Semantic Web Interest Group IRC Chat Logs for 2002-08-06

00:18:43 <tansaku_xr> .google rdfig

00:18:44 <datum> rdfig: http://rdfig.xmlhack.com/

00:48:38 <AaronSw>http://www.google-watch.org/krane.html

00:48:38 <dc_rdfig> A: http://www.google-watch.org/krane.html from AaronSw

00:48:44 <AaronSw> A:|Google's Privacy Policies

00:48:45 <dc_rdfig> Titled item A.

00:48:54 <AaronSw> F:"With the new Patriot Act, the use of the GET instead of the POST method [will let] authorities can claim that the search terms are part of the URL, and that they get logged with the URL in normal httpd logging."

00:48:54 <dc_rdfig> Label F not found.

00:49:05 <AaronSw> A:"With the new Patriot Act, the use of the GET instead of the POST method [will let] authorities can claim that the search terms are part of the URL, and that they get logged with the URL in normal httpd logging."

00:49:06 <dc_rdfig> Added comment A1.

00:49:17 <AaronSw> A:An interesting anecdote for the GET vs. POST debate. ;-)

00:49:17 <dc_rdfig> Added comment A2.

00:50:10 * danbri blinks

00:51:56 * tansaku_xr thinks AaronSw has developed an irc allergy

00:53:51 * danbri nods; he's trying to kick the habit

00:56:05 * DanC_jam waves from Montreal

00:56:27 <DanC_jam> ok, well, I've got cwm put back together, after I scrubbed its relative URI handling

00:57:07 <danbri> hi there

00:57:25 <danbri> did you fix the thing I was grumbling about the other day, where it passes thru <> without absolutising?

00:57:37 <DanC_jam> I was looking at (a) support for typed literals, and (b) convergence with eikon's code. I couldn't figure out the module interfaces, so I started cleaning.

00:57:55 <danbri> Oh, interesting.

00:58:02 <DanC_jam> fix: doubt it; with one exception, the test results are the same.

00:58:05 * eikeon waves

00:58:05 <danbri> I'm looking at that too, in a different context.

00:58:12 <DanC_jam> in general, cwm doesn't absolutize

00:58:32 <danbri> Ruby and Python are close enough that I'm thinking it feasible to transliterate eikeon's rdf parser, for sake of having a native ruby pares

00:58:36 <danbri> parser, i mean

00:58:40 <danbri> hi eikeon

00:58:47 <DanC_jam> the llyn.py module was importing thing.py, but only for diagnostic gizmos: verbosity() and progress()!

00:58:57 <danbri> ...in which case, since my api is rather fluid, using something similar makes sense

00:59:10 <DanC_jam> crap; conflicts during merge

00:59:36 * danbri looking at mozilla APIs again; things like attaching observers to graphs looks useful

00:59:47 <DanC_jam> a big question is: can URIrefs be shared across triple stores. in cwm/llyn: no. But I think they can be in eikon's code

01:01:02 <danbri> I'm thinking about doing everything via a Node.getResource() and Node.getLiteral() call, to avoid having multiple objects with the same content floating around

01:02:24 * DanC_jam thinks getResource is a use/mention bug. should be getURIRef or getSymbol

01:03:09 <danbri> hmmm

01:03:15 <danbri> works for bnodes too

01:03:32 <danbri> and I want each resource to expose its URIrefs as (generated) properties, if it has any

01:03:34 <DanC_jam> yup

01:03:48 <DanC_jam> ?

01:03:50 <danbri> I'm happy with nodes having multiple uriref labels (not sure RDF MT likes that...)

01:03:50 * eikeon Still thinking about URIRefs shared across triple stores or not... what is the case for not sharing them?

01:04:36 <DanC_jam> not sharing: in llyn, each URIRef knows all the things that are connected to it in its store.

01:04:56 <danbri> I do that with nodes, but not urirefs

01:05:09 <DanC_jam> i.e. you can do ref.occursAs[PRED] and it'll tell you all the triples (quads, actually) where it occurs as a predicate

01:05:26 <danbri> or rather, a node may have a ref to a graph (which may itself be a composite of other graphs, though not ported that bit from the perl yet).

01:06:02 * DanC_jam doesn't care much for the llyn style; prefers sharable URIrefs

01:06:12 * DanC_jam prefers *non* sharable BNodes, though

01:06:36 <danbri> it is handy to be able to call methods on a node that ask about its properties in some database

01:07:12 <danbri> eg mydoc.dc_title exapnding to 'find me the value(s) for http://purl.org/dc/elements/1.1/title' of the resource we call mydoc

01:07:12 * eikeon agrees with *non* sharable BNodes ;)

01:07:42 <danbri> I can leave with bnode identifiers leaking out, so long as under some explicit contract

01:08:29 <danbri> eg. my db might use uuids for bNodes, and vouch to recognise resouces by those ids for, say, 48 hours. Long enough for most interactive query sessions to rummage around the graph with iterated query.

01:08:34 <DanC_jam> another difference between swap/cwm and rdflib: the swap triplestore is actually a quad store; i.e. the store groks multiple contexts/formulas. I can't remember where that showed up, when I was reviewing the rdflib code...

01:09:17 <danbri> btw re 'rdflib', see http://www.xulplanet.com/tutorials/xultu/rdflib/

01:09:20 <AaronSw> oops. forgot process-issues was not public. soz DanC

01:09:21 <danbri> hi AaronSw

01:09:22 * DanC_jam finds cvs-updating big test files in swap really a pain over a slow link

01:09:50 <DanC_jam> ok, easy enough to forget, AaronSw

01:09:51 <eikeon> Adding support for contexts has been on my list for some time... was still researching how to implement it.

01:10:12 <DanC_jam> rdflib has a few closed-world gizmos, too... e.g. untypedResources()

01:11:01 <danbri> eikeon, can you link your cvsweb urls from rdflib.net?

01:11:25 <eikeon> Sure... will do that now.

01:11:30 <danbri> thanks :)

01:12:07 <danbri> also I can't see license terms anywhere... (without downloading)

01:12:21 * danbri trying to find out if can do GPL-only stuff as 'w3c software'...

01:12:32 <DanC_jam> eikeon, the rdflib code is really pretty nifty. Do throw in a comment here or there, though, would ya? docstrings, at least?

01:12:39 <AaronSw> ooh, pretty picture on DanC's webpage

01:12:58 <AaronSw> aha. extreme markup

01:13:04 <danbri> flatterer!

01:13:07 * DanC_jam waves from montreal

01:13:42 <danbri> I want to do something similar re RSS feed sources, map all SemWeb research groups in Europe (and beyond)

01:13:46 <danbri> but especially Europe :)

01:15:21 * DanC_jam wishes the cwm modules were separately unit-testable

01:15:49 <eikeon> DanC_jam: Thank you. Sorry about the weak documentation... I know better. Will set aside the time to bring them at least up to par.

01:17:03 <DanC_jam> clean code is better than docs... but you do a lot of mix-ins and such, and I wonder how certain things are expected to be used

01:19:58 <danbri> something i'd like to know... for those of us doing generic rdf interfaces without muchthought for inferences, rules etc

01:20:17 <danbri> ...how can we do something that allows cwm-like facilities to be layered on top later, cheaply...

01:20:29 <danbri> maybe quads / grouping constructs is key thing?

01:21:02 <AaronSw> statement ids

01:21:24 <danbri> same thing?

01:21:34 * danbri has those in backend store, not manifested in api yet

01:23:22 <DanC_jam> statement ids are almost the same thing, but two statements that share a bnode share a context. dunno if they'd share a statement id

01:23:59 <DanC_jam> hi tim... zat you? I'm doing a bit of cwm cleanup... (replay:)

01:24:01 <danbri> eikeon, in http://rdflib.net/latest/doc/triple_store.html is 'core layer' the 'layer 0' referred to?

01:24:12 <DanC_jam> the llyn.py module was importing thing.py, but only for diagnostic gizmos: verbosity() and progress()!

01:24:24 <DanC_jam> I was looking at (a) support for typed literals, and (b) convergence with eikon's code. I couldn't figure out the module interfaces, so I started cleaning.

01:25:44 <eikeon> danbri: Yeah, started out numbering them... then switched to names. All except for "1" by the looks of it :)

01:25:52 <danbri> heh

01:26:22 * danbri -> zzz

01:26:49 * eikeon waves... g'nite

01:28:18 <danbri> ps re typed literals, see http://lxr.mozilla.org/mozilla/source/rdf/base/idl/nsIRDFLiteral.idl for moz stuff

01:28:21 <danbri> g'nite

01:28:34 <DanC_jam> hmm... swap/test/underbarscope.n3 is missing

01:28:59 <DanC_jam> $ cvs update underbarscope.n3

01:28:59 <DanC_jam> cvs server: nothing known about underbarscope.n3

01:29:27 <DanC_jam> IOError: [Errno 2] No such file or directory: '/home/connolly/w3ccvs/WWW/2000/10

01:29:36 <DanC_jam> ... /swap/test/string/endsWith.n3

01:29:48 <DanC_jam> swap/test/time/t1.n3

01:30:00 <DanC_jam> Passed 41 out of 51 tests.

01:30:04 <DanC_jam> I think that'll do

01:30:22 * DanC_jam reviews changes

01:33:02 * DanC_jam wonders when nsIRDFLiteral.idl was written

01:35:23 * DanC_jam commits a pile of changes...

01:38:50 <eikeon> DanC_jam: Is there anything you would recommend for me to work on to help the convergence? Are contexts near the top of the list?

01:39:13 <DanC_jam> umm...

01:39:57 <DanC_jam> I'm not really sure... I think the thing to do is to look at your parser unit tests and see if you can run them on the cwm parser.

01:40:13 <DanC_jam> i.e. the sax2rdf.py module

01:40:35 <DanC_jam> btw... I noticed your serializer doesn't use any sort of SAX interface. no biggy, I guess.

01:41:24 <DanC_jam> anybody know how to get CVS to show me change logs from, say, 1 Aug to today?

01:42:48 <eikeon> Something along the lines of: cvs diff -D "Mon Jul 29 19:01:12 2002" -D "Mon Aug 5 19:01:12 2002"

01:43:00 <DanC_jam> that's diffs. I want change logs

01:43:21 <eikeon> Not sure what all commands two -D s will apply to... might work for others.

01:43:44 * DanC_jam RTFMs...

01:44:09 <eikeon> "-d datesSpecify dates (D1<D2 for range, D for latest before)."

01:44:20 <xover> DanC_jam: cvs log -D

01:44:24 <DanC_jam> yeah... cvs log -l -d '1 Aug 2002<6 Aug 2002' >,changelog

01:45:04 <DanC_jam> "minor". gee, thanks, timbl

01:46:07 <DanC_jam> "minor" includes a new command-line option, I see, reviewing diffs.

01:46:12 <DanC_jam> cwm --purg-rules

01:46:15 <DanC_jam> cwm --purge-rules

01:53:22 <DanC_jam> hm... having done all the grunt/prep work for typed literals, I no longer have much steam/enthusiasm for the interesting bit.

01:53:49 <DanC_jam> so... maybe some brainstorming... what should dates look like in N3?

01:54:45 <DanC_jam> when I first implemented N3, the idea was to use it in my palmpilot, where I wrote them: 1Aug2002.

01:55:46 <DanC_jam> :jenny :birthday date"2002-08-01".

01:56:25 <DanC_jam> :jenny :birthday date(2002-08-01).

01:56:37 * DanC_jam looks up SQL date syntax...

01:57:16 <eikeon> DanC_jam: BTW, all my sites are now done using an XHTML in RDF approach that I think originated from you and Tim and arrived at me via sbp. See: http://redfoot.net/redfoot/doc/XHTMLinRDF.html which is also has *no* comments :( But is short.

01:57:23 <DanC_jam> select * from STATION where starttime > '04/01/1997 14:23:59'; -- http://www.iris.washington.edu/SeismiQuery/SQ_syntax.html#dates

01:59:18 * DanC_jam reads XHTMLinRDF.html, fails to grok

02:00:58 <DanC_jam> hmm... sql seems to peek into schemas to know that things are dates

02:02:12 * DanC_jam tries to think of other languages that have date literals

02:02:25 <deltab> REBOL

02:04:13 <DanC_jam> "Reschedule exam for 2-January-1999 at 10:30" -- http://www.rebol.com/docs/core23/rebolcore-4.html

02:05:48 <DanC_jam> rebol literals: http://www.rebol.com/docs/core23/rebolcore-4.html#sect4.2.

02:07:25 <deltab> I'd prefer an ISO 8601 format though

02:09:08 <DanC_jam> got a favorite? which of these do you prefer? :jenny :birthday date"2002-08-01". :jenny :birthday date(2002-08-01).

02:09:44 <deltab> of those, the latter

02:11:04 <deltab> letters-before-string-literals is rather uncommon, afaik

02:11:42 <DanC_jam> daytime://everest.cclabs.missouri.edu -- http://www.rebol.com/docs/core23/rebolcore-16.html#_Toc487519750 (a URI scheme I've never seen before)

02:12:09 <deltab> yeah, rebol lets users define their own schemes at runtime

02:13:10 <DanC_jam> 1999/10/5

02:13:28 <DanC_jam> too bad timbl's not around... I gather he has plans for - and () in N3

02:14:17 <DanC_jam> "There can be no spaces within the date." -- rebol docs

02:16:07 <deltab> details of rebol's network protocols: http://www.rebol.com/docs/core23/rebolcore-13.html

02:17:57 <DanC_jam> rebol has some nifty ideas, but I find it hard to get very interested in programming language technology that's not open source

02:18:45 <deltab> yeah, that was my feeling too

02:23:41 <sbp> <deltab> letters-before-string-literals is rather uncommon, afaik

02:23:47 <sbp> tell that to RDF Core

02:24:05 <sbp> (on the xml"" literal thing)

02:25:14 <DanC_jam> don't expect RDFCore to take aesthetic suggestions about n-triples seriously.

02:26:46 <deltab> oh, the xml prefix changes the actual type of node, doesn't it?

02:27:22 * DanC_jam plays with rdfn3.g and yappstest...

02:34:37 <DanC_jam> (:age :jenny 10)

02:34:41 <DanC_jam> ^output from

02:34:51 <DanC_jam> :jenny :age 10.

02:36:24 <DanC_jam> hmm... treat 61.25 as a rational or a floating point number?

02:36:49 * DanC_jam looks up scheme syntax...

02:39:10 <DanC_jam> 6.2.4 Syntax of numerical constants http://www.swiss.ai.mit.edu/~jaffer/r5rs_8.html#SEC54

02:42:39 <DanC_jam> hmm... does python have rationals?

02:42:48 <deltab> not yet

02:43:10 <deltab> it's been suggested for some point in the future

02:43:11 <DanC_jam> ouch.

02:43:19 <deltab> the proverbial Python 3000

02:43:43 <deltab> there's probably third-party implementations

02:44:52 <DanC_jam> hmm... it has complex numbers, but not rationals. odd.

02:45:04 <DanC_jam> 2.4.6 Imaginary literals http://www.python.org/doc/current/ref/imaginary.html#tok-imagnumber

02:46:28 * DanC_jam looks for mpz stuff...

02:46:58 <DanC_jam> "This module implements the interface to part of the GNU MP library, which defines arbitrary precision integer and rational number arithmetic routines." -- http://www.python.org/doc/current/lib/module-mpz.html

02:47:13 <DanC_jam> "upport for rational numbers can be implemented in Python. For an example, see the Rat module, provided as Demos/classes/Rat.py in the Python source distribution. "

02:48:13 <DanC_jam> "An mpz-number is printed like this: mpz(value). "

02:59:13 * DanC_jam wanders off...

04:25:07 <sbp> sbp is now known as sbp`

04:50:46 <sbp`> sbp` is now known as sbp

10:54:15 <JimJibber> JimJibber is now known as JibberJim

12:05:21 <tav> tav is now known as tav`

12:32:36 <real-tim> real-tim is now known as tim-lap

12:41:45 <tim-lap> DanC, that you?

12:41:56 <tim-lap> Sorry I missed you last night

12:42:16 <tim-lap> Sorry I missed you last night-0400Z

12:43:31 <tim-lap> Sorry I hadn't checked in some of teh test files. Should be fixed now.

12:44:09 * tav offline

12:59:44 <pixel> pixel is now known as pixel-meeting

13:09:20 * DanC_jam waves to tim

13:09:27 <DanC_jam> hmm... just about time to head out to the conference

13:09:51 <danbri> Hey DanC_jam, tim-lap.... any reason not to add Cwm to http://www.w3.org/Status ?

13:10:08 <DanC_jam> no, pls do, danbri

13:10:08 * jang catches up (with some relief) on the datatyping discussions

13:10:27 <DanC_jam> as to DV signing it... that deserves fixing, but I'm not sure how

13:10:37 <jang> are people coming around to the idea of throwing more into the value space of lierals than just strings (and xml structures) at last?

13:10:41 <danbri> will do. yeah re dv too. hmm.

13:10:56 <danbri> typed strings

13:10:59 <DanC_jam> at last: I, for one, am.

13:11:16 <jang> bloody hell, it seems like danc and I are in agreement

13:11:19 * DanC_jam doesn't think of them as typed strings

13:11:21 <jang> :-)

13:11:45 <jang> not typed strings. although the distinction is conseptual

13:11:46 * danbri -> sorting thru piles of papers for missing stuff

13:11:48 <jang> conceptual even

13:12:01 <DanC_jam> I was hesitant to change RDF so much. But since we're doing it as a result of requesting community feedback, I'm much more comfortable.

13:12:21 <jang> since these theoretical RDF graphs with numbers in them (say) are still implemented with typed strings, after some fashion, sooner or later.

13:12:31 <DanC_jam> and I'm looking at mozilla code that implements it that goes back to '98. This is not a new idea.

13:12:45 <jang> of course not...

13:13:00 <jang> ... you get anyone who biulds an rdf database and the first thing they do is store non-string objects

13:13:24 <DanC_jam> yup

13:13:31 <jang> back in a tick... work calling

13:14:18 * DanC_jam -> ExML conference

14:14:16 <pixel-meeting> pixel-meeting is now known as pixel

15:53:18 <FUNDRAISING> [Global Notice] Good afternoon to all. As you may be aware, OPN is now part of a nonprofit, Peer-Directed Projects Center, and we've begun our initial fundraising drive. There's information on http://openprojects.net/news.shtml. Please click on one of the PayPal buttons on that page and help us out!

15:54:51 <FUNDRAISING> [Global Notice] We just started yesterday, and our total is at $621.90. We hope you'll keep watching these status messages, which will appear twice a day. If you want to ignore them, though, on most clients, it's /IGNORE FUNDRAISING NOTICE.

17:10:40 <Seth> could anyone tell me why we can't express quantification as triples between a variable and a formula ??? ... really this is beginning to bug me

17:10:56 <Seth>http://robustai.net/mentography/Quantification.html

17:10:56 <dc_rdfig> B: http://robustai.net/mentography/Quantification.html from Seth

17:11:47 <Seth> B:| An animated picture showing how we can express quantification as relationships between a vaiable and a formula

17:11:48 <dc_rdfig> Titled item B.

17:13:01 <Seth> B: Why, oh why, is TimBl trying to eliminate tangeable quantification triples from n3

17:13:01 <dc_rdfig> Added comment B1.

17:45:33 <olmy> seth in the interpretation there is just one object?

17:45:38 <sbp> B:because of this bit of N3: "this :forAll :s, :p, :o, :q . :forAll = log:forAll . { :s :p :o . :p = :q } log:implies { :s :q :o } ."

17:45:38 <dc_rdfig> Added comment B2.

17:45:55 <sbp> B:O.K., there's a bit of a joke in there in that it's solvable by a backwards chainer...

17:45:56 <dc_rdfig> Added comment B3.

17:46:35 <sbp> B:the the fact is that parsers should be expected to make inferences in order parse an N3 document!

17:46:35 <dc_rdfig> Added comment B4.

17:47:01 <sbp> B:the log:forAll and log:forSome predicate break rank: they are parts of the syntax that found their way into the model somehow

17:47:01 <dc_rdfig> Added comment B5.

17:47:38 <sbp> B4:the the fact is that parsers shouldn't be expected to make inferences in order to parse an N3 document!

17:47:38 <dc_rdfig> Replaced comment B4.

17:48:46 <sbp> annoying how leaving out three simple characters can completely reverse the meaning of a sentence... :-)

17:49:45 <sbp> (IIRC, there are examples where leaving out a single bit of punctuation can reverse the meaning)

17:51:14 <sbp> argh, I'm such a pedant

17:51:24 <sbp> B4:but the fact is that parsers shouldn't be expected to make inferences in order to parse an N3 document!

17:51:24 <dc_rdfig> Replaced comment B4.

17:51:38 <sbp> B5:the log:forAll and log:forSome predicates break rank: they are parts of the syntax that found their way into the model somehow

17:51:38 <dc_rdfig> Replaced comment B5.

17:52:11 <sbp> it's like my typos sit on the screen and laugh at me until I correct them...

17:52:28 <sandro> bijan: ths distinction of "making inferences" is fuzzy. All computer programs being executed are "making inferences" in some sense.

17:52:34 <sandro> B:ths distinction of "making inferences" is fuzzy. All computer programs being executed are "making inferences" in some sense.

17:52:34 <dc_rdfig> Added comment B6.

17:54:13 <sbp> B:I mean the specific case of deriving new triples from triples already present in a document, whether that be via rules from some external document, a built in rule, or something else doesn't bother me. the fact is that the meaning of an N3 document should be completely transparent to a parser: a parser should just have to parse the document, not decide things about its contents

17:54:13 <dc_rdfig> Added comment B7.

17:55:38 <sbp> B:If you have to start worrying about the meaning of a document before you can parse it, then you're stuck, because you have to parse it in order to get the meaning. that's why using @prefix as a triple doesn't work either (I've demonstrated in an upcoming article why several models for this fail)

17:55:39 <dc_rdfig> Added comment B8.

17:57:03 <sandro> B:Sorry for the confusion. I agree viewing @prefix as a triple is a lousy design, etc. I was just being pedantic about your use of "inference".

17:57:03 <dc_rdfig> Added comment B9.

17:57:43 <sbp> B:It's a good point: I am rather poor when it comes to my correct use of terminology such as "inference" (I believe that bijan has taken me up on this before)

17:57:44 <dc_rdfig> Added comment B10.

17:58:39 <sbp> B:But I hope/believe that the jist of what I said above is discernable in amidst the terminology confusion :-)

17:58:39 <dc_rdfig> Added comment B11.

17:58:47 <sandro> B:Something like n3 should have a simple and clear grammar, such that you or I or Bijan can implement correct parsers easily. No question.

17:58:47 <dc_rdfig> Added comment B12.

17:59:35 <sbp> B:As DanC is keen in noting, the fact that N3 is LL(1) makes it easier for people to implement parsers. in fact, there is some confusion as to whether it is LL(1)

17:59:35 <dc_rdfig> Added comment B13.

17:59:56 <bijan> Actually, LL grammars generally suck.

18:00:04 <bijan> But they are efficient :)

18:00:24 <sbp> heh, well, I'd like to find out more, but my apologies: I have to go to dinner etc. now

18:00:36 <sandro> B:The issue of parsing quantification gets conflated between (1) n3 making it look like log:forall is not magic syntax, and (2) people thinking n3 parses to triples when it really parses to logical formulas (which I contend can be nicely mapped to triples, but that's an aside).

18:00:36 <dc_rdfig> Added comment B14.

18:00:55 <bijan> The particular grammar is what's LL(1) or not, though I suppose if a language* can* be represented with an LL(1) grammar, it doesn't hurt to call it LL(1)

18:01:08 <bijan> basically, LL is very restrictive but very easy to make fast.

18:05:02 <bijan> Heh, CycCorp is recruiting at UNC's phil dept :)

18:35:50 <Seth> is {log:equalTo daml:equivalentTo n3:= } ? ... where is = defined in a n3 schema ? ... its not in log

18:44:34 <Seth> B: [http://ilrt.org/discovery/chatlogs/rdfig/2001-10-19.html#T21-34-43|for a previous discussion of this topic]

18:44:35 <dc_rdfig> Added comment B15.

18:50:56 <Seth> B: TimBl says "That if :x = :y in the daml:equivalentTO sense, then you can't assume from forAll x that froAll y too."

18:50:56 <dc_rdfig> Added comment B16.

18:53:04 <Seth> B: But why not? ... with {:x daml:equivalentTo :y} you just said that :x and :y denote the very same thing, so it follows that from forAll x you can assume forAll y. I don't get this :(

18:53:04 <dc_rdfig> Added comment B17.

18:54:24 <Seth> B: if {:x daml:equivalentTo :y} then you could *always* :x for :y and visa versa.

18:54:24 <dc_rdfig> Added comment B18.

18:54:57 <Seth> B18: if {:x daml:equivalentTo :y} then you could *always* substitute :x for :y and visa versa.

18:54:57 <dc_rdfig> Replaced comment B18.

18:57:25 <Seth> i guess im missing something ... could someone please please put me out of my misery

18:58:14 <Morbus> /cs op #rdfig;

18:58:19 <Morbus> hrm. didn't work. sorry.

18:58:20 <Seth> there must be some unstated assumption that you guys are holding that i am not, but what is it?

19:13:47 <Seth> take http://robustai.net/mentography/quantification_n3.gif in rebuttal or explain yourself, sir Palmer !

19:15:18 <Seth> B: in rebuttal to sbp's above [http://robustai.net/mentography/quantification_n3.gif|mentograph of this :forAll :s, :p, :o, :q . :forAll = log:forAll . { :s :p :o . :p = :q } log:implies { :s :q :o } .]

19:15:18 <dc_rdfig> Added comment B19.

19:17:54 <Seth> B: so the question remains: Why can't we express quantification as arcs between the variable and a formula ??

19:17:54 <dc_rdfig> Added comment B20.

19:27:36 <Seth> B: apparently there is a bug in CWM that barfs on this, so apparently instead of removing the bug, TimBl and company just decide to sweep the problem under some rug. I say we need to see quantification as triples, otherwise its not labeled directed graphs, but just wierdness.

19:27:36 <dc_rdfig> Added comment B21.

19:32:24 <Seth> B: KIF does a better job, at least it allows us to see the forAll: (forAll (?x ?y)

19:32:25 <Seth> (dmal:equivalentTo ?x ?y))

19:32:25 <dc_rdfig> Added comment B22.

19:33:54 <Seth> B22: B: KIF does a better job, at least it allows us to see the forAll: (forAll (?x ?y) (dmal:equivalentTo ?x ?y))

19:33:54 <dc_rdfig> Replaced comment B22.

19:35:15 <Seth> hi tim

19:36:13 <tim-lap> Hi seth

19:36:27 <Seth> as you can see i've been defaming you

19:36:50 <Seth> care to put me out of my misery ?

19:36:52 <tim-lap> I haven't been reading the logs

19:37:13 <tim-lap> Should I or would it be better not? ;-)

19:37:35 <Seth> just read the last chump, my question ... it pretty clear there, i think

19:37:50 <tim-lap> ok

19:39:28 <tim-lap> wow . must be the biggest chump in a while.

19:40:20 <Seth> well its been weighing heavy on my head since last time we talked .. and now that im getting triples from a n3 parser, im even more disssatisfied cause i dont get triples for the quantification

19:42:20 <tim-lap> I'm in a meeting now... i'll respomng to this when i can. The problem is *not* a bug -- it is logical.

19:42:48 <Seth> well what is the logical explanation?

19:45:02 <FUNDRAISING> [Global Notice] Hi all. OPN has been adopted by a Peer-Directed Projects Center, a Texas nonprofit corporation, set up for that project and related community projects. An application for US 501(c)(3) tax-exempt status is pending.

19:46:58 <FUNDRAISING> [Global Notice] We've started a fundraising drive and are currently at $628.07 in verified donations. There's information and there are paypal button on http://openprojects.net/news.shtml . If you can help, please do. Help us get through this drive as quickly as possible.

19:47:58 <FUNDRAISING> [Global Notice] We'll send these global notices twice a day. We hope you'll continue to read them and keep up with our progress, but if you feel strongly about not seeing them, your client should take: /ignore fundraising notice .... Thanks for your patience, and thank you for using OPN!

20:23:15 <tim-lap> Seth, RDF triples are assertions, and there are various axioms about them. Eg, One is that you can take away any triple and the formulais still true (if it was true).

20:23:48 <tim-lap> Consider: { this forSome Seth. Seth a Monkey}

20:24:26 <tim-lap> That says that there is at least one mokey. It says nothing about Seth.

20:24:51 <tim-lap> However, if i use the axiom that removing a triple leaves the formula just as valid, I get:

20:25:00 <Seth> what did the 'this' mean again?

20:25:02 <tim-lap> { Seth a Monkey}.

20:25:14 <tim-lap> "this" means "the formula this is in"

20:25:38 <tim-lap> "this log:forSome x" was the way i was using triple arcs to encode quantification.

20:26:20 <tim-lap> Alas, they are psudoproperties, but for the reason just shown, they can't be treated as RDF triples and in fact I am going to phase out their use and introduce asepcial syntax.

20:27:39 <Seth> ok, i think i have a solution for you, then. in { this forSome Seth. Seth a Monkey} you put the triple {this forSome Seth.} in the *same* formula as {Seth a Monkey} ... right?

20:28:11 <Seth> put it outside the formula, and then you dont run afoul of that condition.

20:29:31 <Seth> ?Seth forSome ?context. {?Seth a monkey}.

20:30:50 <tim-lap> logger, pointer?

20:30:50 <tim-lap> See http://ilrt.org/discovery/chatlogs/rdfig/2002-08-06#T20-30-50

20:31:51 <tim-lap> Seth, you can do it sure ... you can say for example { {Seth a monkey} forSome Seth}

20:32:19 <Seth> ok, and your complaint with that ?

20:32:55 <tim-lap> or {{Seth a Monkey} forSome (Seth x y z) } forAll (a b c d). Various flavours.

20:33:22 <tim-lap> And then you can reify that and you have the whole thing in triples. That's bascially what Sandro is doing with his ptl.

20:34:00 <tim-lap> So it's just the way I was doing it with flat triples in the formula itself which doesn't work.

20:34:55 <tim-lap> In practice in the software, it is very clumsy to handle them in these latter ways, though, so cwm will keep them internally in the formula's objects, and then just output whatever the flavor of theday is in a given language.

20:35:11 <tim-lap> No complaint with that.

20:35:39 <tim-lap> But you could retract the bit about the cwm bug in B!

20:35:57 <tim-lap> :)

20:36:30 <Seth> maybe im not convinced yet. thing is that other systems (sailor for example) process n3

20:36:49 <tim-lap> B: See [explanation|http://ilrt.org/discovery/chatlogs/rdfig/2002-08-06#T20-30-50]

20:36:50 <dc_rdfig> Added comment B23.

20:36:53 <Seth> and i dont get the quantification data out of the n3 parser

20:37:06 <sbp> Tim, a related question: I presume you're planning to add @for(All|Some) as substitutes. will these be taking QName/URIref tokens, or some kind of neutral variable syntax (as Sandro, I think, proposed)

20:37:07 <tim-lap> out of which parser?

20:37:16 <Seth> seans

20:37:29 <sbp> (afon)

20:37:39 <tim-lap> you have to ask sean about afon

20:38:14 <sbp> Seth: the quantification data gets passed through into sink.quant

20:38:31 <Seth> ive mentographed the way i'd like to see these ... all chumped today .. even animated a gif to show how the inference works

20:38:49 <tim-lap> Sean, I think I would like to add normal qnames etc, and seperately add ?xxx as a shortcut.

20:39:48 <tim-lap> ?xxx being something like :xxx but where it is assumed quantified in the parent context.

20:40:35 <sbp> I was wondering if your concern about quantifying URI-refs in the graph (that one can't be substituted for another) would carry over

20:41:03 <sbp> ("<timbl-lap> The difference betwen forAll and an arc in the normal sense? That if :x = :y in the daml:equivalentTO sense, then you can't assume from forAll x that froAll y too" from http://ilrt.org/discovery/chatlogs/rdfig/2001-10-19)

20:41:25 <tim-lap> Might need to quote the x in some way.

20:41:26 <Seth> ... see http://robustai.net/mentography/Quantification.html ... the quantificion is expressed between the variable and a formula or a variable fro a formula: ?seth forAll ?context

20:42:43 <Seth> dont quote the x just put the quantification triple outside the formula where it is used !!!

20:44:19 <Seth> sbp, how do i pick up the quantification from the sink.quant ?

20:44:22 <tim-lap> With @forAll we are talkiing syntax ... no need to add extra quoting, but in the reification to triples, yes, the thing can't be just <x> it has to be "x" or something.

20:44:23 <sbp> Hmm... I was wondering if the fact that @forAll :x . can appear after the triples that :x is used in is a problem, but now I don't think it is

20:44:46 <Seth> make it ?x

20:44:49 <tim-lap> I was thinking of restricting it so tha the #fors come at teh beginnijng of the formula

20:44:51 <Seth> just like KIF

20:45:02 <sbp> that's a good idea

20:45:30 <tim-lap> So ?x becomes not a uriref symbol at all but a new sort of object, a variable?

20:45:30 <Seth> (forAll (?x ?y)

20:45:30 <Seth> (dmal:equivalentTo ?x ?y))

20:45:36 <sbp> Hmm... but also before any statements in a document? I suppose that's enforcable too, but trickier

20:45:49 <Seth> tim, yes a variable.

20:45:50 <tim-lap> hmmm. true....

20:45:56 <tim-lap> (trickier)

20:46:26 <Seth> note that in the KIF, the quantification is outside the () where it is used.

20:47:17 <Seth> so if we want to translate between n3 and kif, we need those triples ... don't we?

20:47:35 * tav offline

20:47:39 <tim-lap> When you do substituition in a log:implies rule, it is deep substitution. However, when you do = substitution it is not.

20:47:55 <tim-lap> (Unless you have axioms to allow you to go deep in certian cases)

20:48:39 <Seth> well = just means daml:equivalentTo .. which says that you can *always* substitute .. i thought that was what daml:equivalentTo meant

20:49:10 <tim-lap> n3 does not have sepecial things for variables. I suppose it could. I haven't thought what the impact would be.

20:49:42 <tim-lap> If x = y and you know x colour red then you can deduce y color red.

20:50:01 <Seth> well i get a lot of tagging out of afon that sure looks like n3 knows whether a thing is a univar or a exivar

20:50:09 <tim-lap> If x = y and you know I said { x color red} you cannot deduce i said { y color red}.

20:52:00 <tim-lap> However, if @forAll x. {foo bar x } => { I said { x color red}} then you *can* deduce that I said { x color red}

20:52:10 <tim-lap> Just thought I'd mention it.

20:52:37 <Seth> wait a moment ... lets go back to If x = y and you know I said { x color red} you cannot deduce i said { y color red}.

20:52:44 <tim-lap> So /me hates the way chatzilla makes smiley faces out of =>

20:52:54 <sbp> heh, heh

20:53:12 <sbp> it's not too late to change it to "->"!

20:53:21 <tim-lap> :)

20:53:51 <tim-lap> However, if @forAll x. {foo bar x } :) { I said { x color red}} then you *can* deduce that I said { x color red}

20:53:55 <tim-lap> ?

20:54:28 <tim-lap> Maybe -- to go the other way -- all the emoticons should be n3 shortcuts.

20:54:31 <Seth> if x and y are synonyms, then there is absolutely no distinction between them as far as RDF is concerned. so it does not matter whether you said {x color red} or {y color red}

20:54:37 <tim-lap> anyway .. I have to leave now

20:54:55 * sbp too, in a minute

20:54:57 <tim-lap> oh but it does ...

20:55:06 <tim-lap> tim-lap is now known as tim-away

20:55:09 <sbp> thanks for the dialogue, guys

20:55:53 <Seth> tim (gone on now) .. oh but why does it ?

20:56:07 <sbp> Seth: if you have any problems with the quant thing, you can always email me. c'ya

23:55:21 * tav offline