Semantic Web Interest Group IRC Chat Logs for 2001-04-15

NOTICE: #rdfig logs are being turned off 2004-12-03. Please switch to the new and shiny #swig channel for Semantic Web Interest Group chat. Change your client to #swig and enjoy the new experience. Or read the latest #swig logs to see what you've been missing :)

00:29:56 <connolly_> connolly_ is now known as DanC_

00:30:23 <DanC_> chump down?

03:27:02 * bkdelong listens to the pin drop.

03:28:30 <gerald> hey bk

03:32:29 <bkdelong> Heya

03:33:42 * em waves to bk and plans to respond to mail asap

03:34:24 <bkdelong> Hehehe

03:34:27 <bkdelong> No pressure.

03:35:02 :)

03:36:52 <bkdelong> I've suddenly, er, had some encouragement to move forward. (http://www.brain-stream.com/archive/2001_04_08_archive.html#3206559)

03:39:20 * em reading...

03:41:09 :) excellent!

03:41:34 glad to hear (err.. read) things are focusing and moving forward

03:42:24 I'd be happy to help discussing SW/DC stuff more with you

03:42:46 <bkdelong> It'll be updated in about 10min when Sally' writes her On Firday's entry. Last I heard, she had one sentence left.

03:43:05 understood...

03:43:20 <bkdelong> Great! Yeah, the SW/DC stuff feads back to the "blue sky" talks we had at lunch during the plenary

03:43:26 yep :)

03:43:28 <bkdelong> s/feads/feeds

03:44:26 and all of this feeds off various discussions we've had in the past... :) it all ties together...

03:44:35 <bkdelong> Exactally.

03:45:00 <bkdelong> A few things just need a good kick in the pants.

03:45:17 * em has big feet

03:45:52 * bkdelong has steel-toed boots.

03:47:29 * em has to get back to some writing...

03:47:48 take care bk, looking forward to playing catchup/update

03:48:24 <bkdelong> k. later.

03:48:56 * DanC_ waves

03:49:38 * em pops back in to wave to danc :)

03:49:47 * DanC_ reads "The Unknowable"

03:49:58 <DanC_> I was gonna check in my write-to-palmpilot code...

03:50:12 <DanC_> ... but that requires some engineering discipline, and I'm not in that mood just now.

03:50:15 <bkdelong> Is that like KnowNowable?

03:50:49 <DanC_> I found 'the unknowable' on the scratchpad: http://www.cs.auckland.ac.nz/CDMTCS/chaitin/unknowable/

03:51:11 <DanC_> whatcha writin', em?

03:51:19 IBM research book?

03:51:37 <DanC_> yeah, I saw mention of IBM research

03:51:45 this guy: http://www.cs.umaine.edu/~chaitin/ ?

03:52:43 <DanC_> I expect so; dunno.

03:53:32 re writing: has some free time tonight... semantic web white paper / road map

03:53:32 * em does this after loading his rdf/palm db back into EOR

03:53:32 :) haven;t slept in days... :)

03:53:33 <bkdelong> Yeah, the site at Auckland is a mirror of his Maine page

03:53:48 * DanC_ is surprised that em isn't poking me to check in my POST-to-pilot code

03:53:59 * em pokes!

03:54:14 * em puts down the SW paper and pokes again

03:54:36 * em updates his copy of palmagent

03:54:58 <DanC_> hmm... I don't check in untested code, but techinically, this code is tested.

03:55:16 <DanC_> I usually try not to check stuff in with undocumented limitations, but hey...

03:55:27 noticed this... :(

03:57:18 * DanC_ reviews/summarizes changes...

03:58:27 <bkdelong> It's up, em : http://sally.editthispage.com

03:58:33 BK, have you seen palmagent? If you're interested in SW stuff, check out what DanC is doing at http://dev.w3.org/cvsweb/2001/palmagent/

03:58:53 * em checks on sally

03:58:58 <bkdelong> Cool....I'll have to commandeer Kirky's palm.

04:01:35 re editthispage: thanks for the update bk. so I'm assuming from this that you'll be sticking around in boston?

04:02:15 <bkdelong> Yeah...though I'll be looking for a new job.

04:02:32 <bkdelong> I'm hoping to do similar work to what I did with Dublin Core, Mozquito Technologies etc

04:02:59 <bkdelong> I like digging my hands in technology and making it easier for others to grok :)

04:03:34 hmm... none of that around here, eh :)

04:03:47 <bkdelong> hehehe.

04:04:43 * em thinks about this more as he signs off (again)...

04:04:48 take care bk, danc

04:05:22 <DanC_> not gonna test my code, Eric? pdkb.pl v1.6

04:05:24 <bkdelong> Later

04:06:13 * em pops back in to tect code... can't resist :)

04:08:18 * em checks out code

04:08:49 <DanC_> so it works like this, eric: first, start the server: perl pdkb.pl --serve DATEBOOK

04:09:06 <DanC_> then, put an event record in foo.rdf (I can dcc one to you if you like)

04:10:03 <DanC_> then POST foo.rdf to http://127.0.0.1:NNNN/DatebookDB/

04:10:05 * em figures this from code... cant remember where he put medusa

04:10:11 :(

04:10:25 <DanC_> medusa? what's medusa got to do with it?

04:10:32 <DanC_> if you need a POST tool, I recommend curl

04:10:41 ah...

04:11:30 * em installs curl

04:11:36 <DanC_> I don't have the exact command line, but it's something like: curl --data "`cat foo.rdf`" --header "Content-Type: application/rdf+xml" http://.../

04:11:37 done

04:13:19 perl pdkb.pl --serve ~/MyPilot/backup/DatebookDB.pdb

04:13:26 Serving palm datebook </home/em/MyPilot/backup/DatebookDB.pdb> at: <http://birch:35205/>

04:13:35 <DanC_> you're rolling!

04:13:38 ok... now, to find curl

04:13:45 got it

04:13:49 ok...

04:14:20 creating a sample meeting.rdf file

04:14:31 <DanC_> there's a trick to what you put in foo.rdf; did you see that? you have to have <rdf:Description about=""><rdf:value rdf:resource="#myevent"/></rdf:Description> in there

04:14:42 hmm...

04:14:47 can you send me a sample?

04:14:53 <DanC_> yeah... hang on...

04:15:38 <GabeW> Hi people

04:15:46 <DanC_> actually, I should check in the sample, shouldn't I?

04:16:31 yes... examples will be helpful (and do the tidy -i -xml before :)... the frame-happy syntax shouldn't be seen by humans :)

04:16:41 <DanC_> ok... checked in dbrec-test.rdf

04:16:46 yet... I seem to be able to read this...

04:16:54 checking out...

04:17:29 <DanC_> frame-happy is not an indentation thing. frame-happy just means "sorted by subject"

04:17:59 understood... but first glance is scary until you realise this

04:18:16 cvs seems slow... hmm... network prob agina?

04:18:27 <DanC_> if your post succeeds, you'll get ,put.pdb in the current directory (of the server)

04:20:23 got it...

04:20:23 looking into the exact put syntax for adding records

04:20:46 <DanC_> see the curl command above? (should go in the readme...)

04:21:45 curl --data "cat dbrec-test.rdf" --header "Content-Type: application/rdf+xml" http://birch:35205/

04:21:53 <title>403 Forbidden</title>

04:21:58 <h1>403 Forbidden</h1>

04:21:59 hmm...

04:22:18 <DanC_> put /DatebookDB/ at the end

04:22:34 <DanC_> you post to the datebook, not to the server home page

04:22:43 ah ha!

04:22:57 {:rxd "YYYY-MM-DDTHH:MM:SSZ"; :from "192.168.1.3"; :method :POST; :res </>].

04:23:03 <GabeW> Can I ask a basic question about N3 or are you guys too deep in your hacking right this minute..

04:23:03 [:rxd "YYYY-MM-DDTHH:MM:SSZ"; :from "192.168.1.3"; :method :POST; :res </DatebookDB/>].

04:23:03 @@ POST request; content: cat dbrec-test.rdf

04:23:12 syntax error at line 1, column 0, byte 0 at /usr/lib/perl5/5.005/i386-linux/XML/Parser.pm line 168

04:23:23 <DanC_> that content isn't right; you're missing some back-quotes

04:23:36 GabeW: can you hold off on the n3 question for about 2 min?

04:23:49 * GabeW will come back in 5

04:23:52 we're close to transfering brain share :)

04:24:52 hmm... not sure if i'm on a bad perl installation...

04:25:39 looks close..., but I didn;t get ,put.pdb file

04:27:05 <DanC_> do you get a bunch of RDF after ...content: this time?

04:27:48 <DanC_> i.e. in the @@ POST diagnostic?

04:28:12 hold on... the put shut the server down

04:28:25 ok... back up

04:28:35 <DanC_> yeah, the server is very touchy.

04:28:54 same error on put

04:29:00 curl --data "cat dbrec-test.rdf" --header "Content-Type: application/rdf+xml" http://birch:35216/DatebookDB/

04:29:07 Serving palm datebook </home/em/MyPilot/backup/DatebookDB.pdb> at: <http://birch:35216/>

04:29:07 [:rxd "YYYY-MM-DDTHH:MM:SSZ"; :from "192.168.1.3"; :method :POST; :res </DatebookDB/>].

04:29:07 @@ POST request; content: cat dbrec-test.rdf

04:29:11 syntax error at line 1, column 0, byte 0 at /usr/lib/perl5/5.005/i386-linux/XML/Parser.pm line 168

04:29:41 content: cat dbrec-test.rdf seems wrong... this should be rdf data no?

04:30:12 * em looks into curl command line parameters

04:30:12 <DanC_> yes, it should. your curl command line is wrong

04:30:37 <DanC_> the --data bit is touchy: --data "`cat dbrec-test.rdf`"

04:30:44 oh...

04:30:46 <DanC_> that's double-quote, back-quote, cat ...

04:30:51 restarting server...

04:31:00 perl pdkb.pl --serve ~/MyPilot/backup/DatebookDB.pdb

04:31:04 Serving palm datebook </home/em/MyPilot/backup/DatebookDB.pdb> at: <http://birch:35219/>

04:31:11 PUTing data

04:31:34 curl --data "'cat dbrec-test.rdf'" --header "Content-Type: application/rdf+xml" http://birch:35219/DatebookDB/

04:31:43 [:rxd "YYYY-MM-DDTHH:MM:SSZ"; :from "192.168.1.3"; :method :POST; :res </DatebookDB/>].

04:31:43 @@ POST request; content: 'cat dbrec-test.rdf'

04:31:43 syntax error at line 1, column 0, byte 0 at /usr/lib/perl5/5.005/i386-linux/XML/Parser.pm line 168

04:31:50 <DanC_> back-quotes.

04:32:01 hmm... still seems wrong... no instance data

04:32:10 <DanC_> copy/paste this: --data "`cat dbrec-test.rdf`"

04:32:17 <GabeW> yeah

04:32:50 * em gets his hand held one more time...

04:32:52 * DanC_ thinks eric skipped a few years of shell-quoting purgatory

04:32:53 starting server

04:33:08 * em thinks em hasn't slept

04:33:35 Serving palm datebook </home/em/MyPilot/backup/DatebookDB.pdb> at: <http://birch:35223/>

04:33:36 [:rxd "YYYY-MM-DDTHH:MM:SSZ"; :from "192.168.1.3"; :method :POST; :res </DatebookDB/>].

04:33:36 @@ POST request; content: <rdf:RDF xmlns:rdf='http://www.w3.org/1999/02/22-rdf-syntax-ns#'

04:33:38 xmlns:pd='http://www.w3.org/2000/08/palm56/datebook#'>

04:33:38 <rdf:Description rdf:about=''>

04:33:38 <rdf:value rdf:resource='#n'/>

04:33:38 </rdf:Description>

04:33:39 * DanC_ feels a bit guilty for keeping em up. but only a little bit.

04:33:40 <rdf:Description rdf:about='#n'>

04:33:42 <pd:date>2001-04-27</pd:date>

04:33:44 <rdf:type

04:33:46 rdf:resource='http://www.w3.org/2000/08/palm56/datebook#Timeless' />

04:33:48 <pd:description>testing@@</pd:description>

04:33:50 <pd:note>workie?</pd:note>

04:33:52 </rdf:Description>

04:33:54 </rdf:RDF>

04:33:56 @@ fillRecord: subject is #n

04:33:58 @@ fillRecord: date 2001 04 27

04:34:00 ah ha!!!

04:34:02 <DanC_> bingo.

04:34:02 * em feels better :)

04:34:04 <GabeW> Is this success?

04:34:23 has a ,put.pdb file :)

04:34:53 syncs his palm...

04:35:10 <DanC_> well, there you go; if you use coldsync, you can copy that ,put.pdb over your DatebookDB.pdb and sync; dunno if jpilot works that way.

04:36:03 yep! bingo :)

04:36:03 * em will sleep well tonight :)

04:36:30 this is really excellent danc... well done!

04:36:45 * DanC_ rings up his second satisfied palmagent customer

04:36:51 * GabeW feels like he's witnessing something, but he's not sure what...

04:36:55 <DanC_> thanks for alpha testing, Ericm.

04:37:18 my pleasure! :)

04:37:28 <DanC_> GabeW, we're able to POST rdf-encoded datebook records to our palmpilots

04:37:50 <GabeW> oh, thats kwite kool

04:38:15 <DanC_> i.e. we're replacing the palm desktop or outlook or whatever with any software you like that can do HTTP and RDf.

04:38:15 <GabeW> So, what rdf vocab./schema is used to do this?

04:38:34 <DanC_> it's a schema I made up based on Palm::Datebook

04:38:39 GabeW: ... you just witnessed 'hell' for the Semantic Web

04:38:40 only this time... we got 'hello'

04:39:11 <DanC_> on the TODO list is to map my datebook schema to/from vcard using rules.

04:39:15 <DanC_> RDF rules.

04:39:45 <GabeW> Very kewl.. So any app that wants to talk about appointments can use these schema (and of course extend it) and it can be imported into palmos

04:39:57 <DanC_> yup.

04:40:28 arrg.... can't update cvs..

04:40:40 <GabeW> And, if someone else comes up with a schema, because they are schema on a semantic level, the translation to Your Schema is much easier

04:40:44 * em suggests new sample event record to include <pd:description>Watson, come here please.</pd:description>

04:40:52 <GabeW> than lets say translating XML schemas back and forth.

04:41:14 <DanC_> easier: well, the proof isn't in the pudding yet; but that's the hypothesis.

04:41:24 <GabeW> You will start to understand the depth of my RDF grok level.

04:41:29 s/watson/danc

04:41:33 <GabeW> By my comments and questoins

04:42:00 <DanC_> your comments are relatively astute so far, GabeW. What's your n3 question, btw?

04:42:15 <GabeW> Astute because I know to keep my mouth shut without thinking first ;-)

04:42:19 <DanC_> feel free to put watson or whatever in dbrec-test.rdf yourself, eric

04:42:39 <DanC_> ... or to just go nighty-night.

04:42:48 <GabeW> My N3 question is well, a couple. First, is there a way to express collections (Bag, Set, Alt) in n3?

04:43:09 <DanC_> there are two ways...

04:44:13 <DanC_> first, you can write { :myBag a rdf:Bag; rdf:_1 <firstItem>; rdf:_2 <second> }

04:44:27 <GabeW> (eww)

04:44:28 DanC: cvs [server aborted]: "commit" requires write access to the repository

04:45:02 <DanC_> your CVS directory is set up read-only, remember eric? you'll have to check it out again using ssh rather than pserver

04:45:41 * em thought i did... worth checkign

04:45:53 <DanC_> other collection syntax: (<firstItem> <second>). But it's defined in terms of daml:first and daml:rest and we haven't specified the connection to rdf:_1 yet.

04:46:21 <GabeW> OK, so thats what I figured.

04:47:01 <GabeW> Second question is whether (assuming that you can) there is at least one way to express

04:47:25 <GabeW> every N3 expression (assuming in some sort of canonical form) in RDF/XML and vice versa

04:47:33 * em will deal with cvs later... night all, thanks again Danc

04:47:39 <GabeW> night em

04:47:42 <DanC_> night eric. likewise.

04:48:30 <DanC_> there are various corner cases where conversion breaks down; e.g. { <funky-pred/> <obj> } can't be written in RDF 1.0 syntax.

04:49:04 <DanC_> and I thought {} was short-hand for reification, but TimBL hasn't made up his mind about that yet.

04:49:19 <GabeW> OK, so N3 is a "in-flight" thing..

04:50:13 <DanC_> absolutely. N3 was a hair-brained idea last summer; it became a one-night hack last October. it's had maybe a few dozen man-hours put it into it by now.

04:50:49 <GabeW> pretty good for a one night hack - I mean its pretty concise and fairly precise.

04:51:18 <lilo> [GlobalNotice] Good morning, all. The news is up on http://openprojects.net/ along with project philosophy and status. Highlights: Contribution to the fund, initiating the 501c3 process, PayPal and non-US users. Fiddling with the new code and server status. Have a great week! :)

04:51:47 <DanC_> yeah... I think it started with TimBL's frustration that RDF 1.0 syntax doesn't have, at its core, the equivalent of <subj> <pred> <obj>.

04:52:17 <DanC_> i.e. no way to just write the three URIs without a namespace for the predicate or anything like that.

04:52:26 <GabeW> I'm trying to evaluate RDF (and the semantic web stuff) for the legal xml effort

04:53:04 <GabeW> There are some problems with the nature of XML for capturing all the semantics (and even structure) for different types of legal documents used in different contexts

04:54:20 <GabeW> XML schema only allow (without trickery) one real structure which has to align with semantic meaning -- which is dangerous because semantic meaning of text in legal documents

04:54:35 <DanC_> yeah... I remember you were talking about several structures imposed on one document. I think Michael Sperberg-McQueen is still reasearching that topic. RDF would certainly be a reasonably good fit.

04:54:50 <GabeW> may be in dispute or itself be of a different "ontology" depending on the context in which the document is used.

04:56:05 <GabeW> Its been hanging up the effort for probably almost 2 years, I would say. Lots of people saying "you have to use this structure because you have to support my use of the document"

04:56:40 <DanC_> encoding law has got to be frustrating. It's ultimately not formalizable; the rules come down to judgement calls (by a jury of your peers, e.g.).

04:56:46 <GabeW> and others saying "well, your structure doesn't make sense in my context, you have to use mine" -- really this has come to a head on arguing over

04:56:59 <GabeW> what a freaking paragraph should be (vs a clause, vs a sentence, etc)

04:57:13 <GabeW> "not formalizable" - exactly!

04:57:45 <GabeW> I would argue that the very markup you add to a document is itself often as much "content" as any of the words

04:58:05 <GabeW> And therefore should be treated as "first order" content itself. (e.g. signable, referenceable, etc)

04:58:14 <GabeW> And along comes RDF..

04:58:26 <DanC_> while it doesn't look feasible to formalize law, there's a lot of ways you could make doing law more efficient... you can formalize citations, make records more available, etc.

04:59:15 <GabeW> Well, you can capture a lot of useful information - the fact that a court believes paragraph 1 & 2 forms a clause is useful - the fact that a contract author believes paragraphs 1 thru 5 constitue "conditions" is useful

04:59:42 <DanC_> as a general approach, I like to "automate what people do". i.e. don't think up the perfect system; just shave a little here and there off the present system... eventually, you get a non-linear impact (ala ebay)

05:00:24 <GabeW> Yeah, excatly - I'm working for a public B2B company (Ventro) as a contractor and their B2B systems have large pockets of human "workflow"

05:01:17 <GabeW> There has actually been a lot of talk (not sure about actual deployment) of automated systems in courts and legal industry - one big drawback is not having documents in machine useable formats

05:02:27 <DanC_> There are lot of rules about how legal documents are written, but they're all in terms of paper, no?

05:03:08 <GabeW> Well, there are rules (all mostly "local" rules - ie the court you are filing in) about the layout of the page, length of filings, parts of different types of filing required, etc

05:03:28 <GabeW> But thats for court filings

05:04:07 <GabeW> We are also talking about contracts, wills, legislation, court transcripts, judicial/administrative opinions, etc

05:04:12 <DanC_> I'd be looking for some high-volume document-flow to get a critical mass of digitized content... ala the SEC EDGAR thingy; they enacted rules that every quarterly filing has to be in XML (SGML, at the time). It was a closet system for a while, but it's having massive, non-linear impact these days.

05:05:11 <GabeW> Yeah, thats certainly an area that people keep pointing to - the difference there is that the SEC had a unique ability to demand its filers to file electronically

05:05:28 <GabeW> And that the documents that needed filing were very very structured and well understood documents

05:06:16 <DanC_> wills are at the other end of the spectrum: not a large audience, long time between when they're written and when they're executed (leaving lots of time for specs to change, context to change, etc.)

05:06:23 <GabeW> Right

05:07:04 <GabeW> The legalxml effort is divided between different working groups - e.g. contracts, legislation, court filing... Each one has (in some ways) different requirements

05:07:12 <GabeW> and different 'real world' realities

05:07:46 <DanC_> so... who has a vested interest in XML wills? I can't see the business case for it.

05:07:57 <GabeW> There isn't a wills working group ;-)

05:08:01 <DanC_> ah.

05:08:15 <GabeW> Actually, there is a market for "do it yourself" lawyering software

05:08:33 <DanC_> I can see big public-interest benefit to XML legislation, but I don't see commercial motivation.

05:09:09 <GabeW> Oh, there is a huge publishing industry and I know that in some ways, they would love to get their material from primary sources already marked up

05:09:14 <DanC_> DIY laywer software: I was going to mention that. But those vendors don't benefit from standards, do they?

05:10:13 <GabeW> re: DIY lawyer software - It depends - if you integrate your software with one of these online legal referral services, for example, or one of these prepaid legal services,

05:10:21 <GabeW> you could see a business case for making everything electronic

05:10:51 <DanC_> ah... yes, online legal services would benefit. I can see that. Maybe the DIY software vendors are smart enough to realize they're really selling a service.

05:10:59 <GabeW> Really, the big motivators these days are areas affecting courts and legal publishing and contracts (esp commercial contracts where there is lots of volume)

05:11:46 <DanC_> I see.

05:12:01 <GabeW> There is a lot of money in the legal industry - money comes in when the economy is booming (IP law, finance) and when its crashing (bankruptcy, shareholder lawsuits)

05:12:46 <DanC_> but money is not exactly the mother of invetion ;-)

05:12:47 <GabeW> But I'm more interested in the fact that there is a lot of legal information that is STILL hard to get to and even harder to use that is ostensibly supposed to be easily available to the public

05:13:35 <DanC_> in fact, folks with lots of money are rarely motivated to change anything, no?

05:13:44 <DanC_> hence the term "conservative"

05:13:47 <GabeW> In the court world, they are desperate to change things because they are overwhelmed..

05:14:12 <GabeW> Law firms are big business and like any other big business, they would love to save money.

05:14:22 <GabeW> I donno

05:15:27 <GabeW> Anyway, so I'm interested now more in capturing information about legal documents in whatever format they be in rather than trying to force them into XML schema

05:15:36 <GabeW> thats not really true

05:15:40 <GabeW> What I meant to say was

05:16:29 <GabeW> I'm interested more in capturing as much info as possible about legal documents using XML, but not (at this point) expecting a clean XML schema-based solution

05:16:38 <DanC_> I see...

05:17:12 <GabeW> RDF has made me face up to the fact that XML (as we now know it) is too limited to capture all the information we need or want to

05:17:37 <GabeW> (when I say XML, I mean XML without any higher layer of abstraction like RDF)

05:17:58 <GabeW> Which is a slight problem when your effort is called LegalXML

05:18:08 <DanC_> XML Schema is in some sense more mature and less flexible than RDF. If you're trying to say "every X can contain one or more Y's followed by exactly three Z's" then you can say that in a little XML schema and get tool support pretty much for free...

05:18:39 <GabeW> Right. Well, I wasn't neccesarily referring to the W3C XML Schema Rec itself, but that category of specification

05:18:49 <DanC_> .... you could also say it in RDF, but not in any wide-known vocabulary. RDF is less constraining, but requires that you do more infrastructure yourself, at this point.

05:18:55 <GabeW> Exactly

05:20:07 <DanC_> I don't think the name "legalXML" requires to you constrain yourself. I sometimes refer to RDF as "smart XML"; sometimes as a joke, sometimes more seriously.

05:20:11 <GabeW> The thing is, the legalxml effort is probably very suited at defining the RDF vocabulary/schema, once they get beyond the RDF/XML syntax

05:20:25 <DanC_> also, I intend to show how XML Schema and RDF can get along much better presently.

05:20:50 <GabeW> That would be something I'd be **very** interested in seeing.

05:21:26 <GabeW> The other issue is that the legalxml folks tend to be a little Luddite for XMLers - very conservative with technology

05:21:57 <DanC_> the story goes like this: The XML Schema spec has a surface syntax, but a schema is actually a set of "components" represented by that surface syntax. The components have properties in -- you guess it -- a directed, labelled graph.

05:22:40 <GabeW> Right, there seems to be a parallel, but unconnected, set of relationships, at least in terms of classes and types..

05:22:50 <DanC_> and the XML Schema developers have come up with an interop-testing dump format for the component abstraction. I'm (slowly) convincing them that RDF is a natural fit for that.

05:23:24 <DanC_> It's the way I intended XML Schemas to work in the first place (check out the DCD submission from way back, which used RDF to write XML Schemas)

05:23:58 <GabeW> I could definitely understand how XML Schemas could have been based on RDF.

05:25:07 <GabeW> So, one way of integrating RDF into a legal document (assume the legal document was marked up in XHTML, for example)...

05:25:09 <DanC_> but there wasn't enough deployment of RDF to do that; there wasn't (a) enough tools showing how RDF can be useful, nor (b) enough people who thought that RDF was a useful approach.

05:25:46 <GabeW> would be to give each atomic unit of text an ID and then have a separate part of the document describe structure and semantics using the ID of the units of text

05:25:56 <DanC_> yup.

05:26:02 <GabeW> is that insane or rational?

05:26:21 <DanC_> quite rational. Quite widely practiced in the humanities, in fact...

05:26:30 <DanC_> do you know who Robin Cover is?

05:26:35 <GabeW> Oh yeah

05:27:34 <DanC_> then you know Robin Cover's affiliated with the summer institute of linguistics -- basically, a place for studying bible translation.

05:27:54 <GabeW> OK, didn't know Robin's affiliation, but OK.

05:28:03 <DanC_> One of the big uses of SGML is concondances, comparative literature, and all that.

05:28:52 <GabeW> ok..

05:29:16 <DanC_> this separate-document approach is called stand-off markup (or out-of-line markup)

05:29:27 <GabeW> Right, thats something thats come up before

05:29:39 <DanC_> Ted Nelson argues that all markup should be out-of-line (in an article that I selected for an edited collection/book on XML)

05:29:55 <GabeW> Really early on actually, and I wasn't convinced it was the right thing to do for LegalXML and now I've done a complete 180

05:32:12 <GabeW> See, this "history" would be so useful to the LegalXML crowd. They wouldn't think I was so off my rocker then.

05:33:58 <GabeW> So, is anybody using RDF on the scale that I think LegalXML could be?

05:35:05 <DanC_> the deployment of RDF in the dublin core community is pretty substantial. (I gather; ask EricM for 1st-hand data)

05:35:32 <GabeW> Oh right, of course.

05:35:35 <DanC_> the DAML deployment is large, though it's researchy-by-nature

05:35:53 <GabeW> hi tav

05:35:55 <tav> ehm

05:35:56 <tav> hi

05:35:58 <DanC_> then there's NewsML, PRISM, RSS, and all that.

05:36:08 <tav> a channel dedicated to rdf?

05:36:09 <tav> wow

05:36:10 <tav> ;p

05:36:32 <GabeW> Tav, see http://rdfig.xmlhack.com

05:36:39 <DanC_> http://www.xml.com/pub/a/w3j/s3.nelson.html

05:36:39 <dc_rdfig> A: http://www.xml.com/pub/a/w3j/s3.nelson.html from DanC_

05:37:01 <DanC_> A:|Embedded Markup Considered Harmful

05:37:01 <dc_rdfig> titled item A

05:37:09 <DanC_> A:by Ted Nelson

05:37:10 <dc_rdfig> commented item A

05:37:19 <DanC_> A:in an XML book edited by yours truely

05:37:20 <dc_rdfig> commented item A

05:37:35 <DanC_> A:Oct 1997

05:37:35 <dc_rdfig> commented item A

05:37:54 <tav> ehm

05:38:00 <DanC_> but there are more substantial works on stand-off markup; see Robin Cover's bibliography

05:38:13 <GabeW> Oh ok, great - I didn't know this was a great topic of debate...

05:39:10 <GabeW> My forays into RDF over the last few days (and also semantic web) have been really eye opening..

05:40:28 <tav> k

05:40:44 <tav> well thanks for the info

05:40:54 <tav> laters

05:50:51 <GabeW> DanC_ - in a presentation you gave to the ALA in Jan 2000, you mention the Uniform Commerical Code in giving an example of RDf

05:52:00 <GabeW> Do you know of anyone who has attemped to use the UCC's terminology as part of an RDF vocabulary specifically?

06:21:36 <DanC_> " I treasure books that emphasize ideas, not technical

06:21:36 <DanC_> details, and that are as self-contained as possible, for these I can study on my

06:21:36 <DanC_> own." -- http://www.cs.auckland.ac.nz/CDMTCS/chaitin/unknowable/ch1.html

14:20:28 <kenm> A:

14:20:47 <kenm> dc_rdfig: a:

14:20:53 <kenm> dc_rdfig, A:

15:19:30 <danbri> me waves at aaron

15:20:20 <danbri> heh, can't type properly. Just passing through anyway...

15:20:35 <AaronSw> am i supposed to be able to see: http://www.w3.org/2001/sw/RDFCore/members.html

15:20:51 <AaronSw> the acl for it seems confused

15:22:00 * danbri doesn't believe it public yet; will talk to brian after the (extended) weekend

16:28:38 <danbri> danbri is now known as mags

16:45:19 <mags> mags is now known as danbri

17:57:49 <danja> logger grep edited

17:58:20 <danja> anyone?

18:02:07 * bwm waves at danja

18:02:40 <AaronSw> danja, it's logger, grep foo

18:02:55 <danja> danc - what's the T.Nelson book? when's it published?

18:03:38 <AaronSw> err, out of print: http://www.ora.com/catalog/wjfall97/noframes.html

18:06:09 <danja> thanks Aaron - just reading the excerpt, it is a little Old Testament ;-)

18:06:36 <AaronSw> gotta love Ted

18:07:57 <danja> What's he doing now?

18:08:55 <AaronSw> ted: http://www.sfc.keio.ac.jp/~ted/

18:09:11 <danja> ta.

18:13:09 <bwm> danja: you mentioned on irc a few days ago that you'd tried Jena but had trouble getting it to work. I was wondering what the problem was.

18:20:46 <danja> the RDF API? - just checking

18:20:55 <bwm> Yup.

18:21:22 <bwm> I should declare an interest - its mine

18:21:47 <danja> aha - while your there...

18:23:22 <danja> I was putting together some bits to stream in data

18:23:29 <danja> from either file or http

18:23:45 <danja> either just dump the data somewhere

18:23:58 <danja> and/or parse it

18:24:07 <danja> made a couple of factories

18:25:35 <danja> to provide parsers/datasources for raw char, HTML, XML & RDF

18:26:03 <danja> the RDF was getting clunky using Megginson's filter

18:26:12 <danja> Jena looked a better model

18:26:24 <danja> but I couldn't get an adapter to work

18:26:30 <danja> then I broke everything...

18:26:42 <danja> and went 10 steps back...

18:26:55 <bwm> What sort of adapter? Should I go chasing a bug?

18:27:14 <danja> nah, I'm sure the fault was my end - hang on...

18:29:02 <danja> can't find anything specific

18:29:48 <bwm> k. I'm keen to learn about any bugs in it. If I don't know they're there, I can't fix em.

18:29:49 <danja> but what I was trying to do was to make 'DataParser' objects

18:30:15 <danja> k - see what you reckon -

18:30:34 <bwm> Ahh - interesting idea. So one could have multiple readers for different kinds of input.

18:30:48 <danja> DataParser interface just has setHandler & parse methods

18:31:13 <bwm> The reader and writer stuff was kinda hardcoded as an afterthought. It should be generlized so that different readers and writers can be plugged in.

18:31:53 <bwm> re methods: yes thats exactly the sort of thing thats needed.

18:32:06 <danja> I was trying to make an adapter so I could glue in your reader into that interface

18:32:10 <bwm> must make sure its on my todo list.

18:32:37 <bwm> What did you want out of it - a stream of triples?

18:33:00 <danja> essentiallly, yep

18:33:24 <bwm> Thanks for the info

18:34:15 <danja> all I was wanting to do at this point was get URLs of DMOZ sites

18:34:29 <danja> and pull off page content

18:34:36 <danja> along with metadata

18:34:36 <bwm> well off to update the issues list. its a public holiday so I get time to do some rdf stuff :)

18:34:52 <danja> have fun

18:34:56 <bwm> bye

19:14:08 * sbp laments the wandering off of William and Libby...

21:25:03 * DanC_ waves, having completed his taxes

23:58:11 Disconnected from irc.openprojects.net (ERROR :Closing Link: logger[tatooine.ilrt.bris.ac.uk] by sagan.openprojects.net (Ping timeout for logger[tatooine.ilrt.bris.ac.uk]))

23:58:11 Attempting to reconnect

The IRC chat here was automatically logged without editing and contains content written by the chat participants identified by their IRC nick. No other identity is recorded.