Semantic Web Interest Group IRC Chat Logs for 2002-03-06

00:04:17 <AaronSw> anyone know how to turn a next cube on?

00:05:44 <AaronSw> oh, it's on the keyboard

05:35:27 <dilbert> [Global Notice] Hi all. We're experiencing some connectivity problems, please bear with us.

05:37:33 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

05:40:48 <lilo> [ServerNotice] Hi all. The server you are on is asimov, a new server under test. Unfortunately, it's split, indicating that its connectivity may not prove to be entirely adequate. It's been removed from the rotation for now, and you may want to disconnect and reconnect to 'irc.openprojects.net' if you see any further problems. Sorry for the inconvenience, and thank you for your patience.

05:45:48 <connolly> ok... dbview seems to be working.

05:45:53 <connolly> connolly is now known as DanC

05:46:03 <DanC> now... to do show tables etc.

05:46:12 <DanC> cf 4.5.6.1 Retrieving information about Database, Tables, Columns, and Indexes

05:46:13 <DanC> http://www.mysql.com/documentation/mysql/bychapter/manual_MySQL_Database_Administration.html#SHOW_DATABASE_INFO

05:57:46 <danbri> this is an SQL 2 RDF dumper?

06:04:33 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

06:26:07 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

06:29:07 * danbri gets soap2rdf script working, dumps out the Ruby Application Archive metadata for testing

06:29:27 <danbri> (my script being an app/vocab specific hack, currently)

06:38:03 <DanC> ah... hi

06:38:11 <DanC> yes, SQL 2 RDF dumper.

06:38:39 <DanC> wanna try it?

06:38:43 <DanC> danbri?

07:59:43 <mnot> hmmm.... /me finally caught up with what 'smush'ing is...

08:00:00 <mnot> it would be nice if there were a Semantic Web jargon file (in n3, natch)

08:00:23 <mnot> with references, etc.

08:00:39 <mnot> so people would no longer wonder about the meaning of bNode, etc. ;)

13:43:26 <danbri> dajobe? should redland (and perl/python wrappers) build from CVS currently?

13:44:56 <dajobe> possibly

13:45:16 <dajobe> the latest snapshots seem to be OK http://www.redland.opensource.ac.uk/dist/snapshots/redhat/

13:46:32 <danbri> I tried to make a SWIG wrapper, but when I ran it my test script barfed with

13:46:34 <danbri> ./danbri.rb:4:in `require': ./Redland.so: undefined symbol: RDF_ParserCreate - ./Redland.so (LoadError)

13:46:53 <danbri> So I tried to the Perl and Python wrappers and got errors there. I may have screwed up though...

13:46:56 <dajobe> dunno where RDF_ParserCreate is frome

13:47:02 <dajobe> repat probably

13:47:22 <danbri> rdf_parser_repat.c: scontext->repat = RDF_ParserCreate(NULL);

13:47:49 <dajobe> so somehow, repat isn't compiled/linked into to the above Redland.so

17:08:38 * bwm curses loudly

17:08:59 <bwm> Amaya crashes repeatedly editing RDFCore f2f minutes

17:09:29 <bwm> Can anyone suggest an alternative xtml compliant wysisyg xhtml editor

17:10:24 * JibberJim laughs at the very notion...

17:10:48 <JibberJim> something else and tidy is probably your only hope.

17:11:48 <bwm> Arghhhh - I've got two hours a day to work on this stuff and I have to spend my time patching up tools.

17:12:02 <bwm> ok - emacs it is

17:44:11 <DanC>http://www.w3.org/2000/10/swap/dbork/dbview.py

17:44:11 <dc_rdfig> A: http://www.w3.org/2000/10/swap/dbork/dbview.py from DanC

17:44:29 <DanC> A:|dbview -- view an SQL DB thru RDF glasses.

17:44:29 <dc_rdfig> titled item A

17:45:24 <DanC> A:implements much of TimBL's design note, [Relational Databases on the Semantic Web|http://www.w3.org/DesignIssues/RDB-RDF.html]

17:45:24 <dc_rdfig> added comment A1

17:45:45 <DanC> A:see [alpha test/peer review chatlog|http://lists.w3.org/Archives/Public/www-archive/2002Mar/0010.html]

17:45:45 <dc_rdfig> added comment A2

17:46:51 <mnot> MarkB, you around?

17:47:13 <MarkB> yup

17:50:33 <mnot> I was going to say these would be appropriate for restRDF, but maybe they're more appropriate for WRDL: http://rdfweb.org/2001/01/swipe/ and the sherlock services (referenced within)

17:52:10 <MarkB> ooh, cool. this is *exactly* what I've been looking for, for some work I'm doing at PF. Not sure how it would relate to RestRDF - not meta enough 8-)

17:52:34 <mnot> yeah, restRDF is way meta - just the way I like it! ;)

17:52:45 <MarkB> party on!

17:52:49 <mnot> heh

17:52:52 <DanC> what's cool?

17:53:00 <MarkB> Swipe

17:53:02 <mnot> Swipe

17:53:23 <DanC> er... but not cool enough to tell dc_rdfig about?

17:53:28 <mnot> I was looking at it in the context of xpath2rss and friends, but then realised it was interesting from a REST / WRDL standpoint as well

17:53:58 <mnot> well, it's on rdfweb

17:54:16 <mnot> and we're mostly interested in it from a rest standpoint, i think; isn't it old news in the RDF world?

17:54:26 <dajobe>http://rdfweb.org/2001/01/swipe/

17:54:27 <dc_rdfig> B: http://rdfweb.org/2001/01/swipe/ from dajobe

17:54:33 <dajobe> B:|SWIPE

17:54:33 <dc_rdfig> titled item B

17:55:10 <DanC> B:hm... I could use an example earlier in the spec

17:55:11 <dc_rdfig> added comment B1

17:55:29 <dajobe> shall I add? B:<MarkB> ooh, cool. this is *exactly* what I've been looking for

17:55:35 <mnot> heh

17:55:50 <DanC> B:<BasicSpec rdf:about="" method="GET"> <-- method attr is unqualified

17:55:51 <dc_rdfig> added comment B2

17:55:55 <MarkB> now I see an example, I'm not quite as impressed.

17:56:30 <MarkB> heh, no thanks 8-O

17:56:56 <mnot> I like it as a use case

17:57:19 <DanC> btw... dbview is pretty REST-y

17:57:37 <DanC> maps sql select to HTTP GET, pretty much.

17:57:55 <DanC> haven't done the hard part, i.e. update/insert => POST, yet.

17:58:07 <MarkB> update->PUT, no?

17:58:18 <DanC> umm...

17:58:30 <DanC> not in general.

17:58:32 <MarkB> calling it db*view* gives you an out, DanC 8-)

17:58:33 <mnot> wouldn't think so

17:58:55 <DanC> PUT only works for update TBL set COLs where PRIKEY=val

17:59:18 <DanC> POST handles the general update TBL set COLs=EXPRS where whatever-you-like

17:59:56 <MarkB> if you had a URI to a row, PUT would replace the row

18:00:12 <DanC> yes, a URI to a row is the case where you have a primary key (and its value for the row).

18:01:43 <DanC> the fun bits would be to support detailed mappings like mysql datestamps to http:last-modified. Or in general, generation numbers to HTTP etags.

18:02:03 <MarkB> what did you have in mind for an update that would use POST?

18:02:09 <mnot> that *would* be nice. I started to do that with LDAP...

18:02:11 <MarkB> neat idea

18:02:50 <DanC> for POST update, just marshall and sql update statement any way you like. POST it to the database.

18:02:59 <DanC> s/and sql/an sql/

18:04:07 <MarkB> sounds like tunneling, if the word "update" is in the body of the POST

18:04:18 <MarkB> (if that's what you meant)

18:04:48 <DanC> well, the POST handler on the database has to somehow distinguish update from insert.

18:05:01 <MarkB> I *think* PUT would suffice for that distinction

18:05:36 <MarkB> probably need to run through scenarios - haven't thought about mapping SQL ops to HTTP methods much before

18:05:46 <DanC> no... consider: update employee set salary = salary * 1.03 where performanceEval>4

18:06:00 <DanC> that's clearly not a PUT

18:06:50 <MarkB> good example

18:07:20 <MarkB> but set salary = 40000 where level=4 would be PUT

18:08:36 <DanC> I don't think so.... you need where employeNum=43 to do PUT.

18:08:59 <DanC> i.e. you have to identify a single row. or ... hmm... maybe not... maybe you could address a column and PUT to it.

18:09:24 <DanC> ok, now my brain is hurting.

18:09:26 <MarkB> that's what I was thinking. might be awkward though

18:09:55 <DanC> logger, pointer?

18:09:55 <DanC> See http://ilrt.org/discovery/chatlogs/rdfig/2002-03-06#T18-09-55

18:10:16 <DanC> A:see also: [discussion of mapping update/insert to PUT/POST|http://ilrt.org/discovery/chatlogs/rdfig/2002-03-06#T18-09-55]

18:10:16 <dc_rdfig> added comment A3

18:10:28 * DanC wanders off...

18:11:09 * MarkB heads out to work on his delts

18:38:16 <MysticOne> [GlobalNotice] Hey everyone, please bear with us while we get things sorted out. Apparently, some of our hubs have lost connectivity in numerous geographic locations. Not sure what the deal is, but we're working on it. We'll give more information as it becomes available. We appreciate your understanding, and as always thank you for using Open Projects.

18:49:13 * danbri catches up

18:50:29 <MysticOne> [GlobalNotice] Okay, most of the servers and all are back up now. A few hubs have lost connectivity, as well as a few normal servers, and unfortunately services. It looks like the possibility of a few backbone providers having a few lines cut, though I'm not completely sure on that.

18:51:12 <MysticOne> [GlobalNotice] Anyway, we're still working on it, though the stability will have hopefully returned for now. Again, we really appreciate your understanding in this matter, as well as your patronage for the Open Projects Network.

18:53:36 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

18:58:12 <MysticOne> [GlobalNotice] Okay, last one unless anything else pops up. Services are back and looks like we're in business. Though we couldn't have forseen it, apologies for the interruption in services. Every have a great day! :)

19:21:48 <tim-gone> DanC, new http://www.w3.org/DesignIssues/RDB-RDF table -- yellow rows are documents, white are not

19:22:04 <tim-gone> s/RDB-RDF/RDB-RDF.html/

19:35:03 * DanC is now on the phone...

20:37:12 <tim-gone> tim-gone is now known as tbl

20:38:14 <sbp> s/personell/personnel/

20:40:28 <JibberJim> Sean?

20:40:30 <sbp> ooh, running code

20:40:32 <sbp> Hi Jim

20:40:55 <JibberJim> Do you think there's any scope for improving on the EARL model, or can we push for a stable EARL 1.0 ?

20:41:25 <JibberJim> oh and we need a schema for EARL, for my EDITOR.

20:41:52 <sbp> Hmm... I think that's something that the group as a whole certainly need to discuss. Although, IMO, the model has been solid for quite a while - don't think it needs much tweaking

20:41:57 <sbp> EDITOR?

20:42:27 <JibberJim> The RDF Editor.

20:42:35 <sbp> aha

20:42:52 <JibberJim> - http://jibbering.com/rdf/editor.html - it's almost stable enough to "release" in the new version.

20:43:27 <sbp> we have a schema: http://www.w3.org/2001/03/earl/0.95.rdf

20:43:47 <sbp> the dropdown box is a great idea...

20:44:07 <JibberJim> It doesn't seem to be able to get that schema :-(

20:44:34 <sbp> Hmm... you're right. I wonder what borks it?

20:44:54 <JibberJim> well my schema parsing is all a bit of a hack.

20:54:06 <sbp> quotable notable for the week: "It may be that RDF will fill a simple role in simply expressing the equivalence of the terms in each database schema." - http://www.w3.org/DesignIssues/RDB-RDF

20:56:16 <JibberJim> It seems to be that WWW-authenticate: Basic realm="W3CACL" header that mucks it up.

20:57:26 <sbp> why is that mucking it up?

20:57:53 <sbp> Hmm... where'd that 1.0 draft go to...

20:58:03 <JibberJim> dunno, a bug in MS's xmlhttp object presumably.

20:58:03 <sbp> oh, do we have an ERT meeting this week?

20:58:19 <JibberJim> Dunno Sean, I'm assuming not, but will pop in at midnight just in case.

20:58:33 <sbp> O.K.

20:58:37 <JibberJim> - It's not got any http://www.w3.org/2000/01/rdf-schema#isDefinedBy predicates - that's how I parse Schemas.

20:59:05 <sbp> I'll convert the latest to XML RDF...

20:59:29 * JibberJim doesn't understand schemas

21:00:48 <sbp> look for things for type rdfs:Class or rdf:Property

21:01:38 <sbp> well, it's at http://lists.w3.org/Archives/Public/www-archive/2002Mar/att-0014/01-1.0-test.rdf

21:01:49 <sbp> but no isDefinedBy arcs. I just assume the racine

21:04:38 <JibberJim> so it's either object of the triple with predicate rdfs:isDefinedBy or object of those with subject:rdfsClass and rdf:property ?

21:05:34 <sbp> it's the subject of things with rdf:type rdfs:Class or rdf:type rdf:Property

21:06:12 <sbp> so, for example <rdfs:Class elements with rdf:about attributes

21:07:06 <sbp> or <rdf:Description> elements with <rdf:type rdf:resource="&rdfs;Class"/> elements in them, etc.

21:07:37 <sbp> wouldn't it be easier to hook up a proper JS RDF parser? :-)

21:07:44 <JibberJim> I have one!

21:07:57 <JibberJim> I think, well it creates the triples, but I don't know what to do with triples.

21:08:01 <sbp> ah

21:08:10 <sbp> ?x rdf:type rdfs:Class .

21:08:16 <sbp> ?x rdf:type rdf:Property .

21:09:48 <JibberJim> and I can not bother with the isDEfinedBy>

21:10:05 <sbp> yep

21:15:57 <sandro> Is there anyone on RDF Core here who could comment on the likelyhood of N-Triples getting "fixed" -- finishing defining wide character handling, and allowing all three identifier types in all three positions.

21:16:05 * JibberJim can import the 1.0-test earl now.

21:16:42 <AaronSw> sandro, why would we do the latter?

21:17:34 <sbp> and even literals as predicates?

21:17:39 <sandro> AaronSw, so that all RDF graphs could be serialized with it, and RDF graphs would be closed under reasonable (all?) operations.

21:18:14 <sandro> Yes, all -- because otherwise, it's not closed under daml:EquivalentTo, which is an obvious & necessary operation.

21:18:53 <sandro> That is , Yes, even literals as predicates.

21:18:58 <dajobe> don't hold your breath sandro. I'm editor of Ntriples and I'm not expecting it "fixed" as you describe it in this WG lifetime.

21:19:10 <AaronSw> blast you, dajobe

21:19:12 <AaronSw> ;-)

21:19:29 <AaronSw> sandro's reasoning is the best i've heard yet.

21:19:35 <dajobe> I've not heard a good use case for literals as predicates

21:19:48 <AaronSw> so that equivalentTo can be closed under RDF graphs.

21:19:51 <sandro> That's what I thought. Fortunately, I don't need something blessed by a WG. Still, sad.

21:20:14 <AaronSw> sandro, mind writing something up and I'll send it to the WG?

21:20:22 <dajobe> RDF doesn't have daml:equivalentTo - that would be a WebOnt / OWL thing

21:20:28 <sandro> How about finishing the unicode question, dajobe?

21:20:34 <dajobe> finishing?

21:20:45 <AaronSw> dajobe, we have subClassOf in both directions, which has the same meaning...

21:21:01 <AaronSw> and subPropertyOf

21:21:10 <dajobe> but we luckily have no processing model to use them :)

21:21:18 <AaronSw> we have the RDF Model Theory

21:21:37 <dajobe> I don't have to apply the rdfs rules if I don't want to

21:21:37 <AaronSw> so that'd require bNodes as predicates

21:21:44 <AaronSw> yes, but you should be able to

21:21:50 <AaronSw> if you do want to

21:21:54 <sandro> http://www.w3.org/2001/sw/RDFCore/ntriples/ leaves open the issue how to encode wide chars, right?

21:21:54 <dc_rdfig> C: http://www.w3.org/2001/sw/RDFCore/ntriples/ from sandro

21:22:00 <sandro> arg.

21:22:09 <AaronSw> C:|

21:22:10 <dc_rdfig> titled item C

21:22:11 <dajobe> sandro: write a use case, send to www-rdf-comments for the group

21:22:24 <AaronSw> sandro, that doc is out of date

21:22:24 <dajobe> as a comment to the model theory , that is

21:22:32 <AaronSw> to ntriples!

21:22:48 <dajobe> ntriples only changes when the model does

21:23:00 <dajobe> so if you need literal predicates, ask to get the model changed

21:23:05 <AaronSw> oh, ok then.

21:24:18 <dajobe> so going back to unicode sandro ...

21:24:34 <dajobe> ntriples can express any unicode character string

21:24:43 <dajobe> (in fact too many; I need to cut back)

21:26:14 <sandro> How, dajobe?

21:26:26 <sandro> Did you pick 1, 2, or 3?

21:26:36 <dajobe> eh? how what?

21:28:06 <dajobe> C:[http://www.w3.org/TR/rdf-testcases/#ntriples|NTriples format]

21:28:07 <dc_rdfig> added comment C1

21:29:22 <sandro> Ah, you picked 1. I missed the note about being superceded.

21:29:40 <sbp> C:|.

21:29:40 <dc_rdfig> titled item C

21:30:00 <dajobe> well, it's in the status section :)

21:31:45 <sbp> C:|<> a :Obsolete .

21:31:45 <dc_rdfig> titled item C

21:32:10 <dajobe> C:|NTriples original version (now superceded)

21:32:10 <dc_rdfig> titled item C

21:32:31 <dajobe> add a comment if you want sbp

21:35:21 <sandro> So the MT says that you can't use a URI to denote a string?

21:35:28 <sbp> Hmm... if I put in an anon node, the chump will just think it's a URI. Pff

21:35:30 <dajobe> strictly, no

21:35:48 <sandro> And everyone knows that

21:36:01 <sandro> And everyone knows that's wrong, but out-of-scope to fix?

21:36:22 <dajobe> no idea

21:36:58 <dajobe> it probably should encourage the use of strings to denote strings

21:37:05 <dajobe> rather than say you can't use URIs

21:37:28 <dajobe> is that right, or is this late night coffee I'm drinking too strong

21:37:38 <sandro> So if I want to transmit an RDF Graph that's been through an inference step, I should define my own format, eg S-Triples (defined as N-Triples, where the subject and predicate productions are like the object production.)

21:38:01 <dajobe> sure

21:38:18 <dajobe> probably S-Triples is-a-subset-of N3, so probably you are OK

21:38:19 <sandro> Yeah. Okay.....

21:38:53 <dajobe> making the triple uniform has been discussed several times on rdfcore but we haven't moved there yet

21:39:05 <dajobe> I can see literal subjects being added though

21:39:09 <DanCon> which n-triples infelicity are you asking about sandro? literals in the predicate slot seems like one of the less painful ones.

21:39:27 <DanCon> bnodes as predicates has more bite, for me. And literals as subjects.

21:39:43 <dajobe> bite=good or bad?

21:39:51 <DanCon> bite=bad. i.e. lack of it hurts.

21:40:04 <dajobe> bnode predicate sfor query I assume?

21:40:05 <DanCon> or rather: I use it (anonymous predicates) all the time in N3.

21:40:11 <sandro> DanCon, I'm thinking about all of them -- from the angle of wanting to serialize the output of horn/cwm-style inference, no matter what the rules were.

21:40:47 <dajobe> well, that only leaves literal predicates

21:40:58 <DanCon> for query: more or less... for the same reasons most triplesMatching(S, P, O) APIs allow the null-wildcard in the P slot as well as the S and O slots.

21:41:20 <sandro> And literal predicated might have invalid semantics, but I want the syntax there, so I can separate the modules.

21:42:04 <DanCon> yes, I asked for literal predicates to be syntactically OK, even though such a triple would be false in all interpretations. On the grounds of simplifying the model theory. I didn't get much support.

21:42:29 <DanCon> but to be fair, my only argument was aesthetics.

21:42:31 <dajobe> I think Pat said it was crazy

21:42:37 * dajobe looking for quote

21:43:08 <DanCon> frap; my nifty new (just a few months old) LCD monitor is flickering madly. I wonder if a cable is out somewhere.

21:43:38 <sandro> Pat's come a long way. :-)

21:43:57 <dajobe> lots of education in many directions

21:43:59 <sandro> He's also know to overstate cases and be wrong, of course.

21:44:35 <dajobe> and change his mind once shown that

21:44:40 <DanCon> oh... which reminds me... I meant to send a message to rdf-core, mainly to pat, about the difference between foo:bar and <http://foo> and <foo:bar>...</foo:bar>.

21:45:29 <dajobe> yeah, I think I'll have to do a bit more n-triple policing too

21:46:26 <bijan> What's the canonical way to serialize n3 contexts in ntriples? not to do it?

21:47:03 <AaronSw> I know sandro's a fan of bags of reified statements

21:47:26 <sbp> or a single reified statement if the formula only contains one triple

21:47:43 <bijan> Just wondering, as I'm dorking around with a pretty printer for CWMClone.

21:47:53 <bijan> Seems like I might as well aim for N3 output.

21:48:01 <dajobe> is canonical==what cwm does?

21:48:03 <bijan> Rather than doing ntriples.

21:48:33 <bijan> dajobe, eh. Well, what it does - obslete stuff - bugs - things we're all embarassed about, sure.

21:48:35 <sbp> CWM outputs broken NTriples in many cases

21:48:50 <sandro> Bijan, Tim and I have talked about this, and the most recent view we've had is based on log:Conjunction.

21:48:58 <dajobe> yeah, one of the regexes is different from ntriples

21:49:00 <sandro> But cwm doesn't implement anything.

21:49:06 <DanCon> cwm has a --flatten thingy. it's broken. TimBL's never really answered the question of how to reduce {} to triples.

21:49:18 <bijan> I don't might being ahead of the CWM curve on this :)

21:49:21 <bijan> er..mind

21:49:23 <sandro> He has, DanCon, in e-mail with me. Let me try to find it.

21:49:32 <dajobe> regexes: I mean, the one for _:THING is different from n3/ntriple

21:49:51 <sbp> I think that was fixed

21:49:54 <sbp> the main problem is line breaks in literals

21:49:58 <sbp> and triple quoting

21:50:20 <dajobe> I've just updated my ntriples parse to roundtrip all that and xml("<b>...</b>") stuff

21:50:39 <DanCon> did you implement c14n while you were at it, dajobe?

21:50:40 <dajobe> which is ahead-of-the-curve given I haven't made the syntax doc change yet

21:50:48 <sandro>http://lists.w3.org/Archives/Public/www-rdf-logic/2001Dec/0031.html

21:50:48 <dc_rdfig> D: http://lists.w3.org/Archives/Public/www-rdf-logic/2001Dec/0031.html from sandro

21:50:51 <dajobe> DanCon: no, jeremy says I don't need to :)

21:51:07 <DanCon> really?

21:51:08 <dajobe> however Daniel Veillards' (+others) libxml got a new c14n module this week

21:51:33 <dajobe> but that would be in the RDF/XML parser anyway

21:51:42 <sandro> D:|Discussion of how to flatten n3, reifying a {formula} into an RDF Statement or log:Conjunction of RDF Statements.

21:51:42 <dc_rdfig> titled item D

21:52:31 <DanCon> er... I don't see any {}'s reduced to triples in D. what am I missing?

21:53:15 <DanCon> dajobe, do you have an ntriples-comparison implementation? do you expect c14n to be there? or, as you said, in the RDF/XML parser?

21:53:19 <sbp> Statements are subclasses of formulae

21:53:37 <dajobe> DanCon: I've got jang's C++ one 'ntc' that I'll probably convert to C, extend

21:53:50 * bijan hopes the formulae/context terminology settles at some point soon :)

21:54:09 <sandro> I think it's settled Bijan.

21:54:11 <DanCon> that terminology _is_ settled, to my satisfaction.

21:54:15 <dajobe> for c14N: not sure where it would go since the words jeremy is intending to propose allow late c14n at the point of comparison

21:54:30 <sandro> There are some old uses of "context" meaning "formula" floating around, alas.

21:54:51 * dajobe found pats quote on literal predicates

21:55:16 * bijan changes his hope to that a pointer to the Definitive Guide to Formulae/Context terminology crops up soon :)

21:55:18 <sandro> DanCon: The main thing in D is "{ <a> <b> <c>. <d> <e> <f>.} log:conjunction ( { <a> <b> <c>.} {<d> <e> <f>} )."

21:55:21 <sbp> so... statements are subclasses of formulae with a single triple in them, but luckily all >1 triple formulae can be expressed as a conjunction of single triple formulae

21:55:28 <sbp> snap

21:56:03 <DanCon> yeah, well, reduction to previously unsolved problem: {<d> <e> <f>}

21:56:08 <sbp> Gotta run

21:56:09 <sandro> DanCon: ANd he agrees { <a> <b> <c>.} is simply an RDF Statement. He doesn't say how to reify an RDF statement, it's true.

21:56:09 <bijan> right.

21:57:05 <DanCon> rdf:predicate/subject/object vocabulary is mighty screwey. How do you reify _:foo, for example?

21:57:10 <sandro> right. at the time of that discussion, I think we thought it was solved. but yes -- it's not quite solved.

21:57:36 <DanCon> I think you had a working/workable FOL schema that didn't have the bugs that rdf:predicate/subject/object has.

21:57:57 <bijan> Sandro, the log:conjunction partial solution suggests that formulae/contexts are order sensitive, is that right?

21:57:59 <sandro> agreed. My best solution is in that FOL schema. But it is even MORE verbose.

21:58:39 <DanCon> well, more verbosity -- or more information of some kind -- is necessary, from what I can tell.

21:58:59 <sandro> bijan, I think Tim would say they are not order sensitive. I'm tempted to keep the order around for debugging and performance, but it should never effect semantics.

21:59:35 <DanCon> we could get away with subjectConstant, subjectExistential, predicateConstant, predicateLiteral, etc. i.e. the reification vocabulary has to be as expressive as the n-triples grammar. (at least)

21:59:51 <sandro> Right -- that's the other option.

22:00:13 <bijan> Indeed, the shuffling CWM allows is killing me :)

22:00:45 <bijan> Well, not killing, really, but adding several layers of complexity.

22:00:56 <sandro> Huh.

22:01:23 <bijan> Being as I translate an N3 rule to a prolog rule, and prolog rules *are* generally order sensitive.

22:01:31 <sandro> dajobe: But it doesn't say how to reify an RDF Statement

22:01:40 <sandro> arg. sorry dajobe

22:01:42 <sandro> D:But it doesn't say how to reify an RDF Statement

22:01:43 <dc_rdfig> added comment D1

22:02:34 <sandro> D:For which we need something more than RDF M&S, because we need to distinguish uriref, nodeID, and literals.

22:02:34 <dc_rdfig> added comment D2

22:02:41 * jonb was wondering how you were going to convert _:foo into URIrefs

22:03:17 <sandro> This WOULD kind of go away if we just used data: for literals and urn:local: for bNodes.

22:03:48 <bijan>http://www.isg.sfu.ca/~hak/documents/wam.html

22:03:48 <dc_rdfig> E: http://www.isg.sfu.ca/~hak/documents/wam.html from bijan

22:03:51 <sandro> But mostly that just moves the problem around, without really solving it.

22:04:03 <bijan> E:|Warren's Abstract Machine: A Tutorial Reconstruction

22:04:03 <dc_rdfig> titled item E

22:04:29 <bijan> E:Prolly chumped before, many a time, but the classic description of the Prolog implemenation that made Prolog a practical tool.

22:04:29 <dc_rdfig> added comment E1

22:04:44 <jonb> urn:local sort of works except that software needs to be specially aware of this, right?

22:05:16 * tbl ctaches up

22:05:30 * bijan decides to go with n3 output :)

22:08:51 <sandro> Run, jonb. I think the code ends up being about the same, really, either way. Objects are still denoted with local-scope names, global-scope names, or (as a special case, if they are unicode strings) their contained characters in sequence.

22:08:57 <sandro> s/run/RIght!

22:09:39 * jonb was running but stopped

22:10:29 <DanCon> prolog order sensitive? really? pure prolog?

22:10:43 <bijan> Hmm.

22:10:49 <bijan> I believe so, yes.

22:11:37 <DanCon> hang on... the output of a pure prolog computation can change as a result of reordering the KB?

22:11:48 <DanCon> or reordering the premises in a rule?

22:12:25 <DanCon> my intuition says otherwise. I'd need an example to be convinced.

22:12:50 <larsbot> recursion affects this, at least

22:12:50 <DanCon> E:doesn't sound familiar...

22:12:50 <dc_rdfig> added comment E2

22:12:51 <bijan> Yes, I believe so, if what you mean is "Is a pure prolog computation running on a real implemenation order senstivie"

22:13:13 <bijan> E:It's the classic. But it's obscure. Nice that it's on the web now.

22:13:13 <dc_rdfig> added comment E3

22:13:27 <bijan> A sequential implementation of Prolog is characterized by a depth-first, left-to-right visit of the computation tree. This fixes the order in which each subgoal is processed.

22:13:39 <bijan> From: http://www.cs.nmsu.edu/~epontell/adventure/node50.html

22:13:42 <DanCon> E:note: "If you retrieve it, all I ask is that you let me know who you are, and for what purpose you intend to use it.

22:13:42 <DanCon> "

22:13:42 <dc_rdfig> added comment E4

22:14:39 <DanCon> E:chuckle... copyright in the PDF version seems to conflict with the license on this web page

22:14:39 <dc_rdfig> added comment E5

22:15:02 <bijan> DanCon: better reference: http://www.cs.nmsu.edu/~epontell/adventure/node9.html

22:15:12 <sandro> em: I would say pure prolog is not generally order sensitive. Almost nobody writes pure prolog, of course.

22:15:25 <dajobe> hehe

22:15:30 <dajobe> do X:nospacehere

22:15:56 <bijan> Well, it depends on what you mean by "pure prolog". If pure prolog has an operational semantics, and ought to respect it, then I'll bet it's order senstive.

22:16:57 <DanCon> node9 says, as I expected, "In general, no restrictions are imposed on the order in which the subgoals and the clauses to be used are selected

22:16:57 <DanCon> "

22:17:41 <DanCon> meanwhile, "The semantics of Prolog are heavily based on the ordering adopted by the two selection functions--which is fundamental for the correct behaviour of the various extra-logical features of the language.

22:17:41 <DanCon> "

22:18:08 <bijan> Yes, the "In general" doesn't apply to prolog :)

22:18:42 <bijan> Pure Prolog - Prologian operational semantics ~= order sensitive

22:19:10 <DanCon> yeah... I was thinking of mapping N3 <-> algernon. Algernon cares about order too. I tried to imagine convincing timbl that N3's formulas should be ordered. I couldn't even make the case to myself.

22:19:31 <bijan> Mercury and some other lp languages aren't order sensitive.

22:20:13 <bijan> For the most part.

22:20:28 <bijan> Mercury does agressive reordering of conjuncts in clauses.

22:20:40 <larsbot> for optimization?

22:20:45 <DanCon> E:ugh... for something called "tutorial", there are an awful lot of greek letters in here.

22:20:45 <dc_rdfig> added comment E6

22:20:47 <bijan> yep.

22:21:02 <larsbot> right. I do the same in my prolog-like TMQL

22:21:08 <sandro> Do any SQL implementations encourage you to think about ordering your terms in the WHERE clause?

22:21:21 <larsbot> is that a rhetorical question?

22:21:34 <bijan> E:Yes, the "tutorial" is notoriously brutal. I went through in in a class. I survived. I think.

22:21:35 <dc_rdfig> added comment E7

22:21:43 <sandro> No, I don't know the answer.

22:21:49 <DanCon> SQL: I don't think so. that would be like trying to outsmart an optimizing fortran compiler.

22:21:49 <dc_rdfig> Label SQL not found.

22:21:55 <larsbot> as far as I know all SQL implementations do heavy optimizing

22:22:11 <sandro> Sounds like n3 should be non-order-sensitive.

22:22:22 <bijan> Eh..

22:22:23 <tbl> Re literals as Properties, I am happy for Literals to be distinct data type jsut as Literals and formulae are distinct.

22:22:55 <tbl> Bijan, Formulae are defintly order-free. I generally sort them by s then p then o on output.

22:23:05 * tbl is still reading way up th epage...

22:23:31 <bijan> *Really*...

22:23:38 <bijan> That's interesting.

22:24:36 <sandro> And the aparent ordering in our reification goes away when you realize that log:Conjunction's semantics say it doesnt care about order.

22:24:39 <bijan> In a make the implementation harder kind of way :)

22:26:41 <sandro> why? can't you just pick any order you want, bijan?

22:26:54 <bijan> No.

22:26:59 <bijan> Well, maybe.

22:27:01 <sandro> why not?

22:27:03 <sandro>http://www.w3.org/2002/03/s-triples/

22:27:04 <dc_rdfig> F: http://www.w3.org/2002/03/s-triples/ from sandro

22:27:14 <bijan> Because you'll get more failures on some orderings.

22:27:26 <bijan> er..more failed inferences.

22:27:28 <sandro> F:| A name for N-Triples without restrictions on the type of terms in each position

22:27:28 <dc_rdfig> titled item F

22:27:59 <bijan> I.e., if the semantics of log:implies was "Some order of the antecedent, up to the compiler which"

22:28:09 <sandro> Why would it fail -- they'll all be tried on backtracking.

22:28:56 <tbl> When it comes to the antecedent of a rule, cwm (slowly and carefully ;-) looks at the various statements it is trying to find and figures out which ones are going to work fastest, and does those first.

22:29:02 <sandro> Since the antecedents are just conjoined, how could their order matter?

22:29:23 <bijan> Because some fail if their vars arent' fully bound.

22:29:36 <bijan> So if the binding statement comes after that conjuct, the conjunctionalways fails.

22:29:44 <sandro> But they fail in a back-trackable way.

22:29:55 <sandro> (they should)

22:30:32 <tbl> The order you try to find them matters. { :x a :targetDocument. :x :semantics :F. :F log:incldues :G } for example you don't start by finding the semantcis of all the documents inthe universe and then seeing whcih ones were doeumnts you had selected as targets.

22:30:44 <tbl> They don't fail, they just getpostponed till later.

22:30:45 <bijan> {?x startsWith: "B". :bijan name: ?x} will always fail.

22:30:51 <bijan> If it's order senstive.

22:31:02 <bijan> string:startsWith:

22:31:12 <bijan> See atom_prefix/2.

22:31:21 <sandro> You guys are using impure code hacks.

22:31:36 <tbl> sring startswith is a light built-in. It can only run when it has no unbound variables.

22:31:49 <sandro> You should make the semantics of them such that they don't look like impure code hacks. :-)

22:32:08 <tbl> I don't regard this as a fudge at all.

22:32:25 <bijan> er..trying all working orderings seems to get reasonable purity.

22:32:35 <bijan> Prolog doesn't do that, however.

22:32:38 <bijan> So I have to :)

22:32:45 <bijan> CWM does.

22:32:50 <sandro> How do you reconcile string:startsWith caring about terms being bound AND order not mattering?

22:33:07 <bijan> there's two things that CWM shuffles, for performance and for bindings.

22:33:09 <tbl> There are very clear distinnctions. You can basically make a list, of all the unbound variables. Then you make a list of al;l the unbound varibales which need to be cleared before you can run each builtin.

22:33:22 <bijan> Well, if ?x *never* got bound, that's a proper failure, yes?

22:33:26 <tbl> The order in the original formula makes no difference.

22:33:40 <DanCon> sandro, since order doesn't matter, cwm can choose an order where string:startsWith isn't evaluated until everything's bound. (or give up trying)

22:33:43 <tbl> How you go eveluate a query will always involve a bit of thouhgt.

22:34:00 <bijan> Prolog just requires you to do such orderings manually.

22:34:15 <bijan> CWMCLone could do a dynamic shuffling a la CWM.

22:34:18 <tbl> (broing!)

22:34:22 <tbl> (boring!)

22:34:32 <DanCon> manually? really?

22:34:36 <bijan> Or a static analysis of the formula and generate a set of rules that cover the feasible shuffles.

22:34:36 <tbl> Must take a lot of thought, and not always the same.

22:34:44 <bijan> DanCon: yes.

22:34:50 <DanCon> blech.

22:35:17 * tbl should definitely patent cwm ;-)

22:35:30 <bijan> There are times when the order of clauses in the KB matter, as well as the order of terms in a clause.

22:35:32 <sandro> Okay -- I guess I was thinking code hacks like string:startsWith could hide the reording inside themselves, but the other way works, too.

22:35:36 <bijan> Hence asserta assertz, etc.

22:35:46 * tbl really has to look for sbp's performance test message and profile cwm.

22:35:54 * DanCon remains blisfully unaware of asserta, assertz, etc.

22:36:20 <bijan> asserta/1 asserts the clause at the beginning of the kb, assertz at the end.

22:36:23 <MarkB> A: I tossed this question to [rest-discuss|http://groups.yahoo.com/group/rest-discuss/message/863]

22:36:23 <dc_rdfig> added comment A4

22:36:36 <sandro> But Bijan, again, that's impure prolog. (which matters in prolog programming, but not in analogy to cwm.)

22:36:38 * tbl reminded of <a> :outputString ;x

22:36:51 <dmiles> yay prolog talk

22:37:02 <DanCon> LOL

22:37:02 <tbl> beginning? end? aaagh

22:37:24 <bijan> sandro, well duh, I was just pointing out how order matters in prolog :)

22:37:30 <sandro> Most prolog programs care a LOT about the order in which solutions are found.

22:37:34 <dmiles> beginning or end is in a KB is is heuristic

22:37:43 <tbl> Ok, from now on log:asserta assrets a formula F at the beginning of the web.

22:37:52 <sandro> :-)

22:37:57 * dmiles keyboard stuttered

22:38:30 <bijan> Not all logic programming languages are like this. Prolog makes these tradeoffs because they're reasonably practical.

22:38:34 <bijan> Especially historically.

22:38:56 <bijan> Prolog has *always* been billed as a hybrid language.

22:39:13 <tbl> I remember DanC how you had to point out to emiller that his rules wouldmn't work as cwm had nowhere to start. So it obviously is possible to make rules which don't work for that reason, though we get so used to makeing ones which do, we forget.

22:39:16 <dmiles> yes a good viewpoibnt of prolog is that it is a priogrtamming language.. and not an infernce tool

22:40:06 <dmiles> i think when one approaches prolog at first they think it will do their heuristics but they learn that asserta vs attertz is very important practivcally

22:40:35 <connolly> or one goes "blech. if I want to tell the machine exactly what to do, I'll go back to c/python/java, thanks." ;-)

22:40:48 <connolly> connolly is now known as DanC

22:41:23 <bijan> <shrug/>*Nobody* says prolog is the final programming language :)

22:41:46 <sandro> Warren (David S.) argues that prolog performance is generally better if you avoid assert/retract and other impure constructs.

22:41:58 <bijan> Most people say that.

22:42:02 <bijan> Well, many.

22:42:19 <bijan> O'Keefe's *Craft of Prolog* is *dedicated* to the proposition :)

22:42:23 <dmiles> it is becasue it requires copying it is much faster to pull the term into some structure that you cary with you

22:42:29 <sandro> I was replying to dmiles saying "asserta vs attertz is very important practivcally"

22:42:49 * DanC wonders how tbl's profiling is going

22:43:03 <bijan> Well, practically, it's important because of the determinante search order.

22:43:09 <dmiles> so i avoid assert retract unless it is globally important .. but bnevber use them during a backchain

22:43:54 * tbl was distracted by DanC's dbview!

22:44:11 <dmiles> generally when the user does something to make a state chainge i'll assert .. buit if i am dealing in a hypothetical construct.. for a durration of an inferance i wont use assert/retract

22:44:15 <DanC> hmm... family time in T-15... crack open dbview again? hmm...

22:44:45 * sbp catches up

22:45:49 <tbl>http://www.w3.org/2000/10/swap/db.n3

22:45:50 <dc_rdfig> G: http://www.w3.org/2000/10/swap/db.n3 from tbl

22:45:58 <DanC> frap frap frap. my net connection to MIT is becoming unliveable. I'm gonna have to give up IMAP for POP pretty soon.

22:46:11 <tbl> That is a schema for exposing a databse on teh web as served by dbview

22:46:18 <tbl> I forget how to chump that

22:46:23 <DanC> you just did.

22:46:25 <DanC> oh...

22:46:33 <DanC> G:|a schema for exposing a databse on teh web as served by dbview

22:46:33 <dc_rdfig> titled item G

22:46:52 <sbp> Tim: 403, but available at http://dev.w3.org/cvsweb/2000/10/swap/db.n3 anyway...

22:46:54 * tbl notes DanC not alone with netprobs

22:47:23 <tbl> 403 is weird ... I have cganges teh access many times but I get 403 too. bug ...

22:48:00 <DanC> say, tim, please use complete sentences for rdfs:comment. i.e rather than rdfs:comment "The document which describes the form of this database";

22:48:27 <DanC> say: for databaseSchema(DB, S), read: database DB is described by document S.

22:48:54 <DanC> or just: database SUBJ is described by schema document OBJ.

22:48:54 * sbp ponders a string:canonicalized

22:49:11 <dmiles> generally i agree with what everyone said about prolog :) the use differnce with asserta/z is for the practicallity of their program.. but most of the time it is better to avioid their use unless you really are trying to modify behavour of the program long term

22:49:30 <tbl> I had at one time institutionalize "this" and "that" for SUBJ and OBJ but i couldn't remember shich!

22:49:33 <tbl> which!

22:50:07 <DanC> soc:Work? hmm... been a while since I looked at soc

22:50:31 <DanC> did you mean doc:Work?

22:50:32 <tbl> A docuemnt by any oterh name

22:50:34 <sbp> Hmm... soc: isn't declared as a prefix

22:50:42 <tbl> Oh ... ooops s/soc/doc/gg

22:51:59 <DanC> does swap/pim/doc.n3 capture pretty much all your ideas about generic resources? I've long wondered how to formalize generic resources.

22:52:00 <tbl> r1.3

22:53:24 * DanC is still here

22:53:50 <sbp> on that note, the terms in http://www.w3.org/DesignIssues/Generic prefixed with u: could be moved to doc.rdf/doc.n3

22:54:38 <DanC>http://www.w3.org/DesignIssues/Generic

22:54:38 <dc_rdfig> H: http://www.w3.org/DesignIssues/Generic from DanC

22:54:44 <sbp> then they'd have a datespace URI, and I'd be happier recommending referencing them from the EARL schema, etc.

22:54:52 <DanC> H:|Generic Resources

22:54:52 <dc_rdfig> titled item H

22:55:16 <DanC> H:under the principle that designissues and implementations should point to each other, this one should be linked...

22:55:16 <dc_rdfig> added comment H1

22:55:45 <DanC> H:to [swap/pim/doc|http://www.w3.org/2000/10/swap/pim/doc.n3]

22:55:45 <dc_rdfig> added comment H2

22:56:05 <DanC> sean, pls chump any compare/contrast notes you find

22:56:45 <DanC> MarkB, still around? does u:TimeInvariant look familar?

22:56:46 <sandro> (re H) Isn't the King James Version necessarily in English? Anything else would be a translation of the KJV.

22:56:54 <tbl> I haven't looked to see whetehr doc has all I would want to give URIs to. After all,in principle, all kidsn of otehr relationships can arise. (Is doc:reprinting etc)

22:57:23 <tbl> That depends on wt athe person sho gives a URi for the KJV defines it to stand for.

22:57:41 <tbl> Whether it is necessarily in emnglish depends on wt athe person sho gives a URi for the KJV defines it to stand for.

22:57:56 <DanC> I have a hard time seeing the bounds of what's related to what by this "generic" relationship.

22:58:37 <sbp> H:IMO, The terms (properties and classes) under "Using RDF to model this" could be moved to the PIM doc.n3 schema, since describing the genericity of documents is useful for specifiying the semantics of predicates such as :archivedVersion, :persistentCopy, etc.

22:58:37 <dc_rdfig> added comment H3

22:58:39 <sandro> (wow, old W3C logo. Old document!)

22:58:39 <tbl> It is up to the publisher and therefore ratehr unbounded and theerfore mcuh safer when described with a nice bit of RDF.

22:59:23 <tbl> Aaaar, if yer wanna see what oid call a *real& w3c logo, me boy, oi could show yer summat! ;-)

22:59:28 <DanC> ok, so neither do you have any general-purpose, formalizable notion of "generic". Glad I'm not missing anything. ;-)

22:59:44 * sbp giggles

23:00:05 <DanC> frap; there went my 15minutes.

23:00:24 <sandro> Wait -- doesn't u:TimeInvariant talk about the mapping from URI to resource? In which case you can't say that in RDF?

23:00:28 * timbl wndres if family time ends

23:00:35 <DanC> sbp, do you know what I'm talking about when I ask MarkB about TimeInvariant? pls chump his internet draft if you find it.

23:00:39 <sandro> when they turn 18, I hear.

23:00:46 <timbl> <db.n3> :ERROR_PREDICATE_NOT_DECLARED_AS_PROPERTY <http://purl.org/dc/elements/1.1/creator>,

23:00:46 <timbl> <http://purl.org/dc/elements/1.1/title>;

23:00:46 <timbl> :ERROR_TYPE_NOT_DECLARED_AS_CLASS daml:uniqueProperty .

23:00:52 * sbp goes searching

23:01:34 <sbp> DanC, is this it?: http://www.ietf.org/internet-drafts/draft-baker-http-resource-state-model-01.txt

23:01:44 <timbl> don't you belive this "18" stuff Sandro... after 35 they still come back but they just bring kids too!

23:01:47 <DanC> I think so, sbp

23:02:09 * timbl don't you belive this "35" stuff Sandro... after 60 they still come back but they just bring grandkids too!

23:02:15 <DanC> seen Parenthood, everybody? "I'm 56. He's 37. And he's still my son."

23:02:31 <DanC> great line.

23:02:36 <timbl> .google Parenthood

23:02:38 <xena> Parenthood: http://www.plannedparenthood.org

23:02:42 <DanC> no, no!

23:02:44 <sbp> H:cf. [http://www.markbaker.ca/|MarkB]'s [http://www.ietf.org/internet-drafts/draft-baker-http-resource-state-model-01.txt|draft-baker-http-resource-state-model]

23:02:44 <dc_rdfig> added comment H4

23:03:02 <DanC> .google parenthood steve martin

23:03:04 <xena> parenthood steve martin: http://www.the-movie-times.com/thrsdir/actors/smartin.html

23:03:15 * DanC wanders off...

23:03:55 <sandro> Am I wrong that RDF predicates talk about resources, not the logical symbols (URI's) used to denote them? Unless they're really magic syntax, which isn't RDF.

23:04:57 <sbp> if it were all about symbols, what would the objections to literals-as-predicates be? :-)

23:05:00 <sandro> H: These are magic predicates, like log:uri, which somehow know what URI is being used to denote some resource. Not normal RDF model stuff.

23:05:00 <dc_rdfig> added comment H5

23:05:05 * timbl Hmmm. not surprised if the DC schema is a mess, but daml:uniqueProperty ... aah spelling error.

23:05:25 * timbl validate.n3 straikes again .. fighing for bad typers everywhere.

23:05:30 <sbp> lol

23:05:37 <DanC> 6rttdfftdftvgvyfgtf ddddddddddddddddddddrserssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssyhfgtfffffffdtddddddd5dcessrseeeeeeeeeeeaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa

23:05:37 <DanC> ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssweaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaqytvgftrffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff

23:05:38 <DanC> fffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffrrrrrrdcrdxdddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddffersdddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddressesssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss

23:05:45 <DanC> ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssswqaqwSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSS

23:06:02 <DanC> DanC is now known as KyleC

23:06:12 <timbl> Hiya Kyle!

23:06:15 <KyleC> llllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllll

23:06:16 <sandro> sbp, literals are the objects themselves. The four characters "DanC" are not a relationship between 2 objects -- they are a string!

23:06:21 <sbp> heh. Thought he'd gone to sleep on his keyboard for a second then...

23:06:35 <timbl> happens to the best of us

23:06:42 <sandro> Gee, family time is hard on ALL of us now.

23:07:45 * AaronSw waves to KyleC

23:08:10 <AaronSw> so kyle, # or / ?

23:08:15 <sbp> heh, heh

23:08:35 * timbl goes off too

23:08:50 * sandro tries to imagine "hash" vs "slash" in a child-raising context.

23:09:32 * sandro wanders off, too.

23:09:42 <KyleC> photos of me, about a year ago: http://dm93.org/y2001/kyle1/ksimle

23:10:18 <sbp> ooh, baby photos

23:10:28 <sbp> well, toddler

23:12:34 <tbl> Cool, kyle

23:20:08 <KyleC> jjjjjjjjjjjjjjjjjjjjjjjjnjmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmjn

23:20:25 <KyleC> jjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjj

23:20:37 <KyleC> dddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddd

23:20:52 <AaronSw> Now put them together... and you've got an RDF triple!

23:21:04 <KyleC> sssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss

23:21:14 <KyleC> wwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwww

23:21:29 <AaronSw> exactly: rdf is part of the sw

23:21:35 <dajobe> at this juncture, a good time to note this channel is nearly 1 year old

23:21:38 <KyleC> qqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee

23:21:41 <dajobe> first log was 2001-03-09

23:21:54 <KyleC> zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz

23:22:12 <KyleC> rrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr

23:22:27 * KyleC wanders off to #KyleScribble

23:22:51 * KyleC pitches a fit, returns

23:22:59 <AaronSw> Heh.

23:23:08 <KyleC> sssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss

23:23:20 <KyleC> xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

23:23:38 <KyleC> ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff

23:24:00 <KyleC> xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

23:24:09 <KyleC> zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz

23:25:19 <AaronSw> DanC: your job is in jeopardy ;)

23:25:50 <AaronSw> he just needs a few more months to learn formal logic...

23:28:06 * KyleC accepts cookie bribe, wanders off...

23:32:44 <sbp> dajobe: 1 "logging year" that is :-) it was around for a few weeks before logger got here

23:33:37 <AaronSw> Log file opened at: 2001-04-24 7:30:16 AM

23:35:23 <dajobe> aaron registered this channel 1yr 3 days ago

23:48:05 <sbp> my personal logs start as:-

23:48:07 <sbp> Start of #rdfig buffer: Sun Mar 04 23:59:14 2001

23:48:09 <sbp> *** Now talking in #rdfig

23:48:13 <sbp> but we'd been in here previous to that, I believe

23:49:07 <sbp> s/previous/prior/

23:50:08 <dajobe> my first log is Feb 15

23:50:13 <dajobe> but yes

23:54:31 <sbp> heh: "@@ no #". [Tim on http://purl.org/dc/elements/1.1/ in http://dev.w3.org/cvsweb/2000/10/swap/db.n3?rev=1.5&content-type=text/x-cvsweb-markup]

23:54:45 <AaronSw> heh!