00:02:09 <bijan> A:
00:02:09 <dc_rdfig> http://www.xml.com/pub/a/2001/07/25/prologrdf.html?page=1
00:02:10 <dc_rdfig> RDF Applications with Prolog
00:02:11 <dc_rdfig> (bijan) By Yours Truely.
00:02:12 <dc_rdfig> (bijan) And it's very very very long.
00:02:13 <dc_rdfig> (bijan) I must learn to write short articles about small topics that I don't have to figure out as I go along :)
00:02:14 <dc_rdfig> (bijan) But it has a cool line by line comparsion of a Prolog DCG transformation and the equivalent XSLT sheet.
00:02:15 <dc_rdfig> (bijan) Puts Prology, inferential queries side by side with XPath, structural queries.
00:02:16 <dc_rdfig> (AaronSw) dc:title "Foo" .
00:02:17 <dc_rdfig> (sandro) Is great for people who know Prolog already; I'll be surprised if many newcomers to Prolog (who the article addresses at many points) can really make sense of it.   Maybe very very smart ones.   :-)
00:02:20 <dc_rdfig> (danbri) There are a bunch of us on the borders of Prolog; we know what it does and how, but never really jumped in. I find it a nicely juggled piece, introducing Prolog and RDF folks to one another...
00:02:21 <bijan> Sorry.
00:02:23 <dc_rdfig> (sbp) I very much agree with Dan; I downloaded SWI-Prolog after Bijan's first article, but could only take it so far. This one opens up a lot of doors
00:02:34 <SeanP> SeanP has joined #rdfig
00:02:57 <sbp> sbp has quit
00:03:19 <bijan> A:Thanks to my able defenders. One point is that it's a *series*. If you go back to the first article, you'll find, Sandro, I think, that it gives you enough Prolog (and RDF) to follow the second.
00:03:20 <dc_rdfig> commented item A
00:03:30 <SeanP> SeanP is now known as sbp
00:03:48 <bijan> A:Indeed, I made that recommendation to someone already (who had no Prolog or RDF). Seemed to work.
00:03:48 <dc_rdfig> commented item A
00:17:01 <AaronSw> http://flarelang.sourceforge.net/
00:17:01 <dc_rdfig> M: http://flarelang.sourceforge.net/ from AaronSw
00:17:11 <AaronSw> M:|Flare Programming Language
00:17:11 <dc_rdfig> titled item M
00:17:38 <AaronSw> M:Brought to you by [the Singularity Institute for Artificial Intelligence|http://singinst.org/]
00:17:38 <dc_rdfig> commented item M
00:17:47 <bijan> bijan has quit
00:18:26 <AaronSw> M:An XML-based proposal for the first "annotative" programming language, inspired by LISP and friends
00:18:26 <dc_rdfig> commented item M
00:21:24 <AaronSw> M:"""Annotative programming, if done properly, has the potential to be the successor to object-oriented programming, in the same way that object-oriented programming succeeded procedural programming, procedural programming succeeded assembly language, and assembly language succeeded raw hexadecimal numbers."""
00:21:24 <dc_rdfig> commented item M
00:29:25 <sbp> M:Heh
00:29:26 <dc_rdfig> commented item M
00:31:00 <sbp> Why use Python/C++/Perl etc. when you can use annotative programming languages?
00:31:12 <AaronSw> I know! Let's go rewrite cwm.
00:31:17 <sbp> :-)
00:31:21 <AaronSw> ;-)
00:47:48 <sbp> * sbp wonders if all frogs found in sandwiches have to be named "Joe"
00:48:28 <tdxdave> tdxdave has quit
00:50:53 <sbp> q.v. http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Jul/0411
00:51:22 <AaronSw> But what color is it?
00:51:44 <sbp> If by that you mean, "what color is 'Joe'", I have no idea
00:52:12 <sbp> And are all frogs named Joe found in sandwiches of the same color?
01:04:47 <sbp> sbp has quit
01:07:55 <sbp> sbp has joined #rdfig
01:12:20 <AaronSw> * AaronSw hacks an HTTP archiver in Python
01:13:12 <sbp> Hey! I was gonna do that too
01:14:55 <AaronSw> Well you just wait until I finish mine...
01:15:06 <sbp> I might just do that
01:17:10 <danbri> danbri has quit
01:23:15 <AaronSw> * AaronSw considers building HTML Tidy into his archiver ;-)
01:24:05 <sbp> Heh!
01:24:17 <sbp> How far have you got? Where will the results be available, if at all?
01:24:58 <AaronSw> I've got a working proxy (hacked away at the MojoNation one).
01:25:07 <AaronSw> Results will be at: http://logicerror.com/archiverProxy
01:25:27 <sbp> Cool
02:03:37 <AaronSw> Version 0.1 of the archiver is at: http://logicerror.com/archiverProxy-code
02:03:57 <AaronSw> All it does right now is log to STDOUT.
02:04:26 <GabeW> GabeW has quit
02:04:39 <sbp> Great!
02:05:15 <tdxdave> tdxdave has joined #rdfig
02:07:40 <sbp> sbp has quit
02:56:42 <DanC_> * DanC_ waves
03:05:26 <AaronSw> * AaronSw waves
03:10:23 <tdxdave> tdxdave has quit
04:12:21 <DanCon> DanCon has quit
04:12:21 <em> em has quit
04:16:26 <GabeW> GabeW has joined #rdfig
04:16:34 <GabeW> GabeW has quit
04:17:25 <em> em has joined #rdfig
04:17:25 <DanCon> DanCon has joined #rdfig
04:17:30 <DanCon> DanCon has quit
04:17:30 <em> em has quit
04:17:31 <em> em has joined #rdfig
04:17:31 <DanCon> DanCon has joined #rdfig
07:14:42 <OpenKern> OpenKern has joined #rdfig
08:09:10 <basicbsd> basicbsd has joined #rdfig
08:09:49 <basicbsd> basicbsd has quit
09:37:44 <OpenKern> OpenKern has quit
11:01:12 <sandro> sandro has joined #rdfig
11:45:34 <barstow> barstow has joined #rdfig
12:09:31 <barstow> * barstow wonders why the July 26 scratchpad is missing from http://rdfig.xmlhack.com/index.html
12:57:22 <danbri> danbri has joined #rdfig
12:57:42 <danbri> * danbri waves, wonders whether to print out on dead trees for RDF f2f or stash files on laptop
13:00:53 <barstow> * barstow says save the trees :-).  BTW DanBri, if you get a chance, visit Muir Woods National Monument http://www.nps.gov/muwo/  It's not the rain forest you've mentioned before but should be very nice.
13:00:57 <em> * em is bringing sowa book for plane flight rereading by stashing all ofhter files on laptopn
13:01:23 <em> * em waves welcome back to barstow
13:01:39 <danbri> I was going to come into MIT to do some printing today; probably won't now. Busy thinking about model theory and anonymous nodes etc
13:02:10 <em> * em thinks danbri should get away from his laptop more often :)
13:03:08 <danbri> Well, when I work onsite @MIT I use my laptop there too! Might try using the Palm+keyboard for a bit...
13:03:52 <danbri> BTW palm fans: Libby made a version of the rdfweb dataset that works in PalmDAML. Very fun, though its the only thing that crashes my Palm.
13:31:28 <em> em has quit
13:33:41 <danbri> http://lists.w3.org/Archives/Public/xml-dist-app/2001Jul/0311.html
13:33:42 <dc_rdfig> N: http://lists.w3.org/Archives/Public/xml-dist-app/2001Jul/0311.html from danbri
13:33:57 <danbri> N:Msg from Glen about XML-Protocol issues
13:33:57 <dc_rdfig> commented item N
13:35:39 <danbri> N:|XML Protocol bindings, extensions and blocks discussion
13:35:39 <dc_rdfig> titled item N
13:40:40 <em> em has joined #rdfig
14:02:49 <danbri> danbri is now known as danb_scribe
14:04:19 <tdxdave> tdxdave has joined #rdfig
14:54:01 <danb_scribe> danb_scribe has quit
14:55:31 <danb_scribe> danb_scribe has joined #rdfig
15:10:29 <danb_scribe> http://ilrt.org/discovery/chatlogs/rdfcore/2001-07-27.txt
15:10:29 <dc_rdfig> O: http://ilrt.org/discovery/chatlogs/rdfcore/2001-07-27.txt from danb_scribe
15:10:42 <danb_scribe> O:|RDFCore 2001-07-27 raw IRC log
15:10:43 <dc_rdfig> titled item O
15:12:51 <danb_scribe> danb_scribe is now known as danbri
15:19:39 <sbp> sbp has joined #rdfig
15:22:01 <sandro> sandro has quit
15:26:46 <AaronSw> Hello
15:27:08 <danbri> Hello
15:30:08 <danbri> hmm, http://xml.apache.org/axis/ seems hosed 
15:30:20 <sbp> sbp has quit
15:40:37 <bwm> bwm has joined #rdfig
15:47:28 <sbp> sbp has joined #rdfig
16:00:38 <lasDesk1> lasDesk1 has quit
16:06:14 <lasDesk> lasDesk has joined #rdfig
16:23:36 <sandro> sandro has joined #rdfig
16:27:37 <DanC_> continuing saga of my broken palmpilot...
16:28:06 <DanC_> I added enough vcal export to pdkb.pl to transfer my data to korganizer
16:28:21 <DanC_> now I'm trying to integrate korganizer into my semantic web...
16:28:41 <sbp> your semantic web?
16:28:46 <DanC_> one step achieved: hacking on vcal2xml to get my data to be well-formed XML.
16:29:34 <DanC_> I have a (broken) palm IIIx.
16:29:47 <DanC_> I think it still computes, but the display is broke.
16:30:25 <sbp> Ugh. I've been looking for some Wiki software for Palm, and wondering if it could export it as RDF, but no such luck
16:30:51 <sandro> sandro has quit
16:30:53 <DanC_> wiki/palm integration was on my TODO list; wanted to sync my palm memos with a Zope-powered wiki.
16:31:07 <sandro> sandro has joined #rdfig
16:31:31 <DanC_> well, pdkb.pl could easily be extended to export memos as RDF.
16:31:55 <DanC_> pdkb.pl started out exporting the datebook as RDF, but I added addressbook export in a few minutes.
16:32:16 <sbp> Oh, and what's up with http://www.w3.org/2001/04/lrns ?
16:32:30 <sbp> (from one of your Datebook hacking pages)
16:32:50 <DanC_> lrns? oh yeah... it's an idea I have not gotten into tangible form.
16:33:30 <sbp> Something to do with QNames in rel/rev. I wrote an XSLT script some while ago that processes QNames in attributes
16:34:13 <AaronSw> QNames in REL/REV? Blogspace does that
16:36:33 <danbri> * danbri tries to repair mangled his PalmDAML installation (it's complaining 'no database' when I run it)
16:37:11 <danbri> anybody run into this? I deleted homework.pdb to save space; installed the compressed version. I'm now re installing homework.pdb, see if that makes it happy.
16:37:32 <danbri> Now I can bore people with RDF demos even without my laptop :)
16:37:38 <sbp> :-)
16:37:56 <DanC_> hmm... I'd like to say "good job" to this Narval guy that developed http://www.logilab.org/vcalsax/ but I can't find his email address!
16:39:10 <AaronSw> hmm, contact@logilab.com for the company
16:39:57 <DanC_> the by mont view of chump... http://rdfig.xmlhack.com/2001/07/index.html it shows the topics, which are rarely interesting. I'd like a few that shows all the items chumped in the month
16:40:42 <danbri> closest i've got to a general overview is queries against libby's database
16:41:21 <danbri> ie http://www.w3.org/RDF/Interest/#irc -> http://swordfish.rdfweb.org:8085/rdfig/index.html
16:41:55 <DanC_> L:at a glance, that does look cool. Hope to check it out in detail...
16:41:55 <dc_rdfig> commented item L
16:42:01 <DanC_> K:in-your-face URLs. grumble.
16:42:01 <dc_rdfig> commented item K
16:42:14 <danbri> * danbri wonders why the tomcat port's no longer hidden behind apache/proxy; suspects a redirect is screwing things up
16:43:45 <sbp> K:For printing
16:43:45 <dc_rdfig> commented item K
16:44:28 <DanC_> J:I wonder what value SOAP adds in this service; could it be done with traditional forms ?query=params technology?
16:44:29 <dc_rdfig> commented item J
16:48:06 <DanC_> * DanC_ checks in vcal2xml.pl capable of handling my korganizer calendar
16:48:23 <danbri> J:That was the purpose of the exercise! The value of the SOAP::Lite perl *module* to me was pretty clear; I got to pretend the network wasn't there (not always wise!). Apparently XML and SOAP let me do that, though another implementation might've used ?query=params
16:48:24 <dc_rdfig> commented item J
16:48:31 <sbp> sbp has quit
16:48:34 <sbp> sbp has joined #rdfig
16:48:48 <sbp> Hmm... if UAs supported CSS2, we could do: @media print { a:after { content: " (" attr(href) ") "; } }
16:50:08 <danbri> J:I felt unless I got my hands grubby with a SOAP app, I wasn't really qualified to wade into xml-dist-app discussions. So I thought I'd earn my opinions by seeing what, if anything, SOAP bought me.
16:50:08 <dc_rdfig> commented item J
16:50:52 <danbri> J:One win: I didn't have to define a data format for serialising the result-set structure; SOAP did that for me. Presumably in the (RDF-ish) Section 5 format.
16:50:53 <dc_rdfig> commented item J
16:52:17 <danbri> J:Actually Glen Daniels tells me that SOAP doesn't itself specify how to serialise associative arrays (maps, hashtables), but there's some implementor concensus around a certain design that uses SOAP's array structure. And that Apache/Axis and SOAP::Lite both use this.
16:52:17 <dc_rdfig> commented item J
16:52:44 <sbp> sbp has quit
16:55:10 <danbri> J:Might be instructional to compare this with [Pasqualino "Titto" Assini's RDF/protocol|http://www10.org/cdrom/posters/p1131/] approach.
16:55:11 <dc_rdfig> commented item J
17:00:46 <em> AaronSw, where is that online cwm service again?
17:00:54 <AaronSw> purl.org/swag/n3tordf
17:01:00 <em> * em start noodling on swws demo
17:01:04 <em> thanks!
17:03:02 <em> is there a way I can submit multiple urls?
17:04:54 <GabeW> GabeW has joined #rdfig
17:05:21 <AaronSw> yes, add them to the URL
17:05:34 <AaronSw> uri=foo&uri=foo
17:05:40 <em> ok
17:05:42 <em> thanks
17:25:30 <danbri> * danbri gets the grey screen of death (PalmDAML still crashing; giving up for now)
17:33:08 <em> arrggg.... the n3 export of http://www.w3.org/2001/07/26-swcal/week-31.2001.rdf yields http://www.w3.org/2001/07/26-swcal/week-31.2001.n3 which when reparsed in http://swag.semanticweb.org/n3tordf produces an error
17:33:56 <AaronSw> * AaronSw looks...
17:33:57 <bwm> bwm has quit
17:34:10 <danbri> are they running different versions of cwm? 
17:34:12 <AaronSw> Ahh, yes -- it's a cwm bug...
17:34:27 <AaronSw> cwm exports files with dashes in them but won't read them for some reason.
17:34:35 <AaronSw> i'll go fix it for you
17:34:43 <em> same error on my version as well...
17:34:48 <sandro> * sandro was driving behind a car with license plate "CWM 328" today.  :-)
17:34:56 <AaronSw> ;-)
17:35:39 <AaronSw> Hmm, looks like I already fixed that bug... odd.
17:35:48 <em> AaronSw, the error is where exactly?
17:36:06 <AaronSw> the line:
17:36:06 <AaronSw>     _namechars = string.lowercase + string.uppercase + string.digits + '_-'
17:36:12 <AaronSw> needs that dash on the end
17:36:15 <danbri> Does cwm have a bunch of tests that run against it? Eating its own exports should be an easy test to write...
17:36:24 <AaronSw> Yes, it does...
17:36:31 <em> yes to test cases... 
17:36:39 <em> 2000/10/swap/test
17:37:14 <em> * em try's feeding rdf file to cwm via AaronSw's service
17:37:24 <danbri> can we get it via CVS yet? my cwm installations are all hand HTTP GET'd, and I've not been pulling the subdirs...
17:37:36 <AaronSw> yes, danbri, see dev.w3.org
17:37:52 <em> AaronSw, 
17:37:54 <em> File "/web/le-dev/www/n3/cwm.py", line 2389
17:37:54 <em>     =======
17:37:54 <em>      ^
17:37:54 <danbri> cool; i thought it was in w3.org doc space for some reason
17:37:55 <AaronSw> i know
17:37:58 <AaronSw> I'm fixing
17:38:27 <em> danbri:, dev.w3.org and www.w3.org in 2000/10/swap are mirrored
17:38:40 <em> dev is behind about 10 min in updates
17:38:58 <AaronSw> OK, finally fixed.
17:39:04 <em> retesting
17:39:29 <em> excellent!
17:39:31 <AaronSw> Great!
17:39:33 <em> n3 aqnd rdf work
17:39:37 <em> err...
17:39:44 <em> rephrase
17:39:52 <em> n3 and xml representation of rdf work :)
17:40:09 <AaronSw> ;-)
17:40:10 <em> * em smacks himself for not clarifying this
17:40:50 <em> ok... now to try integrating results from Aarons service with libby's search tool 
17:43:41 <danbri> do you have a URL of some data file and a SquishQL query?
17:44:41 <GabeW2> GabeW2 has joined #rdfig
17:45:05 <GabeW> GabeW has quit
17:45:23 <GabeW2> GabeW2 has quit
17:45:24 <GabeW2> GabeW2 has joined #rdfig
17:45:34 <GabeW2> GabeW2 has quit
17:45:43 <GabeW2> GabeW2 has joined #rdfig
17:45:54 <GabeW2> GabeW2 has quit
17:46:12 <AaronSw> AaronSw has quit
17:46:27 <GabeW> GabeW has joined #rdfig
17:46:39 <em> em has quit
17:46:42 <danbri> danbri has quit
17:46:42 <em> em has joined #rdfig
17:47:02 <em> but will eventually fetch this file and  http://www.w3.org/2001/07/26-swcal/travel.rdf, run though filter and the results of which passed off and integrated into the SWWS queryhttp://swordfish.rdfweb.org:8085/swws/index.html 
17:47:49 <em> arrg... http://swordfish.rdfweb.org:8085/swws/index.html
17:48:04 <AaronSw> AaronSw has joined #rdfig
17:48:11 <em> still gives me servlet error loading http://www.w3.org/2001/07/26-swcal/travel.rdf
17:48:32 <danbri> danbri has joined #rdfig
17:48:58 <AaronSw> * AaronSw would really like it if the servers stopped doing that...
17:49:09 <bwm> bwm has joined #rdfig
17:49:10 <danbri> any idea what's up?
17:49:28 <AaronSw> server rehubbing
17:49:36 <em> danbri, http://swordfish.rdfweb.org:8085/swws/index.html is failing on all querries
17:49:40 <em> libby in flight?
17:49:58 <danbri> crap; libby fixed something up there real quick before leaving. Yes, think in flight already.
17:50:06 <em> damn
17:50:20 <em> and I almost had this working the other day!
17:50:42 <AaronSw> That's OK, i'm sure you can do it all with cwm.
17:50:54 <em> but thats not the point!
17:50:59 <AaronSw> ;-)
17:51:10 <lasDesk> lasDesk has quit
17:51:34 <AaronSw> err, what was the  point? did I miss it while I got disconnected?
17:51:43 <em> dajobe? jang? you guys around?
17:52:02 <danbri> chances are we can fix this from stanford. Do we have ethernet in our accomodation or just onsite for meetign? (eg. are these wired student rooms)
17:52:10 <danbri> no: same flight
17:52:15 <em> onsite for meeting
17:52:21 <danbri> ILRT's eggs in one basket ;-)
17:52:21 <em> not sure abut rooms
17:52:25 <em> :)
17:52:38 <AaronSw> well, we can have a hack-a-thon during eric's talk -- it'll be fun.
17:53:22 <em> yeah... i'm not getting to this demo until something like slide 20, you  guys should ahve plenty of time
17:53:25 <em> :)
17:53:33 <danbri> I've got access to libby's server, just need to know what to type. shouldn't be a big deal to fix.
17:53:59 <danbri> * danbri just got Apache Axis/SOAP running, remembering the joys of java classpathing...
17:54:19 <AaronSw> I don't know java but it looks like something pretty easy to fix -- just need to convert the postgresql results into jdbc or some such
17:54:21 <em> danbri, can you tell if anything is gone south?
17:54:34 <em> performance 100%, etc.
17:55:22 <danbri> box is fine; tomcat servers misconfig'd I think.
17:55:33 <em> hmm... 
17:55:48 <em> :8085 is your servlet envionrment
17:55:54 <danbri> yeah
17:56:09 <em> http://swordfish.rdfweb.org:8085/rdfquery/index.html
17:56:09 <dc_rdfig> P: http://swordfish.rdfweb.org:8085/rdfquery/index.html from em
17:56:13 <em> seems to be working fine
17:56:18 <em> arrgg!!!
17:56:31 <AaronSw> Eric, you know java... fix it: http://swordfish.rdfweb.org:8085/swws/query.txt
17:56:37 <em> which is also in your servlet context
17:56:38 <AaronSw> you've got full source
17:57:03 <danbri> it's running against a postgresql or mysql store though, isn't it?
17:57:05 <em> AaronSw, thats a jsp file... i think the problem is behind this
17:57:29 <AaronSw> well, some source... ;-)
17:57:49 <em> i'm not sure..  http://swordfish.rdfweb.org:8085/rdfquery/index.html i think is in memory, correct?
17:57:52 <danbri> * danbri snoops libby's bash history
17:58:09 <danbri> the live queries from the web operate in memory, yup
17:58:16 <em> i was under the impression that http://swordfish.rdfweb.org:8085/swws/index.html was in memory to
17:58:27 <em> nothing to do with DB's at all
17:59:03 <danbri> * danbri doubts it, though not sure. There was some nifty rdfq-to-sql code floating around...
17:59:10 <em> loads the SWWS calander information in rdf, and merge with (e.g) my calander, search responses, etc.
17:59:24 <AaronSw> the fact that i get a postgresql resultset error would sort of indicate a database, right?
17:59:33 <AaronSw> """java.lang.ClassCastException: postgresql.ResultSet"""
17:59:37 <danbri> right.
17:59:51 <bwm> bwm has left channel
17:59:52 <em> well, yeah... but i thought it was slightly different
18:00:29 <em> so i'm now wondering if libby is building small postgress DB's, infusing data, quering for results and then destroying database
18:00:51 <em> and the fact that this was listed on www.semanticweb.org/swws page got too many hits and shut this down
18:01:41 <danbri> I don't think so, what i've seen either is in-mem or a longlived pre-established sql db
18:01:44 <em> we're all gussing here, not knowing how she's got this set up...
18:01:57 <AaronSw> umm, look at the source, em?
18:02:06 <danbri> I can try shutting down and restarting tomcat
18:02:11 <AaronSw> she seems to have code for both: in-memory db and jdbc...
18:02:14 <em> pre-established sql db, but can infuse more data? e.g. You can also search this data with other xml/rdf data; put the url(s) here (I've
18:02:15 <em> included an invented example)
18:02:42 <em> danbri, that might be a start... can you do this
18:03:23 <em> AaronSw, yes... just don't know which she was using for this demo :)
18:03:34 <em> perhaps different use of time...
18:04:43 <AaronSw> aha, seems to cache contents of db in memory
18:07:09 <danbri> I don't think its a good use of our time trying to fix this now; I've tried a few things but it's not my box so wary of fiddling w/ servers. We'll live.
18:07:34 <em> yep... i've gone back to working on presentation and workflow demo
18:08:12 <em> the calander stuff would be cool, i'll try and get as much stuff together without being able to test it as I can
18:08:34 <em> AaronSw, please don't shut down you service over the next several days :)
18:08:38 <AaronSw> :-)
18:08:51 <AaronSw> * AaronSw wonders exactly how many servers em's demo will depend on...
18:08:58 <em> 3
18:09:00 <em> www.w3.org
18:09:15 <AaronSw> well, that's good
18:09:16 <em> swordfish.rdfweb.org
18:09:22 <em> logicerror.org
18:09:29 <AaronSw> err, .com
18:09:37 <em> :)
18:09:49 <AaronSw> * AaronSw isn't out for profit -- really!
18:10:43 <em> slurp files and rules off www.w3.org via logicerror, pass results into rdfweb.org, and query via sqish to find out when the best time DanC and I can shoot some pool
18:11:06 <AaronSw> :-)
18:11:47 <AaronSw> this is what life will be like in the future: dependent on some programmer in a plane
18:11:54 <em> ooppss... also slurp in sw calander to boot... so ilrt.org is another dependant
18:12:59 <danbri> ilrt.org: bash-2.01$ uptime
18:12:59 <danbri>   7:12pm  up 385 day(s),  9:47,  2 users,  load average: 0.07, 0.04, 0.05
18:13:10 <em> excellent :)
18:13:19 <AaronSw> Hmm...
18:13:36 <AaronSw> Bob Truel, Dmoz.org: [[[We have not changed the RDF file format, and currently
18:13:36 <AaronSw> have no plans to do so.]]]
18:13:40 <em> 385 day uptime quite impressive
18:13:58 <AaronSw> Haven't you heard of the 388 day uptime legend?
18:14:00 <AaronSw> ;-)
18:14:35 <DanC_> * DanC_ waves
18:14:56 <AaronSw> * AaronSw waves
18:15:05 <em> * em waves wondering where DanC_'s been :)
18:15:41 <AaronSw> shooting some pool, of course
18:17:02 <em> AaronSw, is the cwm bug you mentioned something that DanC knows about?
18:17:15 <AaronSw> Yes, he's the one that explained it to me.
18:17:22 <danbri> http://dmoz.org/rdf/tags.html
18:17:23 <dc_rdfig> Q: http://dmoz.org/rdf/tags.html from danbri
18:17:46 <danbri> Q:|Dmoz RDF 'tags',  used in the current RDF files (2001-06-22)
18:17:47 <dc_rdfig> titled item Q
18:17:56 <danbri> Q:"new tags are highlighted"
18:17:56 <dc_rdfig> commented item Q
18:18:21 <danbri> Q: This message is being sent primarily to keep odp-rdf-announce alive as a
18:18:21 <danbri> yahoo group.  We have not changed the RDF file format, and currently
18:18:21 <danbri> have no plans to do so.
18:18:22 <dc_rdfig> commented item Q
18:18:35 <danbri> Q: ...yahoo group.  We have not changed the RDF file format, and currently have no plans to do so.
18:18:35 <dc_rdfig> commented item Q
18:19:07 <AaronSw> Q:[ChefMoz tags|http://chefmoz.org/rdf/elements/1.0/] *Tasty!*
18:19:08 <dc_rdfig> commented item Q
18:19:40 <danbri> Q: "There is a relatively new file, http://dmoz.org/rdf/tags.html, which provides a listing of each tag that we use in each of the RDF files. This is produced by reading the RDF files, and it highlights any tags that it doesn't expect.  By checking this file weekly, you can easily see any tags that have changed.
18:19:41 <dc_rdfig> commented item Q
18:23:39 <em> AaronSw, re error... DanC did inded fix this, and you can get fix by updating 2000/10/swap 
18:23:56 <em> * em version of cwm now works
18:23:58 <AaronSw> OK, that's probably why it worked when i did a cvs update ;-)
18:24:40 <danbri> Q: I'm puzzled how they're saying, "nothings changed" and "here's what changed" at same time.  See [odb-rdf-announce|http://groups.yahoo.com/group/odp-rdf-announce/] archives for [details|http://groups.yahoo.com/group/odp-rdf-announce/message/6].
18:24:40 <dc_rdfig> commented item Q
18:25:21 <AaronSw> Well, Yahoo implemented a new policy to shut down inactive groups.
18:25:30 <AaronSw> So he had to send some sort of message out.
18:25:47 <AaronSw> I suppose he figured it'd be better if he sent one out with something useful.
18:29:13 <AaronSw> Q:Well, Yahoo implemented a new policy to shut down inactive groups so he had to send some sort of message out. I suppose he figured it'd be better if he sent one out with something useful.
18:29:14 <dc_rdfig> commented item Q
18:29:36 <jonb> jonb has joined #rdfig
18:30:08 <jonb> * jonb waves, wonders if anyone knows the country code for telephoning france?
18:30:13 <danbri> aaron: I understand that, i'm just puzzled by the msg content. Is their format changing or not?
18:30:35 <AaronSw> * AaronSw looks in em's direction
18:31:41 <danbri> * danbri replys to bob's msg
18:32:12 <DanC_> +44, I think
18:32:58 <AaronSw> M:David McCusker [responds|http://wmf.editthispage.com/discuss/msgReader$4968
18:32:58 <AaronSw> ]
18:32:59 <dc_rdfig> commented item M
18:33:02 <AaronSw> M:David McCusker [responds|http://wmf.editthispage.com/discuss/msgReader$4968]
18:33:02 <dc_rdfig> commented item M
18:33:08 <AaronSw> what, danc?
18:33:16 <danbri> +44 is uk
18:33:20 <DanC_> +33, actually. http://thelist.internet.com/countrycode.html
18:33:25 <DanC_> (via google, of course)
18:33:35 <jonb> thx
18:34:05 <DanC_> hmm... cwm needs some sort of HTML output support.
18:34:28 <jonb> google is frequency flaking out for me :-((
18:34:48 <jonb> s/frequency/frequently/
18:35:33 <DanC_> * DanC_ wonders if danbri is around... about jsprolog
18:35:39 <em> DanC - here's my travel itin for next week - http://www.w3.org/2001/07/26-swcal/travel.n3
18:36:03 <DanC_> * DanC_ wonders what it would take to recode cwm in javascript, so it would run client-side.
18:36:04 <danbri> * danbri is around, though talking to jang next week f2f might be a better bet
18:36:09 <danbri> ooh
18:36:12 <em> DanC - a public dump of my calander for week 31 (next week) - http://www.w3.org/2001/07/26-swcal/=week-31.2001.n3
18:36:36 <em> trying again.. http://www.w3.org/2001/07/26-swcal/week-31.2001.n3
18:37:11 <danbri> Jan and I have talked a bit about how tabling prologs do things (ok, he's repeatedly tried to explain to me :), and discussed hacking something in that vein into js-prolog
18:37:39 <DanC_> * DanC_ adds links from http://www.w3.org/2001/08swws67/
18:38:46 <em> http://www.w3.org/2001/07/26-swcal/week-31.2001.rdf
18:38:46 <dc_rdfig> R: http://www.w3.org/2001/07/26-swcal/week-31.2001.rdf from em
18:38:51 <em> arrgg....
18:39:07 <danbri> http://ioctl.org/logic/jsprolog-history
18:39:07 <dc_rdfig> S: http://ioctl.org/logic/jsprolog-history from danbri
18:39:14 <jonb> perhaps better to write client side in java, micro version, for the phones?
18:39:15 <danbri> S:|History of Js-Prolog
18:39:16 <dc_rdfig> titled item S
18:39:19 <em> R:|em's public calander for week 31
18:39:19 <dc_rdfig> titled item R
18:40:02 <danbri> S:"When you see a link from this section, it'll point to a tabling imlpementation that might even use a WAM variant.
18:40:02 <danbri> Yeah, right. 
18:40:02 <dc_rdfig> commented item S
18:40:03 <AaronSw> * AaronSw tries to get his itinerary out of AA.com...
18:40:10 <danbri> S:"Yeah, right. "
18:40:11 <dc_rdfig> commented item S
18:41:37 <DanC_> counting weeks is an art.
18:42:15 <cdlu> [GlobalNotice] We would like everyone's opinion on this matter. Our new irc daemon is ready to go on line, and we are looking for the least possible intrustive time to make the change. Should we make the change as soon as possible, or wait until the end of the weekend? The change may cause up to 20 minutes of down time on the network. Please tell us your opinion in #openprojects, or mail your opinion to support@openprojects.net. T
18:42:19 <DanC_> my weekly-email-archive thinks we're in week 29... based on `date +%Y-%m`
18:43:15 <DanC_> odd... that's supposed to be YYYYMMWW=`date +%Y-%m`
18:43:25 <DanC_> wierd! what happens to percent-U?
18:43:31 <DanC_> anyway...
18:49:22 <DanC_> an observation re http://209.198.125.242/itdsw/pages/resource_edit.asp?resource=http%3A%2F%2Fwww%2Eintellidimension%2Ecom%2Fitdsw%2Fsites%2Fgsc%2FSWWSSched
18:49:32 <DanC_> rdfql is to rules ala cwm --think is to --filter
18:50:07 <DanC_> other way around...
18:50:18 <DanC_> rdfql is to rules ala cwm --filter is to --think
18:53:08 <GlobalMessage> The dancer-ircd server conversion will take place at 3AM EST (7AM GMT). The services database will be exported in another 3 hours. Please make adjustments to your channel lists and nickname passwords.
18:54:42 <danbri> hmm, talking of dates, shouldn't http://rdfig.xmlhack.com have rolled over to a new day by now?
18:58:18 <AaronSw> perhaps it's catching up for not logging the 26th?
18:58:55 <AaronSw> hmm, never mind
18:59:30 <AaronSw> There seem to be some edge cases that really confuse chump
19:00:05 <GlobalMessage> Clarification about the services database... Changes made after 6PM EST (10PM GMT) will not be reflected in the new services database.
19:01:20 <DanC_> * DanC_ adds a new item to http://www.w3.org/2001/08swws67/Makefile
19:01:29 <em> ok, gotcha
19:01:33 <bijan> bijan has joined #rdfig
19:02:19 <danbri> hi bijan!
19:02:24 <bijan> Hey dan!
19:02:47 <danbri> * danbri was thinking about fwding your SWI-Prolog msg to the XSB list, to encourage them to RDFize more...
19:02:57 <bijan> Good idea!
19:02:57 <danbri> any objections?
19:03:01 <bijan> Not at all.
19:03:11 <bijan> I've been interested in XSB but it seems a bit tricky to get into.
19:03:17 <DanC_> let's make an rdf-swws channel, Eric
19:03:39 <DanC_> folks that want to hack on SWWS calendar stuff, pls join #rdf-swws
19:03:41 <em> make sense... anyone that wants to look into swws demo  
19:03:45 <em> pls join
19:03:51 <bijan> Ooo, the jsprolog.
19:04:28 <bijan> I worked up a demo of the stuff from the first article in that.
19:04:35 <bijan> I don't think I got permission to use it.
19:04:37 <bijan> Or something.
19:04:41 <bijan> It's around somewhere.
19:05:04 <bijan> I.e., you could try out all the examples from the first article online.
19:05:13 <danbri> permission from jang? I'm sure he'd love that. We talked about making an executable xml.com article using it
19:05:24 <bijan> That's what I had.
19:05:44 <bijan> We dropped him a note but it may have gotten lost in the flurry.
19:06:00 <danbri> he was pretty busy for a while there; end in sight now I think.
19:06:10 <bijan> I'd happily rework it and add it to that article.
19:06:16 <bijan> It'd be a nice addition.
19:06:25 <bijan> So nudge him! :)
19:06:38 <danbri> Maybe enough there for a separate piece? but using egs from 1st article...
19:06:58 <bijan> There may be.
19:07:03 <danbri> i'll ask him tommorrow in stanford...
19:07:11 <danbri> can you dig out what you did?
19:07:20 <bijan> I'd have thought that the separate article focus would be on the implementation...unless you all are using for soem apps?
19:07:22 <bijan> Yes.
19:07:25 <bijan> it's on disk somewhere.
19:07:48 <leigh> leigh has joined #rdfig
19:07:52 <bijan> Now that I think of it, there are bunch of cool apps that come to mind.
19:08:02 <bijan> Maybe stuff with the calendering schema!
19:08:08 <leigh> evening
19:08:11 <bijan> All embedded in the page.
19:08:14 <bijan> Hello.
19:08:35 <danbri> I'd _love_ to see some prolog stuff reasoning about dates. DanC and I did a bit last year using XSLT-scraped stuff, things have moved on since then...
19:08:48 <danbri> hi leigh!
19:08:54 <bijan> Yes, I have big stack of event and situation calculus papers ;)
19:09:11 <leigh> hi dan, thought I'd stop by :)
19:09:14 <bijan> From back when libby was starting things up.
19:09:55 <bijan> Next article (whichis a go! the series continues!) is likely to be about "Expert system back websites"
19:11:59 <bijan> Other possible subjects: explicating various vocabs (e.g., DAML+OIL), SemantiChimp, calendering...and...
19:12:03 <bijan> Ah yes, CWM.
19:12:11 <leigh> cool
19:12:20 <danbri> xsb list: pinged!
19:12:24 <bijan> Great!
19:12:26 <bijan> Thanks.
19:12:41 <danbri> i mentioned http://rdfweb.org/~pldab/rdfweb/closure.P in passing too
19:12:47 <bijan> I've been getting great feedback.
19:12:52 <danbri> find paths through rdfweb dataset
19:12:58 <danbri> jan again, not me
19:13:03 <bijan> Ooo, neat!
19:13:09 <danbri> would it run in swi?
19:13:39 <bijan> Quick scan reveals nothing off...
19:13:39 <bijan> Ah!
19:13:41 <bijan> Oops.
19:13:46 <bijan> rdf(Property, X, Y) ; rdf(Property, Y, X)
19:13:54 <bijan> The rdf_db is SPO
19:14:01 <bijan> So a quick swap, not bad.
19:14:27 <danbri> A: see also [ping to XSB Dev list|http://lists.w3.org/Archives/Public/www-archive/2001Jul/0022.html]
19:14:27 <dc_rdfig> commented item A
19:14:34 <bijan> yes, otherwise it should run as is.
19:15:21 <bijan> * bijan fires up SWI.
19:15:46 <danbri> some data: http://rdfweb.org/~pldab/rdfweb/allfactoids.P
19:16:22 <danbri> ...a few months old. Dunno if libby's rdf indexer/robot leaves Prolog droppings, suspect it doesn't
19:17:14 <danbri> I'm really getting bored of my rdfweb database being hosed, going to have to take a day later this month and fix it all. Really this time.
19:17:41 <bijan> Heh.
19:17:44 <danbri> bijan, the web interface to the rules thing is currently http://rdfweb.org/rweb/logic?who=genid:jang&predicate=http://xmlns.com/foaf/0.1/knowsWell
19:18:03 <AaronSw> fix rdfweb: yay!
19:18:11 <danbri> ...a perl script that runs an instance of xsb when called. ick. A good candidate for replacing w/ your HTML generation stuff
19:19:48 <bijan> Yes, we're going to move over to DCGs, Ithnk, for a lot of monkeyfist stuff.
19:20:21 <bijan> It's just loads less tedious than XSLT for this kind of thing.
19:20:34 <bijan> For me.
19:23:09 <bijan> Hmm. A few trickinesses.
19:23:40 <danbri> what's tricky?
19:24:20 <bijan> Just aligning SWI's builtins with XSB.
19:24:42 <bijan> And he's got some singleton variables.
19:24:47 <bijan> Which SWI warns about.
19:24:49 <bijan> no biggie.
19:30:07 <bijan> Got part to run at least.
19:35:23 <danbri> oooh! cool
19:36:20 <bijan> Turns out that no one knows you and those who do won't admit it :)
19:36:39 <bijan> Can't argue with AI.
19:37:34 <danbri> tell me something I don't know :)
19:37:39 <leigh> heh
19:38:03 <danbri> leigh recently appeared in my palm-pilot database, thanks to libby's rdfweb.pdb file...
19:38:21 <leigh> cool
19:38:32 <leigh> still working on squish/inkling
19:38:44 <bijan> Aha! Some html clutter was confusing it.
19:39:29 <bijan> Still no friends, dan.
19:40:47 <jonb> * jonb waves to leigh
19:41:02 <leigh> hiya
19:41:02 <lasDesk> lasDesk has joined #rdfig
19:41:13 <jonb> wacha working on?
19:41:36 <leigh> junit test harness + code refactoring/cleanup of squish
19:41:45 <danbri> * danbri cheers
19:42:11 <leigh> means I get familiar with the code, whilst hopefully doing something useful
19:42:42 <danbri> leigh, saw some msgs about that earlier. The 'Tellable' class is indeed of solely historical interest. It gold folded into Janne's original SiRPAC API as whatever the class is called that eats triples
19:43:02 <leigh> my work log is at: http://www.bath.ac.uk/~ccslrd/inkling/log.txt
19:43:08 <leigh> OK thanks - its history :)
19:43:10 <danbri> RDFConsumer? whatever. And the MCF and SiLRI things are both, sadly, not worth following...
19:43:26 <bijan> Dan: is the same query at the top of closure.P supposed to work?
19:43:44 <bijan> * bijan realized he didn't have a known good query :)
19:43:58 <danbri> yes. The older org.desire.rudolf.* stuff was very organic. Think: student project (java applet viewer for MCF) folded with 'danbri learns rdf hacking' plus misc other things.
19:44:03 <jonb> RDFHandler?
19:44:21 <danbri> bijan: not sure, but the example in the url i gave above should work
19:44:30 <danbri> ie http://rdfweb.org/rweb/logic?who=genid:jang&predicate=http://xmlns.com/foaf/0.1/knowsWell
19:45:27 <bijan> Hmm. Is there prolog for that?
19:45:58 <danbri> just those two args...
19:46:29 <bijan> So same form as in closure.P?
19:46:41 <bijan> I.e., symmetricClosure(_S, 'genid:danbri', 'http://xmlns.com/foaf/0.1/knowsWell'), map(_S, 'http://xmlns.com/foaf/0.1/mbox', Answers)
19:46:44 <danbri> yes, the cgi script just invokes xsb with those args
19:47:07 <bijan> Ah, jsut the first.
19:47:48 <danbri> my $input = "symmetricClosure(_S, '$who', '$predicate' ), map(_S, '$foafmbox', Ans
19:47:48 <danbri> wers).\n";
19:47:58 <bijan> Aha.
19:48:01 <danbri> ...from the cgi. ugh. does that help?
19:48:56 <danbri> * danbri up to his elbows in SOAP. Annoyed at having had to *read the spec* to find out how method calls are serialised.
19:49:41 <bijan> Yes.
19:50:23 <bijan> I messed it up :)_
19:54:22 <bijan> Got it.
19:54:23 <bijan> Works fine.
19:55:37 <bijan> S = [ ('genid:gid', 2),  ('genid:libby', 2),  ('genid:michaelm', 2),  ('genid:philhall', 2),  ('genid:pldms', 2),  ('genid:sarahm', 2),  ('genid:craig', 1),  ('genid:emiller', 1),  (..., ...)|...]
19:55:38 <bijan> Answers = ['mailto:Ian.Sealy@bristol.ac.uk', 'mailto:craigd@netgates.co.uk', 'mailto:d.m.steer@lse.ac.uk', 'mailto:daniel.brickley@bristol.ac.uk', 'mailto:emiller@oclc.org', 'mailto:gid@litebase.com', 'mailto:jan.grant@bristol.ac.uk', 'mailto:libby.miller@
19:55:54 <danbri> Neat! Is a version feasible that'd work with both xsb and swi prologs? (and js-prolog, and cwm, and ... ;-)
19:56:06 <bijan> CWM, no :)
19:56:19 <bijan> Otherwise, certainly XSB and SWI.
19:56:34 <bijan> 'not' isn't predefined as an operator in SWI.
19:56:36 <danbri> I wonder how well cwm'd cope with the combinatorial explosion. I'm assuming it wouldn't...
19:56:49 <bijan> But it should be a normal predicate two.
19:56:59 <bijan> If not, you can add a line making that explicit.
19:57:13 <bijan> If you want ot use RDF/XML you'll need a slight mod.
19:57:20 <bijan> I.e., because SWI uses SPO order.
19:57:29 <bijan> While the factoids and closure use PSO.
19:57:32 <bijan> But that's trivial.
19:58:17 <danbri> yes, trivial (and annoying! Those SPO ba****ds will never win me over ;-)
19:58:25 <bijan> Heh.
19:58:31 <bijan> I sent a note to Jan about it :)
19:58:34 <danbri> * danbri finds it hard to care
19:58:45 <danbri> he's probly forgotten he wrote it
19:58:55 <danbri> it was a while ago...
19:58:56 <bijan> The other jan :)
19:59:01 <bijan> SWI-Jan.
19:59:04 <bijan> Not JS-Jan.\
20:01:01 <bijan> Does XSB parse RDF/XSB?
20:01:34 <bijan> BTW, a simple conversion of the factoids to SPO order means you get RDF/SML generation free.
20:01:40 <bijan> As I wrote in my article.
20:01:44 <bijan> Page 2 or 3 I believe.
20:01:52 <bijan> Which may make it worth it.
20:02:50 <danbri> It can, yes. They use the libwww RDF parser, a C transliteration of (and old version of) Janne's sirpac, by John Punin, and patched by Art Barstow (since its used in Amaya).
20:03:02 <bijan> Ick.
20:03:02 <bijan> :)
20:03:11 <bijan> Be nice to get some benchmarks on that.
20:03:33 <danbri> Hopefully Dave's Rapier will be able to replace it shortly. SiRPAC-based code isn't very 2001...
20:03:34 <bijan> * bijan  has a hate onfor amaya.
20:03:53 <sandro> When I want RDF in XSB I run cwm to convert it to n3, then my yaccn3 to turn it into prolog.
20:04:03 <bijan> Ouch.
20:04:05 <danbri> Problem with XSB's RDF last time i looked was documentation
20:04:16 <bijan> Get them to fund me! ;)
20:04:21 <sandro> Of course, I've never wanted to read RDF/XML, actually, so I've never really done that.
20:04:24 <jonb> i've been noodling on something regarding s-expressions and XML: http://www.openhealth.org/XSet/SExpressions.html
20:04:25 <leigh> hmmm, so do you think it'd be best to limit the sirpac dependencies in inkling?
20:04:35 <danbri> they shipped a new version recently. Most of my XSB hackings have been w.r.t. the SQL interfaces, which imho hold huge promise.
20:04:39 <sandro> But I do use n3 that way (no need to run cwm).
20:04:56 <bijan> sandro: is your dcg n3 parser still around?
20:05:01 <bijan> I thought I might beat onthat.
20:05:12 <danbri> leigh: yes. it's been a real pain (worsened by having two flavours of sirpac). Jeremy's new java parser looks a good bet.
20:05:26 <sandro> Hrm.  I never got it working.  You want me to dig up the bits?
20:05:31 <bijan> Sure.
20:05:34 <leigh> OK. Its on my list.
20:05:38 <bijan> It's on my list to write one.
20:05:46 <danbri> ...the only strong dependency between Inkling and sirpac is w.r.t. the API; Inkling shouldn'
20:05:53 <bijan> And it'll prollysave me from figuring out hte N3 grammar ;)
20:05:54 <danbri> t care which parser sits behind the api...
20:06:17 <barstow> * barstow agrees with danbri
20:06:25 <bijan> Added bonus...I can get revenge for you dissing my article by mocking your prolog style! ;)
20:07:10 <danbri> art: w.r.t. libwww and rapier, or w.r.t. sirpac and Jeremy's parser? (or both :)
20:07:36 <AaronSw> F:"""The W4C (World Wide Wireless Web Consortium) has just published the specification of SML/NG (Simple Markup Language -- New Generation), a simplified version of XXHTML designed for the new generation of hypertext rendering micro-devices, running on hardware with reduced computational capacity such as wristtop computers, thumbnail-worn PDAs, and internet-enabled ice boxes."""
20:07:36 <dc_rdfig> commented item F
20:07:54 <AaronSw> Watch out W3C!
20:08:11 <leigh> There's an I3C as well now.
20:08:20 <leigh> I think there's a franchise or something... :)
20:08:24 <bijan> Anyone doing the cotnetst.
20:08:29 <barstow> danbri: that Inkling shouldn't depend on the parser but on an RDF API (the Standford RDF API or Jena)
20:08:30 <bijan> I did last year, seemed much harder.
20:08:36 <bijan> But I'm in the midst of moving.
20:08:46 <sandro> Bijan, it with my other XSB stuff, in http://www.w3.org/2001/04/pl/.  I see http://www.w3.org/2001/04/pl/n3_syntax.P and I'm not sure what state it's in.   My prolog/RDF conventions were of course different from those in your article.
20:09:17 <leigh> Inkling needs to be brought up to date with the latest Stanford stuff...
20:09:33 <bijan> Thanks sandro.
20:09:37 <bijan> I'll see what's stealable.
20:09:43 <danbri> and/or the latest W3C stuff? and/or the latest Jena...? :)
20:09:47 <bijan> er..if that's kosher, of course!
20:09:58 <danbri> what I like is that it doesn't mind which rdf api it sits on top of...
20:10:08 <leigh> yes, I've got a long list of stuff to do! Code archeology is fun...
20:10:42 <danbri> what we have to get right though is a way for Inkling to pass through RDF queries intact when it detects a database is capable of handling SquishQL natively, ie. rather than hammering the 'dumb' triple-api interface
20:10:53 <sandro> I'm working on much cooler parser stuff, but it's not ready yet.   (code's public, but I don't encourage anyone to look at it.)   http://www.w3.org/2001/06/blindfold/
20:11:23 <leigh> dan: yep, already considered that. Should be able to abstract over whats there in a much cleaner way.
20:12:12 <AaronSw> * AaronSw now has full flight information at http://logicerror.com/SWWS-plans
20:12:41 <danbri> leigh, do you think we can get away with overloading the JDBC interfaces for rdf query, or is that too cutesy a trick?
20:13:16 <bijan> It's fine so long as you don't add a bow.
20:13:25 <leigh> hmmm, not sure
20:13:42 <danbri> I'd like a lightweight interface that an object implementing any of the fashionable rdf graph APIs could also implement, that allows it to show Inkling it can eat SquishQL queries. Not sure how much such glue needed between this and JDBC...
20:14:04 <danbri> eg the wrapper classes for XSB that stefan wrote...
20:14:17 <danbri> (I think stefan did it...)
20:14:30 <leigh> Well I was thinking of doing away with 'QueryDirect' and just having a Queryable interface. Fallback would be to do it the hard way.
20:14:45 <leigh> But leave JDBC and QE stuff largely intact
20:15:12 <leigh> Haven't got to that level yet though. Just refactoring RDFGraph at the minute
20:15:14 <danbri> I've not looked at the interfaces since was last in bristol, but that sounds plausible
20:15:37 <danbri> that old thing! that was my first rdfdb implementation...
20:15:50 <leigh> Planning to pull out the query syntax parsing as well to allow that to alter as needed
20:15:57 <leigh> Is it not worth messing with then?
20:17:05 <danbri> i'm not sure how its changed in last couple of years. It used to be a pretty plain in-memory RDF store, used two hashtables. Probably created too many String and URI/URL objects but hey. Was optimised for some but not all RDF API calls.
20:18:35 <danbri> Ok there's cvs logs going back to feb '99 in http://cvs.ilrt.org/cvsweb/rudolf-project/org/desire/rudolf/rdf/RDFGraph.java if you didn't see those yet
20:18:51 <leigh> Its a useful starting point for me. An in-memory is good for the minute as I'm not using Postgres and plan to to add Oracle support.
20:20:14 <danbri> I recoded basically the same thing a couple months ago in perl, fwiw. http://cvs.ilrt.org/cvsweb/rudolf-perl/RDF/RDFWeb/MemDB.pm 
20:21:25 <leigh> cool, thx for the pointer. We should do a semweb-sw event soon
20:21:56 <danbri> yes! ideally when i'm in the country... but maybe several...
20:22:09 <danbri> * danbri wonders if there are more rdf folks in bristol or in boston...
20:22:25 <danbri> (counting bath as bristol for argument; and cambridge/somerville as boston :)
20:23:03 <leigh> :)
20:24:18 <danbri> i quite like that ugly perl package, despite lack of polish. Shows rdf handling can be pretty simple....
20:24:44 <leigh> * leigh doesn't consider perl to be simple :)
20:24:52 <danbri> dc_rdfig:view
20:24:53 <dc_rdfig> O: RDFCore 2001-07-27 raw IRC log (http://ilrt.org/discovery/chatlogs/rdfcore/2001-07-27.txt)
20:24:53 <dc_rdfig> P: http://swordfish.rdfweb.org:8085/rdfquery/index.html
20:24:54 <dc_rdfig> Q: Dmoz RDF 'tags',  used in the current RDF files (2001-06-22) (http://dmoz.org/rdf/tags.html)
20:24:55 <dc_rdfig> R: em's public calander for week 31 (http://www.w3.org/2001/07/26-swcal/week-31.2001.rdf)
20:24:56 <dc_rdfig> S: History of Js-Prolog (http://ioctl.org/logic/jsprolog-history)
20:25:27 <danbri> Q:see also [Bob's clarification|http://lists.w3.org/Archives/Public/www-archive/2001Jul/0024.html].
20:25:27 <dc_rdfig> commented item Q
20:27:33 <sbp> sbp has joined #rdfig
20:32:26 <sbp> And here
20:37:36 <rrs> rrs has quit
20:42:07 <sbp> sbp has quit
20:42:15 <sbp> sbp has joined #rdfig
20:48:11 <danbri> wohoo! I just got data back from my Perl SOAP server using the Apache/Axis java library as a client
20:48:48 <leigh> nice.
20:50:52 <danbri> ...and it shows me the 'map' type Glen was telling me about. I mean, I can see the raw response, which i'd not yet found how to do with Perl SOAP::Lite nor attempted by snooping traffic. 
20:51:12 <danbri> hang on...
20:53:05 <danbri> see raw rqst/responce data, w/ ugly line breaks, c/o http://fireball.dyndns.org/~pldab/2001/07/swad/rdfqlog.txt
20:53:45 <tdxdave> tdxdave has quit
20:54:25 <danbri> I should explain the data structure being serialised. It's a Squish result set, an array of hashes, where the keys are column numbers (@@should be variable names) and the values are the values of that (ie the key-th) variable in that row of the resultset.
20:56:54 <sbp> sbp has quit
20:57:01 <danbri> Next job, how to get Axis to rehydrate that serialisation into a Java array (or List?) of Hashtables. Or into whatever Inkling's resultset format evolves into.
20:57:10 <danbri> Am I making sense here?(tm)
20:58:02 <AaronSw> * AaronSw has never heard of rehydrating a serialization, before... is that when you spill milk on a comic book?"
20:58:26 <AaronSw> ;-) no, i understand
20:58:37 <leigh> yep
20:58:48 <danbri> * danbri likes to abuse words
20:58:53 <danbri> phew!
20:59:37 <leigh> doing away with hashtables and switching to Java 1.2 Collections is another TODO
21:00:57 <AaronSw> danbri should learn phenomics. ;-)
21:01:13 <danbri> But whose TODO? SOAP doesn't even properly have hashtables, not sure what Axis currently turns 'maps' into. Or do you mean (the various uses of)Hashtables in Inkling
21:01:15 <danbri> ?
21:01:38 <leigh> Inkling. I can only refactor the world one step at a time :)
21:05:03 <leigh> dan, I'm having trouble understanding something in RDFGraph
21:05:23 <danbri> sure. what's up?
21:05:26 <leigh> oh wait a sec...
21:05:32 <leigh> * leigh scratches head
21:05:58 <leigh> hmm. you've got forward and inverse pointers in hashtables
21:06:25 <leigh> under what circumstances would an inverse pointer be created?
21:06:40 <sbp> sbp has joined #rdfig
21:07:04 <danbri> they're managed in parallel. Look at the perl code in MemDB above for analogy (if you read Perl).
21:07:11 <danbri> example:
21:07:20 <danbri> asserting: leigh -livesIn-> bath
21:07:46 <danbri> I would take the pair (leigh, livesIn) and add 'bath' to the value of that pair in forward-ptr.
21:08:05 <danbri> i would take the pair (livesIn, bath) and add 'leigh' to the value of that pair in inverse-ptr
21:08:22 <danbri> which sets us up to answer lots of common queries (but not all)
21:08:27 <leigh> hmm, well that doesn't seem to be happening
21:08:44 <danbri> what's it doing?
21:09:02 <leigh> I would expect numOfForwardPointers = numOfInversePointers correct?
21:09:40 <leigh> I've added 7 assertions to a graph, but have 0 inverse. Unless I'm making a cockup with the test data somewhere
21:09:41 <danbri> good test. aside from some potential pedantry w.r.t. literals, yes.
21:10:12 <danbri> can you point me at a url for latest .java of it, to jog my memory. its been a while!
21:10:29 <leigh> er, just a sec.
21:10:35 <danbri> is http://cvs.ilrt.org/cvsweb/rudolf-project/org/desire/rudolf/rdf/RDFGraph.java up to date? (I think this stuff might've been oved outof that cvs space)
21:11:29 <leigh> dunno. I'm working in my own CVS tree.
21:11:51 <danbri> wish libby wasn't on aeroplane...
21:12:03 <leigh> try this: http://www.bath.ac.uk/~ccslrd/inkling/RDFGraph.java (current snapshot of my src)
21:13:22 <danbri> ok, reading.
21:13:35 <danbri> addRules() -- I liked that trick... But anyway...
21:13:59 <leigh> I've dumped by test case (RDFGraphTest, TestConstants) there also
21:14:01 <danbri> did you add the interning or was that libby?
21:14:06 <leigh> libby
21:14:26 <danbri> good, i never got to that. should have speeded things up.
21:14:42 <leigh> All I've done so far is switch to using my new RDFConstants class, and separated out the assertLiteral and assertResource methods, ready for refactoring
21:15:11 <leigh> want to get the test cases right before I go further. Suspect my RDF-newbie-ness is beginning to show...
21:16:24 <danbri> where are assertLiteral/Resource now?
21:16:46 <danbri> oh, gotcha
21:17:26 <leigh> output of running test case is there also (out.txt)
21:19:25 <leigh> from what i can see, entries are only made in ip hashtable within assertResource
21:19:33 <danbri> do you know how "values = ( Hashtable )ip.get( makeKey( prop, val ) ); (in assertResource) is behaving
21:20:03 <danbri> it seems reasonable to assume fp and ip should have the same (or at least similar; literals maybe an issue) number of entries
21:23:05 <leigh> well assertResource isn't being called, because I'm only using literals, hence no entires
21:24:25 <danbri> If you test it with URIs, do you get what we expect?
21:24:26 <danbri> danbri has quit
21:26:09 <danbri> danbri has joined #rdfig
21:26:22 <danbri> randomly disconnected again >:-|
21:26:26 <danbri> leigh,
21:26:53 <leigh> * leigh slaps forward
21:27:01 <danbri> in assertLiteral,  TODO: IP is not implemented for literal valued statements, Bug?
21:27:11 <leigh> just noticed it. Doh!
21:27:26 <sbp> s/Doh!/D'oh!
21:27:53 <danbri> ...not sure if 'bug', depends whether the API calls brute-force it by chugging through fp, versus whether there are false negatives cos queries are failing.
21:27:53 <leigh> sbp: which version went into the OED then?
21:28:16 <sbp> Well, both. I think they note that "D'oh!" is the official spelling, but still list it as "Doh!"
21:28:26 <sbp> But the producers have always stated that it's *with* apostrophe
21:28:45 <danbri> this is so cool of you to go through the code btw leigh! wish we'd had #rdfig etc 3 years ago :-)
21:28:58 <bijan> Praise edd!
21:29:09 <leigh> dan, I've switched to using resource, and now have some ip entries, but not as many as I'd expected
21:30:13 <leigh> I've been meaning to do more open source hacking and some RDF, and this is a happy cross-over. I get a perverse pleasure out of it! :)
21:30:55 <danbri> the number of fwdptr keys is the number of unique resource/predicate pairs; the no of invptr keys is the number of unique predicate/value pairs. I forget why we thought they'd be the same now. What (arguably) should be the same is the number of operations on fp and ip.
21:31:24 <danbri> does that answer it?
21:31:36 <bijan> Code spelunking and archeology...great stuff!
21:32:24 <danbri> BTW historical note, it was Guha (explaining abortive berkeleydb based rdfdb system in olde Mozilla) who told me how to store RDF in this style.
21:32:41 <leigh> I'm not sure that its clicked yet...
21:33:45 <danbri> clicked? the indexing?
21:34:35 <leigh> I've got it now.
21:34:44 <leigh> Yep makes sense.
21:34:55 <leigh> So why are literals an issue?
21:35:06 <danbri> imagine you're  an app, and you wonder who lives in bath. that's a triple w/ bits missing, ie. [??] --livesIn->'bath'. So some API call (couched in terms of triple-matching, or graph nav, picking a metaphor...) grabs that and looks in fwdptr or invptr table.
21:35:21 <danbri> think of the possible questions you can ask such an api. 
21:35:26 <danbri> xxx - is this so
21:35:31 <danbri> ooo - dump the database
21:35:38 <danbri> xoo - arcs out of a node
21:35:44 <danbri> oox - arcs into a node
21:35:53 <danbri> xxo - values for some node/arc pair
21:36:03 <danbri> oox - values of some arc into a node
21:36:08 <danbri> er is that it? 
21:36:18 <leigh> * leigh nods
21:36:51 <danbri> the fwdptr/invptr trick optimises for some but not all of these. how useful it is (apple pie) depends on shape of data and likely queries. 
21:37:58 <danbri> why literals an issue? they're not really, it's just easy to futz and confuse 'behind the scenes' optimisations (eg. storing 'john smith' once) with the abstract information model.
21:38:31 <danbri> eg. an API that allows updates on the graph, if john smith has a sexchange operation and you want to s/john smith/joan smith/, you don't want to turn all johns into joans.
21:38:43 <danbri> so needs a little extra care.
21:38:47 <danbri>  hence the TODOs 
21:38:53 <leigh> I see.
21:46:23 <sandro> sandro has quit
21:54:08 <danbri> leigh, quick java question. Which package does List live in?
21:54:26 <danbri> I'm trying to cast the soap results back into java objects
21:54:37 <leigh> java.util.List
21:54:44 <leigh> All the collections are in there
21:54:47 <danbri> thanks!
21:55:46 <leigh> take a look at Arrays and Collections - bound to be some utility methods you can use in there
21:55:55 <leigh> e.g Object[] -> List
21:58:26 <danbri> i'm trying to find out Axis' soap-to-java mapping; reading the manual's proving more productive than guessing (my original instinct!)
21:58:52 <danbri> just stumbled across http://xml.apache.org/axis/docs/SOAPVerse.html -- idea for MOO/MUD-like SOAP test playground
21:59:13 <danbri> Spooky as I'm often found fiddling with RDF/MOO ideas
21:59:46 <leigh> eek, looks like an I idea I had only last week
22:00:42 <leigh> peer-to-peer MOO
22:01:17 <danbri> I've got skeletal outlines of an implimentation...
22:02:31 <danbri> eg http://ilrt.org/discovery/2001/03/rdf-moo/ http://ilrt.org/discovery/2001/06/rdfperl/current/samples/moo.rdf or  http://ilrt.org/discovery/2001/06/rdfperl/current/samples/verbmoo.rdf
22:02:47 <danbri> but I mustn't get distracted!
22:02:50 <danbri> soap.
22:04:52 <danbri> if we asked Rael to name a peer-to-peer MOO... MOOKat?
22:05:26 <leigh> :)
22:06:53 <barstow> barstow has left channel
22:08:32 <sbp> sbp has quit
22:10:07 <sbp> sbp has joined #rdfig
22:22:18 <leigh> dan, when does libby get to the states?
22:28:39 <danbri> about now i think
22:28:50 <danbri> i should go switch my mobile phone on, thinking about it
22:29:17 <leigh> could do with a chat with her at some point soon.
22:30:35 <leigh> actually I'm still slightly confused, but I'm ascribing that to the late hour. e.g. valuesWhere() current results I don't understand, but if I make it use fp rather than ip then it fits. Think I must be thinking upside down! :)
22:30:51 <leigh> need to come at it fresh, methinks.
22:31:36 <danbri> I picked those method names, and even I find them confusing. what are they, nodesWhere and valuesWhere?
22:32:02 <leigh> there are triplesWhere, nodesWhere and valuesWhere
22:32:19 <danbri> valuesWhere is valuesOfSomeNamedArcOfSomeNamedResource()...
22:32:45 <danbri> shouldn't use ip, should use fp I think
22:32:55 <leigh> thats what I thought!
22:33:18 <danbri> Bugs are perfectly feasible, though I'd be suprised how this could be working at all with a bug in these innards...
22:33:31 <danbri> triplesWhere() libby added
22:33:39 <leigh> thats why I'm assuming I have a bug in my understanding not the code
22:34:17 <leigh> for example: I've asserted bized.ac.uk has a dc:subject of economics. ditto for economist.com
22:34:24 <danbri> yes
22:34:47 <leigh> valuesWhere gives me zero results for valuesWhere(dc:subject)
22:34:50 <leigh> ?
22:35:04 <leigh> I expected 2
22:35:36 <danbri> * danbri revisits code
22:36:07 <danbri> with  /** Query method: find values 'of this property for this resource' */
22:36:07 <danbri>     public Vector valuesWhere( Property property, Resource resource )
22:36:31 <danbri> ...I often used to get the args in the wrong order. Doubt that's a prob here as you said switching fp/ip worked.
22:36:46 <leigh> I'm testing valuesWhere(Property property) also
22:37:27 <danbri> oh, sorry, didn't realise there was a 1 arg version too
22:37:28 <leigh> I also have 3 assertions of dc:creator but get a single result from valuesWhere(dc:creator). Odd.
22:38:54 <leigh> hmm, now that I look, the 2 args version uses fp, but the 1 arg ip!
22:38:55 <danbri> the reason that one uses ip instead of fp, is that the ip keys begind with the rdf predicate URI, so it can churn through looking for keys that startwith the chosen property
22:40:24 <danbri> either way it is somewhat brute-force; as i said earlier in my noughts-and-crosses diagram, some query cases are optimised for. this one (which i missed above i think, or maybe its a hybrid ) isn't. 
22:41:20 <danbri> try making economics a URI?
22:41:30 <danbri> you mentioned above that literals weren't going into ip.
22:41:38 <danbri> * danbri thinks you've found a bug
22:44:07 <leigh> If i make economics a Resource then I valuesWhere(dc:subject) returns 1
22:44:17 <leigh> result
22:44:19 <danbri> figures!
22:44:49 <danbri> so, re discussion earlier about unimplemented ip for literals, seems not all the query methods were working around that
22:45:14 <leigh> errr, why not 2?
22:45:35 <leigh> pretend I'm really thick.. :)
22:47:41 <danbri> * danbri didn't think this through... hang on... (brain creaks into gear)
22:48:45 <danbri> 1 result: I guess the semantics of that method call are unclear (at least to me). 
22:49:15 <danbri> there is only one value in the graph that is at the end of a dc:subject arc, ie. foo:economics. So we get one value.
22:49:42 <danbri> If the call was: return all triples whose predicats are dc:subject, we'd have expected two.
22:49:56 <danbri> dunno. It's an odd query method. I deny all knowledge! 
22:50:06 <leigh> thats what I thought the method was doing
22:50:06 <danbri> * danbri blames libby in her absense :)
22:50:11 <leigh> * leigh lol
22:50:13 <danbri> thats=which? the latter?
22:50:33 <leigh> thats= return all triples whose predicats are dc:subject, we'd have expected two.
22:50:52 <leigh> which is how I wrote my test cases, and hence my brainache :)
22:52:30 <leigh> and it follows from that, that results of valuesWhere(predicate, resource) isSubSetOf valuesWhere(predicate)
22:53:10 <danbri> yes
22:54:11 <leigh> think I'll have to chat to libby. Either its a bug, or the semantics of the method are mislabelled.
22:56:10 <leigh> thanks for the help. sorry to distract u from SOAP hackings
22:56:11 <danbri> certainly confusing
22:56:21 <danbri> NotAProblem :)
22:56:29 <leigh> I'm off now. Speak to you soon
22:56:29 <danbri> this is fun actually
22:56:37 <danbri> OK, take care
22:56:56 <leigh> re: fun - don't worry in the words of Arnie: "I'll be back" :)
22:57:01 <leigh> cya!
22:57:09 <leigh> leigh has quit
23:02:56 <sbp> sbp has quit
23:04:27 <danbri> J: see also [Apache Axis mail archives|http://marc.theaimsgroup.com/?l=axis-dev&r=1&w=2] 
23:04:28 <dc_rdfig> commented item J
23:07:57 <sbp> sbp has joined #rdfig
23:14:39 <bijan> bijan has quit
23:21:03 <danbri> http://marc.theaimsgroup.com/?l=axis-dev&m=99617167021772&w=2
23:21:04 <dc_rdfig> T: http://marc.theaimsgroup.com/?l=axis-dev&m=99617167021772&w=2 from danbri
23:21:36 <danbri> T:Apache/Axis-dev discussion on SOAP Serialization rules 
23:21:36 <dc_rdfig> commented item T
23:22:04 <danbri> T:|SOAP serialization in Axis (developer mailing list)
23:22:05 <dc_rdfig> titled item T
23:23:20 <danbri> T:I'm trying to understand how SOAP serialization relates to RDF syntax in practice; having an opensource implemention to nose around is proving helpful.
23:23:20 <dc_rdfig> commented item T
23:26:56 <robster> [GlobalNotice] Bye bye Espernet services, its been nice knowing you ;)
23:36:01 <danbri> N:related [XMLP Working Group Transport Binding Task Force
23:36:01 <dc_rdfig> commented item N
23:36:37 <danbri> N:related [XMLP Working Group Transport Binding Task Force 2001-07-26 meeting|http://www.w3.org/2000/xp/Group/1/07/26-tbtf-minutes] for more b/g.
23:36:37 <dc_rdfig> commented item N
23:56:00 <danbri> T: oh, xmlhack article mentions a [useful survey of xml data binding resources|http://www.rpbourret.com/xml/XMLDataBinding.htm]. StuffToReadLater.
23:56:01 <dc_rdfig> commented item T
23:56:17 <danbri> * danbri back later