobe> the d/l thing was 20k/s for me

13:30:58 <AndyS> We have had CVS file damage (the support team fixed that for us).

13:31:04 <dajobe> eek

13:31:19 <AndyS> Don't panic - just a jammed file. No loss!

13:32:13 <AndyS> D/L was much faster than that when I tested it 90mins ago.

13:32:36 <AndyS> but it is bursty.

13:32:42 * dajobe uses jena cvs read access also for the latest bits

13:32:55 <dajobe> what jdk(s) do you test with?

13:36:22 <AndyS> (Java 1.3, 1.4) x (Linux RH 7/8/9, WinXP, Win2k) x (MySQL 4.1, 4.0?) but may be not everyone Release Candidate :-)

13:37:05 <dajobe> my redhat9 only has mysql 3.23

13:38:08 <AndyS> I understand that the DB is not updated unless you explicitly ask because of compatibility with existing databases.

13:38:32 <dajobe> ah

13:42:34 <dajobe> ok, the test worked

13:42:43 <dajobe> but there are 1574 tests, not ~1400 as the readme says

13:45:42 <_joshua> with a name containing "Boston" it ought to be in boston

13:46:08 <dajobe> er...

13:46:21 <_joshua> i am always annoyed when i pass the "brooklyn diner" because it's not in brooklyn

13:46:43 * dajobe fights java classpaths, looses

13:46:59 <_joshua> as a computer scientist, i like restaurants that name themselves after their addresses, but i don't like it when those restaurants move to a new location and keep their old names.

13:47:07 <_joshua> Just so you know.

13:47:32 <dajobe> they could put "302 Permanently Moved" on the old place

13:49:17 <sbp`> .google "San Pablo" "William Loughborough" site:w3.org

13:49:19 <datum> "San Pablo" "William Loughborough" site:w3.org: http://lists.w3.org/Archives/Public/www-rdf-interest/2001Jun/0129.html

13:49:59 <sbp`> there ya go. "40th and San Pablo Furniture Warehouse - where the name is the address"

13:50:34 <libby> .time

13:50:34 <datum> Thu, 15 May 2003 13:50:34 GMT

13:51:48 <dajobe> AndyS: re release number separate from notes. That's why I build mine from HTML, and have a release step checker that makes sure it's OK

14:11:44 <AndyS> We *should* have the config/changes/etc in RDF and produce all the files. But we don't :-)

14:11:56 <dajobe> possibly

14:12:10 <dajobe> seems like ant could do with some help for that

14:13:50 <danbri> danbri is now known as danbri_lurk

14:14:07 <AndyS> Anyway - I did manage to change the jar file names! FAQ-0: Set the class path. FAQ-1: And now put icu4j on the path.

14:15:24 <dajobe> I didn't see anything (ant target) to build the demo progs (util progs?)

14:15:37 <dajobe> I can't get rdfparse or some of the others to build at the moment

14:15:43 <dajobe> seems the parser needs SWING!

14:22:56 <AndyS> Could you be more specific? Is something missing?

14:23:19 <AndyS> Everything should be prebuilt into jena.jar.

14:24:37 <libby>http://lists.w3.org/Archives/Public/www-rdf-rules/2003May/0009.html

14:24:37 <dc_rdfig> C: http://lists.w3.org/Archives/Public/www-rdf-rules/2003May/0009.html from libby

14:24:50 <libby> C:|agenda for rdf query testcases meet

14:24:50 <dc_rdfig> Titled item C.

14:24:57 <dajobe> $ javac -classpath lib/jena.jar src/jena/rdfparse.java

14:24:57 <dajobe> src/jena/rdfparse.java:11: package junit.swingui does not exist

14:24:57 <dajobe> import junit.swingui.TestRunner;

14:24:59 <AndyS> There isn't a specific target just for the command line apps or for ARP.

14:25:14 <dajobe> src/jena/rdfparse.java:77: cannot resolve symbol

14:25:18 <libby> C:[http://www.timeanddate.com/worldclock/fixedtime.html?day=15&month=5&year=2003&hour=14&min=30&sec=0&p1=0|2003-05-15, 1430UTC]

14:25:18 <dc_rdfig> Added comment C1.

14:25:25 <libby> i.e. in 5 minutes

14:25:46 <dajobe> dajobe has changed the topic to: RDF/SemWeb hack'n'chat http://rdfig.xmlhack.com/ - RDF Query Testcases meet 14:30 UTC Thursday

14:25:58 <AndyS> Put junit.jar on your path. Put all the jars (except jena.jar obviously) to compile. Or use jena.jar

14:26:04 <libby> BLURB: Actions update (rdf query testcases agenda item D)

14:26:04 <dc_rdfig> D: Actions update (rdf query testcases agenda item D) from libby

14:26:31 <libby> BLURB: Report on Topic maps QL and RDF QL BOF at XMLEurope 2003 (rdf query testcases agenda item E)

14:26:31 <dc_rdfig> E: Report on Topic maps QL and RDF QL BOF at XMLEurope 2003 (rdf query testcases agenda item E) from libby

14:27:06 <libby> BLURB: Where are we up to/preparation for WWW2003 birds of a feather meeting (BOF) (rdf query testcases agenda item F)

14:27:06 <dc_rdfig> F: Where are we up to/preparation for WWW2003 birds of a feather meeting (BOF) (rdf query testcases agenda item F) from libby

14:27:30 <dajobe> AndyS: ok, adding junit worked. Not quite sure why I need it for a util

14:27:36 <AndyS> Or use the ant build script. Why do you want to build just one file?

14:27:46 <libby> D:[tests in ntriples|http://lists.w3.org/Archives/Public/www-rdf-rules/2003May/0010.html]

14:27:46 <dc_rdfig> Added comment D1.

14:28:10 <AndyS> Its not a util - its a command line tool - it will run the tests for you if you want.

14:28:23 <libby> F:[http://esw.w3.org/topic/RdfQueryTestingBudapestMeeting|ESW wiki Budapest meeting page]

14:28:23 <dc_rdfig> Added comment F1.

14:28:39 <libby> E:[http://esw.w3.org/topic/TMQLBOF|TMQLBOF]

14:28:39 <dc_rdfig> Added comment E1.

14:29:02 <dajobe> AndyS: ant doesn't build the src/jena stuff

14:29:07 <libby> .time

14:29:07 <datum> Thu, 15 May 2003 14:29:07 GMT

14:29:59 <AndyS> "Doesn't"? In the sense of fails to or in the sense of has no target?

14:30:11 <AndyS> dajobe: <target name="compile-apps" ...

14:30:17 <libby> time for meeting.....

14:30:28 <libby> -------- RDF query tests meet -------

14:30:39 <libby> sorry to butt in...

14:31:21 <libby> C:attending [libby miller|http://ilrt.org/people/libby]

14:31:21 <dc_rdfig> Added comment C2.

14:31:37 <libby> could attendees add their names as I've done please

14:31:55 <libby> hey alberto

14:32:30 <alberto> hello everybody

14:32:52 <libby> ok, let's get started

14:33:02 <alberto> C:[alberto|http://www.asemantics.com]

14:33:02 <dc_rdfig> Added comment C3.

14:33:07 <libby> 1. Actions update (rdf query testcases agenda item D)

14:33:20 <AndyS> C: [Andy Seaborne|http://www.hpl.hp.com/semweb]

14:33:20 <dc_rdfig> Added comment C4.

14:33:36 <libby> I had an action:

14:33:53 <libby> ACTION libby convert squish and RDQL jena tests to ntriples and

14:33:53 <libby> circulate by date of next meet

14:34:21 <libby> Ive done this (see chump); and alberto has also providede tests. thanks :)

14:34:44 <libby> not working completely when I test them on inkling, but 60-75% hit rate

14:34:57 <alberto> yes, I am almost done with rdf:parseType="Collection" here - hope to run your tests soon

14:35:04 <libby> :)

14:35:23 <libby> danbri alos reports that he is waiting until he has impleemnted collections

14:35:37 <libby> I've got a couple of questions.....

14:36:00 <alberto> not too hard but not clear enough from specs how to do it - looking at generated n-triples instead :)

14:36:09 <alberto> go for them libby

14:36:14 <libby> andy, alberto, do you have the orginal manifest file for the rdql jena tests? several from teh perl generated versions are missing number rows...

14:37:01 <libby> this was the 'primary' manifest file I had: http://swordfish.rdfweb.org/rdfquery/tests/tests/rdql-tests-2003-04-10/Manifest.rdf

14:37:09 <AndyS> In the Jena2 preview3 release :-) I'll find the CVS URL ...

14:37:15 <libby> - some of the numrows are missing, maybe in the perl conversion

14:37:19 <libby> thanks andys

14:37:48 <alberto> yes, Andy emailed his jena2 manifest to us some time ago....

14:38:06 <alberto> it could be the perl conversion I wrote....

14:38:13 <libby> most of the errors I get seem to be inkling's fault, but some from rdql jena are because I don;t support datatypes

14:38:25 <alberto> uahu! p800 :) hi dan

14:38:38 <libby> do we want datatypes tests in the 'core' framework?

14:38:39 <danbriP800> hi

14:38:53 <AndyS> Ugly url alert ....

14:39:05 <AndyS> Jena RDQL manifest : http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/*checkout*/jena/jena2/testing/RDQL/rdql-tests.n3?rev=HEAD&content-type=text/plain

14:39:13 <libby> thanks AndyS

14:39:25 <alberto> libby, datatypes are also in the resul set of Andy

14:40:01 <libby> yes alberto - the slightly old arp I'm using just ignores them - doesn;t mean it doesn;t work

14:40:50 <libby> I don;t think it's unreasonable to include datatypes tests in, I guess I'm askign for a quote on it

14:40:57 <alberto> my backend supports them but not the parser yet...need more work on that too - can not we simply use <rs:size/> without rdf:datatype for the moment?

14:41:33 <libby> what do you think AndyS?

14:41:46 <alberto> but we will need to add support for rdf:datatype sometime in the future anyway

14:41:49 <AndyS> Where does it use a datatype?

14:42:01 <alberto>http://www.w3.org/2003/03/rdfqr-tests/recording-query-results.html

14:42:01 <dc_rdfig> G: http://www.w3.org/2003/03/rdfqr-tests/recording-query-results.html from alberto

14:42:13 <libby> in the resultset AndS - number rows

14:42:25 <libby> but also (separately) 4 of your tests test for datatypes

14:42:38 <AndyS> Ah - I ignore row numbers because I do graph testing.

14:42:47 <libby> right

14:42:52 <libby> still handy though

14:43:00 <libby> at least as a preliminary check

14:43:24 <libby> alberto, does the datttypes in resultset mean that you can;t parse a resultset at all?

14:43:25 <AndyS> Except it consumes the results to count them!

14:43:54 <alberto> g:uses rdf:datatype in the example rs:size property

14:43:56 <alberto> lucky :-)

14:43:58 <alberto> and you got ARP fully aligned with specs too

14:44:22 <alberto> anyway, original question of libby about jena manifest and perl converter - need to be checked

14:44:38 <libby> let me make a proposal

14:45:05 <AndyS> People can convert to pre-DT as transition - for the long term ought we really to do things properly?

14:45:06 <alberto> ok

14:45:13 <libby> PROPOSED (p1): to have datatypes tests in the 'core' set of simple conjnctive queries

14:45:35 <libby> PROPOSED: (p2): to continue to use datatypes in the resultset format

14:45:35 <dc_rdfig> Label PROPOSED not found.

14:45:46 <libby> any seconders for either of these?

14:45:51 <AndyS> Well - it isn't a complete test if some RDF graph matching features aren't tested.

14:46:02 <AndyS> We also should talk about lang tags.

14:46:10 <alberto> fine - I will need to implement those anyway

14:46:22 <danbriP800> yes 2 p1

14:46:26 <AndyS> Why not mark tests as to the features tested and impls can skip

14:46:46 <alberto> lybyy, which datatytpe test are you referrring to? do you have a URL to your rep?

14:47:00 <libby> we would need some additional manifest properties, andys

14:47:16 <alberto> we should have thing like RDF test cases like approved/no-approved

14:47:22 <libby> ERROR! file:queries/nt/test-B-01.nt 1 2

14:47:22 <libby> ERROR! file:queries/nt/test-B-02.nt 1 2

14:47:22 <libby> ERROR! file:queries/nt/test-B-03.nt 0 2

14:47:22 <libby> ERROR! file:queries/nt/test-B-04.nt 1 2

14:47:31 <alberto> ok thanks libby!

14:48:03 <libby> in here: http://swordfish.rdfweb.org/rdfquery/tests/tests/rdql-tests-2003-04-10/queries/nt/

14:48:54 <libby> PROPOSED (p1): to have datatypes tests in the 'core' set of simple conjnctive queries and to indicate that they are datatypes somehow (for example in the description)

14:48:58 <libby> oops

14:49:04 <libby> PROPOSED (p1.1): to have datatypes tests in the 'core' set of simple conjnctive queries and to indicate that they are datatypes somehow (for example in the description)

14:49:04 <alberto> just realised that those tests need rdf:datatytpe support into n-triples parser and I have it already :-) good

14:49:17 <libby> heh, I don't...but I should

14:49:20 <AndyS> Suggest that tests try to have most coverage and systems can say which they implement.

14:49:22 <dajobe> heh

14:49:33 <libby> seconders for p1.1?

14:49:50 <AndyS> This might not even need explicitly need recording. Its impl specific.

14:49:55 <danbriP800> indicate in a machine friendy way

14:50:00 <alberto> yes good idea - we need to flag tests with some kind of URL/property

14:50:19 <alberto> point back to RDF test cases URLs perhaps?

14:50:23 <libby> woudl anyone care to take an action to think about a machine-readable way of specifyign the datatypes thing?

14:50:31 <_joshua> ERROR!

14:50:37 <libby> heh

14:50:42 <AndyS> Any tests on real world data will be hard to classify - depends on the data

14:50:47 <_joshua> anyone know how I would convert something like "S&#227;o Paulo" to UTF8?

14:51:28 <AndyS> Libby - are you going to support datatypes anyway?

14:51:38 <libby> 5 guess so...at some point...

14:51:48 <libby> I think maybe there's a wider issue potentially

14:52:11 <libby> if we could come up with a good way of classifying the tests, could reuse the format for more complex query types

14:52:48 * ericP takes over speaking for danbri and myself

14:53:15 <libby> ok, well there seemed to be come consensus for (p1) - keep the datatypes queries, don;t necessarilty specify that they are in some sense special

14:53:19 <ericP> indeed. many tests can claim to use an interaction of several features

14:53:23 <AndyS> If there were a std, it might have levels of compliance. But feature-by-feature choices tends to grow quite large.

14:53:41 <ericP> eg. datatypes+negation+literal substring

14:53:46 <libby> any seconders for p1? can always do seomthign better later

14:53:52 <alberto> joshua - there is some perl module on CPAN for that (can not remember which one)

14:54:00 <libby> (I can;t take on more actions at this stage)

14:54:00 <danbri_lurk> p1 yes i 2nd

14:54:01 <alberto> ok libby for p1

14:54:12 <libby> any objections?

14:54:25 <AndyS> Not clear here about the "indications"

14:54:34 <ericP> p1 vs. p1.1?

14:54:39 <libby> yep ericp

14:54:49 <AndyS> Description is fine - properties is confusing IMHO.

14:54:50 <libby> andys, no 'indications' in p1

14:55:14 <libby> p1 can have descriptions, but not necessarily

14:55:17 <AndyS> Was at 15:49! Or was that p1.1

14:55:33 <alberto> ok andy, how then you code would guess whether or not a test is about datatype for example? free-text?

14:55:45 <AndyS> P1 is OK - as much detail in description is good anyway,

14:55:50 <libby> this oen: PROPOSED (p1): to have datatypes tests in the 'core' set of simple conjnctive queries

14:55:58 <ericP> i think a vote for p1 will allow us, but not constrain us, to add a description onotlogy at our leasiure

14:56:03 <ericP> +1 on p1

14:56:05 <libby> yep

14:56:14 <libby> ok, well I think everyone is ok with that

14:56:14 <ericP> (as opposed to p2)

14:56:19 <AndyS> Persoanlly, I would run my imple manifest - not the master one - if I didn't have all features.

14:56:45 <libby> D:RESOLVED: (p1): to have datatypes tests in the 'core' set of simple conjnctive queries

14:56:45 <dc_rdfig> Added comment D2.

14:56:56 <libby> ok, any opinions on P2?

14:57:10 <libby> - PROPOSED: (p2): to continue to use datatypes in the resultset format

14:57:16 <AndyS> then it is an explicit process to include/exclude tests. Just my style.

14:57:23 <libby> does this cause people deadful problems? (not me)

14:57:42 <ericP> nor i

14:57:50 <libby> alberto?

14:57:52 <AndyS> Any sense of when "most" toolkits wil have datatypes? Lang tags?

14:58:01 <libby> hm

14:58:04 <libby> not sure

14:58:12 <AndyS> Is it a significant barrier? I'm not a good person to ask!

14:58:21 <alberto> ok andy, but if we want some interoperability of tests we need to agree how people can distinguish between basics and advanced tests

14:58:50 <libby> I think we can look at this later

14:59:00 <ericP> we're not on p2 here, are we...

14:59:05 <libby> I am!

14:59:16 * ericP follows libby's fearless lead

14:59:28 <AndyS> I see tests as encouraging systems to interoperate

14:59:54 <AndyS> So P2 is good - subsets of the tests cause fragmentation.

15:00:02 <ericP> p2, datatype format is the ^^datatype in ntriples

15:00:03 <ericP> ?

15:00:11 <alberto> libby, true - let's get all three/four squishql/rdql implementations running on libby tests rep and the we will look at the problems

15:00:20 <AndyS> eric P - yes

15:00:47 <ericP> is there an option to P2?

15:01:01 <ericP> besides inventing a new encoding?

15:01:06 <grault> grault is now known as earle

15:01:15 <libby> well, we could just leave out the datatype bit

15:01:20 <alberto> not problem for me with datatypes in the result set - I will find a solution

15:01:23 <libby> I was thinking of numrows

15:01:41 <danbri_lurk> p2: i think it is reasonable to assume a datatype capable parser, or at least one that'd silently ignore dts and give you the string form ( blush )

15:01:48 <libby> ok, thanks. any objections?

15:02:01 <AndyS> OK - for <rs:size> remove the range constraint.

15:02:24 <AndyS> I can see this as more inconvenient but the test results should cover datatypes.

15:02:27 * ericP doesn't understand this size discussion

15:02:35 <ericP> does size matter?

15:03:21 <ericP> seriously, what's the size thing? something i didn't get from staring at the cover page of the new syntax spec?

15:03:24 <libby> i.e. http://swordfish.rdfweb.org/rdfquery/tests/tests/squishtests2003-04-24/results/rs2.rdf

15:03:46 <AndyS> EricP - see the result set vocabulary. It defines size with range xsd:integer

15:03:52 <alberto> andy, I am open here - if most people like the <rs:size rdf:datatytpe="...#integer">12</> approach I will work to get it working - no problem

15:03:53 <libby> because (I at least) havn;t got a graph matching thing, number rows iis a useful if minimal way of matching resultsets

15:04:03 <ericP> size, roger - tx

15:04:09 * libby too

15:04:15 <libby> ok, and objections p2?

15:04:26 <libby> s/and/any/

15:04:33 <libby> (we should move on...)

15:04:41 <AndyS> P3: change range constrint on rs:size to be literal

15:04:41 <dc_rdfig> Label P not found.

15:04:54 <alberto> fine, let's move on - next question/s

15:05:15 <libby> not sure we need p3....

15:05:37 <alberto> ok I have another question then

15:05:41 <libby> ok

15:05:44 <libby> go ahead

15:06:14 <AndyS> P3 => can write results with no datatypes at all if Q /data has no datatypes.

15:06:15 <alberto> do we need to have in the manifest some way to say which vars/bindings are being asked?

15:06:46 <libby> in the manifest rather than the resultset?

15:06:49 <AndyS> Prob. not in the manifest - its in query and in the results.

15:06:54 <AndyS> Do you have a use case?

15:07:06 * libby seconds P3

15:07:15 <alberto> libby, this is related to the problem you have with one of my tests (0 rows but just n-triples would give 19 matches)

15:07:31 <libby> you have to load in 2 files rather than one to actually make the query

15:07:37 <libby> - can still do it though

15:07:43 <alberto> andy, let me find the URL...

15:07:47 <libby> my code is a bit crufy in that respect

15:07:57 <godseye> does anyone have a log of the geospatial chat?

15:08:10 <ericP> i think we should peper the rs graph with datatypes when they are from XSD or rs namespaces

15:08:16 <libby> - if it doesn't have any variables to use, it has a backup of all of them. so my error

15:08:27 <alberto> http://swordfish.rdfweb.org/rdfquery/tests/tests/rdfstore-tests-2003-05-14/Manifest.rdf and see test num 13

15:08:27 <dc_rdfig> H: http://swordfish.rdfweb.org/rdfquery/tests/tests/rdfstore-tests-2003-05-14/Manifest.rdf from alberto

15:09:46 <alberto> H:see test number 13 which tells 0 numrows

15:09:46 <dc_rdfig> Added comment H1.

15:09:48 <libby> so here, alberto's test says, give me the results for a variable that I did not initially ask for, and so shuld get 0 reesults

15:10:18 <AndyS> Sorry Alberto - don't get it. Why does the manifest need the variables?

15:10:28 <libby> but alberto, since that info is in the resultset and the manifest points to a resultset, can get informtaion from there

15:10:46 <AndyS> For a stray variabe, get rs:undef surely.

15:10:48 <godseye> ok, better question- what was the date of the geospatial discussion?

15:11:12 <alberto> H:the original query instead is like select ?y from <something.rdf> where (?x,<rdf:type>,<rss:item), (?x,<rss:title>,?title)....

15:11:12 <dc_rdfig> Added comment H2.

15:11:38 <libby> see here somewhere for some links godseye: http://lists.w3.org/Archives/Public/public-esw/2003May/0035.html

15:12:28 <libby> the issue ios perhaps, how we contrcut the full query from the ntriples and the bits and pieces in teh manifest and resultset

15:12:33 <AndyS> ?y is rs:undef and result-set has list of query vars. I seem to be mssing something here Alberto.

15:12:40 <libby> which variabless we want ius a very iportant aspect

15:13:04 <libby> I think we need to move on shortly

15:13:22 <AndyS> OK - in the N-triples form there is less info.

15:13:52 <alberto> H:when you convert the query to n-triples you loose the info about which var/bindings was asked - the only place where is mentioned is the result set, which is not inlined into the manifest itself

15:13:52 <dc_rdfig> Added comment H3.

15:14:04 <alberto> andy, yes but if you simply use the manifest you do not have that info - just tq:numRows='0' which is meaning less

15:14:14 <AndyS> Could have extra infor in N-triples form on the <> resource.

15:14:24 <libby> H:you can grab them by loading the results though

15:14:25 <dc_rdfig> Added comment H4.

15:14:47 <libby> we need to move on....alberto, could you summarise this to the list?

15:14:53 <AndyS> Shouldn't have the size there anyway! Modularity says put in one place.

15:15:16 <alberto> ok fine - but then numrows should be removed too from manifest due we have already tq:status (false in this case)

15:15:22 <AndyS> Or at least the same property???

15:15:29 <alberto> ok - move on

15:15:37 <libby> thanks

15:15:40 <AndyS> (for WWW2003 ..)

15:15:54 <libby> sorry to rush you. I think there may be somethign important in what your saying alberto....

15:16:14 <libby> let's skip TMQLBof for now

15:16:26 <libby> I will send round a mail to the list, but I havn;t written it up yet

15:16:37 <libby> so: Where are we up to/preparation for WWW2003 birds of a feather

15:16:37 <libby> meeting (BOF)

15:16:40 <libby> ?

15:16:51 <libby> and - what do we want to do there?

15:16:57 <AndyS> Danbri around?

15:17:15 <libby> not really I don;t think

15:17:16 <ericP> oof, just stepped out for a moment

15:17:40 <libby> we have a page: http://esw.w3.org/topic/RdfQueryTestingBudapestMeeting

15:17:56 <ericP> ringing him...

15:18:25 <libby> [[why: continued collaboration on Test Case formats, methodology and datasets]]

15:18:36 <ericP> danbri says RdfQueryTestingBudapestMeeting has everything he knows

15:18:39 <libby> we need something practical and short do do in a lunchtime bof

15:18:49 <libby> any ideas for what that might be?

15:18:56 <libby> what are peoples' priorities?

15:19:10 <AndyS> Collecting datasets (inc large datasets) would be good. I can host a repository.

15:19:25 <ericP> data set of test results?

15:19:32 <libby> or as inputs?

15:19:46 <alberto> any, good one

15:19:50 <libby> - for higher powered tests, speedtests etc?

15:19:51 <alberto> s/any/andy/

15:19:54 <AndyS> Of data for queries. Inputs

15:19:59 <ericP> don't think we can do that during a lunch BOF, but we can set the scene for it

15:20:18 <libby> good idea andys

15:20:21 <AndyS> I'll collect CDs :-) Or use IR!!!!!!

15:20:27 <libby> heheh

15:20:34 <libby> we can certainly ask participants

15:20:46 <alberto> plus we should agree on the basic set/bunch of tests we need to include in the startup query "kit" :-)

15:20:52 <AndyS> What about getting non-squish systems involved? rule systems?

15:21:12 <ericP> my guess is that the best thing we can hope for in a short BOF is to get more people focused on this discusion forum.

15:21:27 <ericP> that would include telling people what we have to offer and what we are looking for.

15:21:38 <ericP> AndyS, you will be there?

15:21:39 <libby> - for that then, it might be good to set some prioirities for the future of the forum

15:22:11 <libby> do we want to do any demos? tests rolling past...? ;)

15:22:23 <AndyS> Yes, ericP - I'll be there. I volunteer for the "eating lunch" objective

15:22:33 <AndyS> s/objective/goal/

15:22:34 <libby> :)

15:22:35 <ericP> dibs on the desert

15:22:45 <ericP> maybe we get more than one?

15:22:47 <AndyS> Get in line!

15:23:06 <libby> shall we try and approach people beforehand, so we xcan get them to come, show them our approach?

15:23:14 <libby> see if it extends to other langs?

15:23:24 <tav|passedaway> tav|passedaway is now known as tav

15:23:25 <libby> ...get their comments....

15:23:41 <ericP> andy, when you say you wnat to collect test data, is this to exercise/round out the test cases, or tie it to compelling use cases?

15:24:17 <AndyS> A corpus would help with (1) performance (2) scale

15:24:25 <ericP> i can offer algae, and, if i get to it, a bit of cwm.

15:24:46 <libby> that'd be cool ericp

15:24:48 <AndyS> Common data would be needed for test cases and we have some synthetic and real stuff.

15:24:53 <alberto> libby, for demos, what do you have in mind?

15:25:07 <libby> I dunno alberto, nothing very exciting

15:25:09 <ericP> i think we want to separate these from the tests that are good fro finding bugs in query implementations

15:25:26 <libby> maybe just demonstrate that we can interoperate, maybe w slides only

15:25:32 <ericP> a bunch of little tests, good for saying "feature X doesn't work"

15:25:37 <alberto> loading a manifest URL and print out ok/not-ok ?

15:25:45 <libby> be good to have big set as well tho

15:25:52 <AndyS> eric - have you looked at the current testsuite? Any comments?

15:26:01 <ericP> a bunch of other tests good fro benchmarking and saying "see, we can do that".

15:26:11 <alberto> it would be nice to add an xml:base to the query manifest too

15:26:20 <libby> +1 alberto

15:26:23 <ericP> haven't looked for a long time. just responding the this irc text.

15:26:23 <AndyS> +1

15:26:28 <ericP> +1

15:26:41 <ericP> and lang, everything in the spec.

15:26:47 <alberto> then I get URL of a manifest and can resolve inputs/ queries/ results/ dirs

15:27:00 <libby> ok, well how about we update the meteing page with thse ideas, and see if more come up

15:27:03 <AndyS> Think alberto meant to make the RDf manifest portable.

15:27:26 <libby> yes portable good

15:27:50 <AndyS> Spec - as defined last Friday by core WG (change to DT+lang tag)

15:27:52 <alberto> right now we have them relative and if you pick up a URL for it you can not use unless assuming some base

15:27:53 <AndyS> !!!!

15:28:03 <ericP> no, portable bad, too many people to deal with if it's portable

15:28:19 <ericP> (joking)

15:28:23 <libby> hard to match the resultsets unless have base too

15:28:27 <alberto> did not see that :( what's bad about it andy?

15:28:30 <libby> using e.g. arp

15:28:45 <AndyS> No bad - but a change.

15:28:55 <alberto> I am used to those :-)

15:29:12 <alberto> too many actually ;-)

15:29:20 <alberto> joking...

15:29:23 <DanCon> .time

15:29:24 <AndyS> model9 is now illegal.

15:29:24 <datum> Thu, 15 May 2003 15:29:24 GMT

15:29:37 <libby> F:ACTION all: add names to the http://esw.w3.org/topic/RdfQueryTestingBudapestMeeting page

15:29:37 <dc_rdfig> Added comment F2.

15:29:42 <DanCon>http://www.economist.com/surveys/displaystory.cfm?story_id=1747329

15:29:42 <dc_rdfig> I: http://www.economist.com/surveys/displaystory.cfm?story_id=1747329 from DanCon

15:29:49 <AndyS> Trivial change though

15:29:53 <DanCon> I:|Paradise lost

15:29:53 <dc_rdfig> Titled item I.

15:29:59 <ericP> dancon, can you do some cwm work to use the test harness?

15:30:01 <libby> F:ACTION libby - update that page with ideas for discussion

15:30:01 <dc_rdfig> Added comment F3.

15:30:13 <libby> ok, we're out of time

15:30:23 <DanCon> I: a Nick Arnet moment... "cyber-gurus, framed by colourful PowerPoint presentations reminiscent of stained glass ..."

15:30:23 <dc_rdfig> Added comment I1.

15:30:51 <libby> although we're out of time, do we want to suggest a next meeting date now?

15:31:00 <libby> ...or leave it till after www2003

15:31:03 <libby> ?

15:31:16 * DanCon has no idea what ericp is talking about; just stopped by to bookmark my nick arnett moment. we now return you to your regularly scheduled #rdfig chat

15:31:20 <ericP> i'll be looking to chat with folks in budabest

15:31:43 <AndyS> At the BOF - get others involved

15:31:48 <alberto> libby, let's discuss that in Budapest - I will be there

15:31:54 <libby> ok, cool

15:32:03 <libby> then I think we're adjourned.

15:32:05 <ericP> danc, i volunteered to try to get cwm to produce rs test results. this is me trying

15:32:08 <ericP> .

15:32:17 <libby> thanks everyone

15:32:24 <ericP> cheers

15:32:43 <libby> E:action libby summarise in an email to www-rdf-rules

15:32:44 <dc_rdfig> Added comment E2.

15:32:46 <AndyS> eric P - I nearly did! Issues with header/trailer of the N3

15:33:01 <ericP> ah, so there is some invested work

15:33:21 <ericP> can you get around this with unix pipe tools?

15:33:26 <AndyS> but the core results rule was in archived email .

15:33:36 <ericP> kinda ugly, but still, more usable than nothing at all

15:33:49 <alberto> cool ericP!

15:33:51 <alberto> ok thanks libby

15:34:12 <libby> cheers alberto, andys, ericp

15:34:16 <AndyS> Could use multiple rules sets.

15:34:18 <ericP> "ugly" was referring to the unix pipe soln

15:34:24 <AndyS> Thanks Libby

15:34:31 <libby> I think we've got a long way with this, and will be nice to chat in budapest

15:34:58 <DanCon> I:cf [http://archive.ncsa.uiuc.edu/SDG/IT94/Proceedings/Overviews/arnett/Antinet.html|The Internet and the Anti-net], Arnett at WWW '94

15:34:58 <dc_rdfig> Added comment I2.

15:35:02 <libby> - irc can be a rather slow way of doing things!

15:35:06 <alberto> libby, I will work to get my code to run through your/andy's tests - hope to make it on time

15:35:19 <libby> cool alberto :)

15:35:27 <ericP> actually, even if some folks ran some of the test cases and hand-transformed them to rs, we'd be able to compare the tools.

15:35:31 <DanCon> I:stained glass connection is indirect... "Advertisers today, like the church in the Dark Ages, rely on a "push" model of providing information, presented as truth that is not to be questioned."

15:35:31 <dc_rdfig> Added comment I3.

15:35:36 * libby goies back to fiddling with tests

15:36:37 <libby> G:|Recording Query Results from RDF query testcases

15:36:38 <dc_rdfig> Titled item G.

15:36:55 <libby> H:|RDFStore query tests manifest

15:36:55 <dc_rdfig> Titled item H.

15:46:58 <_joshua> heyaschuler

15:50:39 <dajobe> is the meeting over?

15:51:12 <libby> yup

15:51:35 <dajobe> so elsewhere AndyS solved my jena2p3 problems

15:51:52 <dajobe> it was 'cos I need to include the full classpath with all the lib/*.jar jars

15:52:10 <dajobe> that's 7.2M of jars

15:52:22 <dajobe> hmm

15:55:20 <AndyS> And its an 11.5Mb download for a unzip-and-go version :-(

15:55:38 <AndyS> icu4j is over half the lib/ size

15:56:16 <AndyS> We need to have different downloads.

15:56:52 <AndyS> But it is extra work to maintain. For Jena1 one FAQ was using wrong Xerces despite the fact we shipped the correct one each time!

16:20:57 <libby> andys, sill about?

16:21:08 <libby> know anything about http://swordfish.rdfweb.org/rdfquery/tests/tests/rdql-tests-2003-04-10/results/result-1-07.n3?

16:21:21 <libby> says one result but no detail

16:22:39 <libby> ah one binding of no variables: http://swordfish.rdfweb.org/rdfquery/tests/tests/rdql-tests-2003-04-10/queries/test-1-07

16:24:19 <earle> earle is now known as grault

16:24:48 <AndyS> Hi Libby

16:25:21 <libby> I've worked it out AndyS - no worries

16:25:32 <AndyS> The new (and as Chris pointed out) correct answer is one result binding which is empty.

16:26:05 <AndyS> I have changed to using his graph engine.

16:26:31 <libby> right yep.

16:40:31 <danbri_lurk> danbri_lurk is now known as danbri_fr

16:48:04 <JimJibber> JimJibber is now known as JibberJim

17:22:54 <_joshua> hi

17:34:06 Topic now RDF/SemWeb hack'n'chat http://rdfig.xmlhack.com/ - RDF Query Testcases meet 14:30 UTC Thursday

17:34:06 Users on #rdfig: logger_3 anselm zoyd niq ndw Morbus verbosus Bedrock aminorex deusx inkel chrisc em godseye grove_ kms sbp` Esaj arnarl libby golbeck bijan eikeon _joshua dngor mdupont MarkB collord dc_rdfig idoru aml grault tav xower jeriko DanCon bitsko Wack logger_1 ericP eikco merriam grove xover deltab jordan Kake datum sandro sethl

17:34:06 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

17:49:42 <GabeW_> GabeW_ is now known as GbeW

17:49:44 <GbeW> GbeW is now known as GabeW

18:04:49 <DanCon> "I would appreciate a general cooling of

18:04:49 <DanCon> the rhetoric regarding LC, the delay of which is a bogey man whose appearance is now so regular as to be more boring than frightening" -- IanH to www-webont-wg

18:11:19 <ndw> LC?

18:14:24 <dajobe> last call

18:14:30 <dajobe> of OWL test cases I guess

18:15:35 <DanCon> yes

18:18:20 <ndw> ah.

18:18:22 <ndw> ty

18:18:25 <dajobe> so OWL takes a document-based approach for "what is an ontology"

18:18:37 <dajobe> once you get into triples, you don't know what you have

18:18:59 <dajobe> I'm reading that I guess owl:Ontology (the term, not OWL itself) isn't what I wanted

18:19:21 <dajobe> i.e a named set of (terms, axioms, ... list I forget )

18:19:59 <DanCon> nope, we still don't have a standard way to make such a thing (a formula/context/whatever) a 1st-class object

18:20:16 <DanCon> quoting is explicitly postponed in the webont issues list, I think

18:20:30 <dajobe> relating a term to an ontology needs quoting?

18:20:40 <dajobe> what's rdfs:isDefinedBy for (excusing the naming)

18:21:14 <dajobe> well, I know what it is for is rather vague, but people use it as a handy property between properties and <otherURLs>

18:21:20 <dajobe> sigh

18:21:27 <DanCon> hang on, a minute ago you said you wanted a list of axioms. isDefinedBy works just fine for relating a term to a document.

18:21:37 <dajobe> yes it is

18:21:49 <dajobe> (I don't mind about axioms per se)

18:22:01 <dajobe> jus tthat OWL and RDF provide nothing extra to say about that property use

18:22:23 <DanCon> what else is there to say?

18:22:39 <dajobe> "isDefinedBy works just fine for relating a term to a document."

18:22:54 <DanCon> but that's already said, no?

18:22:58 <dajobe> no

18:23:11 * DanCon is befuddled; checks RDFS spec...

18:23:23 <dajobe> ok it says indicate

18:23:31 <dajobe> but there's nothing but prose

18:23:41 <dajobe> "rdfs:isDefinedBy is an instance of rdf:Property that is used to indicate a resource defining the subject resource. This property may be used to indicate an RDF vocabulary in which a resource is described."

18:24:06 <DanCon> yes, and further... "It may be possible to retrieve representations of O from the Web"

18:24:22 <DanCon> that seems like reasonable standards-ese for "isDefinedBy works just fine for relating a term to a document."

18:24:30 <dajobe> maybe I'd like to see that in the OWL mapping to triples

18:24:31 <dajobe> is that too much?

18:24:41 <DanCon> to see what?

18:25:24 <DanCon> you want every OWL DL document to have explicit isDefinedBy triples for everything? (that's perhaps a reasonable practice; but to make it a technical requirement???)

18:25:31 <dajobe> when an owl term is defined it emits an rdfs:isDefinedby property with subject of the owl term, object the owl ontology name/uri

18:25:40 <dajobe> it's not even clear it is practice

18:25:42 <DanCon> "emits"?

18:25:45 <dajobe> I had a look in owl guide

18:25:52 <dajobe> whatever

18:26:05 <DanCon> the mapping works the other way... the mapping tells you what RDF triples you need to put in your document in order for it to be OWL DL

18:26:25 <dajobe> ok

18:26:40 <dajobe> so I guess best practice advice in doing this with say the wine ontology, would be nice

18:26:47 <dajobe> s/best practice// I don't know

18:27:13 <DanCon> hmm... interesting point; it came up in the DAML+OIL discussions; I wrote a contribution to a "cookbook" that we were hoping/planning to build back then...

18:27:47 <dajobe> I guess it's the same way that dc:title etc. are nice to add as owl:Ontology annotations, but not required

18:28:09 <dajobe> (maybe not title, but something like that)

18:28:23 <DanCon> you can even do stuff like { :DublinCoreProperty s:subClassOf [ owl:onProperty s:isDefinedBy; owl:hasValue <...dublin-core-schema>]}

18:29:11 <DanCon> .google connolly joint-committee isDefinedBy

18:29:13 <datum> connolly joint-committee isDefinedBy: http://www.daml.org/listarchive/joint-committee/0419.html

18:29:50 <dajobe> oh wow, owl guide points to Dublin COre's RDF Schema

18:31:17 <DanCon>http://www.daml.org/listarchive/joint-committee/0418.html

18:31:18 <dc_rdfig> J: http://www.daml.org/listarchive/joint-committee/0418.html from DanCon

18:31:28 <DanCon> J:|a recipie for ont:UnambiguousProperty, rdfs:isDefinedBy

18:31:28 <dc_rdfig> Titled item J.

18:31:35 <dajobe> J:this is exactly what I want

18:31:35 <dc_rdfig> Added comment J1.

18:31:38 <DanCon> J:Connolly to joint-committee May 2001

18:31:38 <dc_rdfig> Added comment J2.

18:32:04 <dajobe> J1:""

18:32:04 <dc_rdfig> Deleted comment J1.

18:33:05 <dajobe> J:this show using "rdfs:isDefinedBy to connect classes

18:33:05 <dajobe> and properties to ontologies" in DAML. It would be good to see this in OWL work as advice or more

18:33:05 <dc_rdfig> Added comment J2.

18:33:42 <dajobe> uh oh, chump bug

18:33:46 <dajobe> J:ignore this

18:33:46 <dc_rdfig> Added comment J3.

18:34:06 <dajobe> oops

18:34:32 <DanCon> I thnk I chumped the wrong message.

18:34:47 <DanCon>http://www.daml.org/listarchive/joint-committee/0419.html

18:34:47 <dc_rdfig> K: http://www.daml.org/listarchive/joint-committee/0419.html from DanCon

18:35:10 <DanCon> K:|an rdfs:isDefinedBy recipie

18:35:10 <dc_rdfig> Titled item K.

18:35:24 <DanCon> K:Connolly to joint-committee, 16 May 2001

18:35:24 <dc_rdfig> Added comment K1.

18:36:08 <dajobe> dajobe has changed the topic to: RDF/SemWeb hack'n'chat http://rdfig.xmlhack.com/ - wiki http://esw.w3.org/topic/FrontPage

18:40:13 <DanCon> C:also: [irc log|http://ilrt.org/discovery/chatlogs/rdfig/2003-05-15.html#T14-24-50]

18:40:13 <dc_rdfig> Added comment C5.

18:42:29 <dajobe> oh dear, I'm doomed

18:42:37 <dajobe> "I don't think of Dave Beckett as a Guide reader.

18:42:38 <dajobe> He's happy reading fairly turgid technical material for developers"

18:52:19 <zoyd> does anyone know of a dictionary feed/API .. to return dictionary meaning for words.

18:53:12 <zoyd> REST style would be nicer.

18:54:14 * DanCon uses google interactively...

18:54:36 <DanCon> e.g. http://www.google.com/url?sa=X&oi=dict&q=http://dictionary.reference.com/search%3Fq%3Dambidextrous%26r%3D67

18:55:04 <_joshua> /msg danbri yes, i really liked it. i was shocked at the part where bruce willis turns out to be a robot

18:55:07 <_joshua> FUCK oops

18:55:19 <DanCon> seems to be just a redirect to http://dictionary.reference.com/search?q=ambidextrous&r=67

18:56:13 * DanCon thinks IRC user interfaces leave a lot to be desired

18:56:40 <DanCon> when I want to send a private message, I have to stop, count to 3, and really concentrate on the syntax of what I just typed before I hit enter

18:57:05 <DanCon> the hamming distance from /msg to /me is too small

18:57:56 <_joshua> mmm, ham

19:23:43 <bouncybouncy> bouncybouncy is now known as _anselm

19:26:28 <sethl> is there a way in cwm to ouput only the new assertions it found via rules?

19:37:11 <AndyS> Danbri - you had a question from the SWAD-e phonce conf on language tags I believe.

19:37:49 <danbriP800> l did?

19:37:59 <danbriP800> ah yes

19:38:02 <sandro> interesting question, sethl. --filter would work if you only needed the rules to run one stage, but not for multiple stages.

19:38:27 <danbriP800> l8r!

19:38:56 <sandro> I don't know of a way to do it. I don't think cwm tags provenance in a way that would allow that.

19:39:30 <sandro> It would need to tag where triples came from OR do inference over two formulas as if they were one. I dont think it does either of those.

19:41:48 <AndyS> sandro - the primer slides have access control problems (non-memeber machine)

19:42:06 <AndyS> example: http://www.w3.org/2003/Talks/0520-www-tf1-a1-primer/slide7-0.html => 403

19:48:01 <sandro> thanks Andy.

19:48:31 <sandro> fixed; may take a few seconds to propagate

19:52:37 <sethl> sandro: thanks, I'll try to rewrite it to take advantage of --filter

19:53:02 <sandro> feedback appreciated of course. tutorial is tuesday, and we're mostly busy between now and then. got several more sections to write!

19:53:32 <sethl> here's another interesting one. when cwm has infered a new triple, it adds it to the store. Is there a way to delete the old triple, so that I don't have dupes in the store?

19:54:27 <sethl> such that I don't end up with: :a :b :c ; :b :d . when I have just infered :a :b :d .

19:54:58 <sethl> I would want to delete :a :b :c .

20:06:54 <sethl> sandro: using filter as you suggested worked great. cwm rules.n3 --filter=rules.n3 only spits out the new facts

20:06:57 <sethl> thanks!

20:07:18 <sandro> excellent.

20:07:44 <sethl> er... unless I have multiple rules, and I only care about the second rules... hmm..

20:08:08 <sandro> filter twice? *shrug*

20:10:12 <sethl> ok, got it. I split the last rule (the important rule) out to its own file. Then, I did python cwm.py rules.n3 --think --filter=rules2.n3

20:10:40 <sandro> great

20:11:10 <sethl> or, don't split it up, just do a --think before --filter. now we're cookin.

20:11:27 <sethl> now... to integrate CWM into my python program. must find SPI.

20:12:48 <sandro> spi?

20:13:16 <sethl> some sort of programming interface so that I can use cwm's llyn, primarily

20:13:23 <sethl> I think that's the part I'm interested in

20:13:25 <deltab> API?

20:14:07 <sethl> that would be better

20:14:54 <sethl> I hope I can divorce llyn enough to use it standalone

20:16:05 <sandro> should be doable.

20:16:47 <sandro> ive started try to clean that up, but am lacking bandwidth with tim.

20:17:26 <sethl> yeah, I tried to clean up its treatment of bNodes, but I couldn't find time to get foot in the door. no biggie.

20:21:27 <sethl> looks like a cool CWM/N3 tutorial for the conference

20:26:12 <sandro> we're trying

20:27:01 <sethl> I wish I could be there

20:30:38 <mdupont_> hi all!

20:30:53 <sethl> howdy!

20:31:16 <mdupont_> hey sethl,

20:32:01 <mdupont_> seth, which seth are you again?

20:32:15 <sethl> er... the cool one. :)

20:32:25 <sethl> not the sailor one

20:32:26 <mdupont_> heheh

20:32:37 <mdupont_> realmeasures?

20:32:41 <sethl> nope

20:32:51 <mdupont_> ok, the third, moderate one :)

20:33:05 <sethl> heh

21:05:16 <sethl> would the new tutorials be the most up to date place for n3 info? I'm looking for a definitive grammar of n3 (if there is one)

21:05:51 <inkel> inkel is now known as inkel_away

21:06:43 <dajobe> outside the cwm one, the jena n3 one was written most recently - using JavaCC

21:07:03 <mdupont_> hey dajobe

21:07:19 <dajobe> hi

21:07:26 <mdupont_> i have upgraded to the new redland

21:07:31 <mdupont_> it is nice and easy to use

21:07:32 <dajobe> there's a new one?

21:07:37 <mdupont_> 92

21:07:45 <mdupont_> i am a bit slow