Semantic Web Interest Group IRC Chat Logs for 2002-05-11

00:02:14 * reagle-tu also signs and sends danc's key

00:04:04 <DanC> er... reagle, did I give you a hardcopy of my fingerprint?

00:04:44 <reagle-tu> no, but i trust the w3c file system until i confirm it f2f

00:04:55 <reagle-tu> (and dns)

00:06:07 <reagle-tu> (rather, no need to "until" i trust the w3c filesystem and/or dns)

00:06:08 <DanC> hmm. I see.

00:07:18 <AaronSw> did you visit the w3c filesystem in person?

00:07:18 <AaronSw> did danc?

00:08:23 <reagle-tu> it's in the cvs space that is served

00:08:59 <AaronSw> Ah. But the attacker is between you and the server!

00:09:28 <reagle-tu> i'm not too worried about that

00:10:22 <DanC> yes, well, the attacker would also have to intercept the IRC communication, AaronSw.

00:10:35 <reagle-tu> (though, this week ted abused <grin/> his root priveledges on tux to molest my ssh-agent running there (forwarded) to pop up an xeyes)

00:11:20 <reagle-tu> so i know longer trust root on tux as much as i did ;)

00:11:46 <reagle-tu> s/know/no/

00:12:13 <reagle-tu> (btw: i did actually confirm the fingerprint aaron gave on irc with that imported into my key ring)

00:20:02 <synack_> synack_ is now known as s|a

01:22:23 <AaronSw> slides with weird content-type: http://www.w3.org/2002/Talks/www2002-amaya/

01:35:03 <jhendler> is anyone working on a chat client for amaya?

01:35:44 <edd> i expect dean will show us one implemented in svg soon

01:35:46 <AaronSw> When you're chatting with Amaya, you're not just talking to other people, you're talking to the Web...

01:36:11 <jhendler> edd - dean does seem to be moving awfully fast...

01:36:32 <edd> jhendler: perhaps you can lure him away to MD...

01:36:45 <jhendler> I wish!

01:37:06 <AaronSw> If only he spent more time uploading and less time cd-ripping...

01:37:35 <AaronSw>http://www.w3.org/2002/Talks/www2002-amaya/

01:37:35 <dc_rdfig> A: http://www.w3.org/2002/Talks/www2002-amaya/ from AaronSw

01:38:10 <AaronSw> A:|W3C Technologies in Amaya

01:38:10 <dc_rdfig> titled item A

01:38:24 <jhendler> (is ilrt still down? if so, where does the chump/log go?)

01:38:29 <dajobe-lap_> no

01:38:40 <AaronSw> Quick, someone edit the future slides before he gets to them!

01:38:44 <dajobe-lap_> logger, here?

01:38:44 <dajobe-lap_> See http://ilrt.org/discovery/chatlogs/rdfig/2002-05-11#T01-38-44

01:38:45 <AaronSw> :-)

01:40:03 <AaronSw> A:i18n part might need some [Unicode fonts|http://www.hclrss.demon.co.uk/unicode/fonts_unix.html]

01:40:03 <dc_rdfig> added comment A1

01:42:51 * jhendler thanks dajobe - I just had a bad pointer because of the midnight change over.

02:24:46 <dajobe-lap__> dajobe-lap__ is now known as dajobe-lap

11:20:23 <sbp-> sbp- is now known as sbp

13:22:00 <RM_39> hi everybody

13:22:18 <RM_39> any1 alive and sitting in irc at such a nice day?

13:22:33 <RM_39> (assuming itt's nice everywhere :)

14:31:20 <RM_39> any1?

16:27:17 * sbp waves

17:01:05 <RM_39> hi sbp

17:01:56 <RM_39> tansaku

18:18:23 * edd readies for day's dueling with SOAPies

18:30:34 <jhendler> xxx

18:57:00 <dajobe-lap> dajobe-lap has changed the topic to: WWW2002 Developers Day - Semantic Web Track - http://rdfig.xmlhack.com/

19:04:18 <dajobe-lap> emiller gives intro

19:04:22 <dajobe-lap> welcomes people to this channel

19:04:45 <dajobe-lap> eric is excited

19:04:48 <dajobe-lap> ... as usual :)

19:05:09 <dajobe-lap> introduces speakers

19:05:13 <dajobe-lap> rob correl, adobe

19:05:16 <edd> protocols session starting

19:05:25 <dajobe-lap> emmanual pietriga, xerox

19:05:29 * edd moves commentary to #www2002

19:05:30 <dajobe-lap> mike dean, bbn

19:05:35 <dajobe-lap> libby miller, ilrt/uob

19:05:59 <dajobe-lap>http://www.daml.org/2002/05/www2002-devday/agenda

19:06:00 <dc_rdfig> B: http://www.daml.org/2002/05/www2002-devday/agenda from dajobe-lap

19:06:07 <dajobe-lap> B:|DevDay Semantic Web Track Agenda

19:06:07 <dc_rdfig> titled item B

19:07:10 <dajobe-lap> ron starts

19:07:12 <dajobe-lap> "metadata is good"

19:07:27 <dajobe-lap> sorry, Rob Corell starts

19:07:45 <dajobe-lap> B:[http://www.daml.org/2002/05/www2002-devday/www2002-devday|DAML version]

19:07:45 <dc_rdfig> added comment B1

19:08:07 <dajobe-lap> data we have is rather tenous in its relation to a dtabase

19:08:30 <dajobe-lap>http://www.linkarchy.com/XMP-Presentation.pdf

19:08:31 <dc_rdfig> C: http://www.linkarchy.com/XMP-Presentation.pdf from dajobe-lap

19:08:39 <dajobe-lap> C:|XMP - Street RDF

19:08:40 <dc_rdfig> titled item C

19:08:44 <dajobe-lap> C:Rob Corell, Adobe

19:08:46 <dc_rdfig> added comment C1

19:08:49 <dajobe-lap> C:(PDF of course :)

19:08:51 <dc_rdfig> added comment C2

19:08:58 <AaronSw> who's RM_39?

19:09:14 <dajobe-lap> customers ask us to create open, extensible, universal, intenrational, robust

19:09:23 <dajobe-lap> (which sounds like the web fundamentals to me)

19:09:49 <dajobe-lap> logger, here

19:09:50 <dajobe-lap> See http://ilrt.org/discovery/chatlogs/rdfig/2002-05-11#T19-09-49

19:09:59 <sbp> no idea

19:10:02 <dajobe-lap> C:[http://ilrt.org/discovery/chatlogs/rdfig/2002-05-11#T19-09-49|chat near here]

19:10:03 <dc_rdfig> added comment C3

19:10:30 <dajobe-lap> embeds XML but based on RDF

19:10:56 <jhendler> "Based on RDF, The W3C standard for XML metadata" - woohoo!

19:11:10 <dajobe-lap> "you can query a PDF for dublin core properties"

19:11:15 <dajobe-lap> such as dc:creator

19:11:17 <dajobe-lap> woo!

19:11:20 <danbri> :)

19:11:55 <dajobe-lap> C:see [http://www.adobe.com/products/xmp/main.html|XMP dev area] on adobe site

19:11:56 <dc_rdfig> added comment C4

19:12:18 <dajobe-lap> expose items via the common file info dialog in photoshop, acroat, ...

19:12:22 <dajobe-lap> acrobat, illustrator

19:12:29 <dajobe-lap> "we expect to put into every authoring adobe app"

19:12:58 <dajobe-lap> C4:see [http://www.adobe.com/products/xmp/|XMP dev area] on adobe site

19:12:59 <dc_rdfig> replaced comment C4

19:13:11 <dajobe-lap> "can't stress enough, is open format and source"

19:13:22 <dajobe-lap> significant industry support

19:13:31 <dajobe-lap> demonstrated interop at seybold 2001

19:13:54 <dajobe-lap> ha ha

19:13:59 <dajobe-lap> "XML = RDF -7"

19:14:05 * dajobe-lap captures that in a photo

19:14:12 <AaronSw> s/XML/XMP

19:14:23 <AaronSw> C:[mac/win SDK available|http://partners.adobe.com/asn/developer/xmp/main.html] under "open source" license, but they have [patent claims|C:[mac/win SDK available|http://partners.adobe.com/asn/developer/xmp/main.html] ]

19:14:23 <dc_rdfig> added comment C5

19:14:53 <dajobe-lap> ignore various bits of syntax, see slides

19:14:55 <AaronSw> C5:[mac/win SDK available|http://partners.adobe.com/asn/developer/xmp/main.html] under "open source" license, but they have [patent claims|http://partners.adobe.com/asn/developer/legalnotices.html]

19:14:56 <dc_rdfig> replaced comment C5

19:15:15 <AaronSw> done with goal to be easy to parse, on machines w/ little resources or lots of data

19:15:23 <AaronSw> maybe they should use ntriples :)

19:15:42 * dajobe-lap slaps AaronSw with a pointy bracket

19:16:12 <dajobe-lap> describe the XMP packet format

19:16:40 <dajobe-lap> evidence of RDF/XML angle brackets shown

19:17:07 <dajobe-lap> barriers to adoption

19:17:32 <dajobe-lap> told, want to use XMP but not RDF to which adobe responded "why?"

19:17:54 <dajobe-lap> "Did I hear you say AI?"

19:18:01 <dajobe-lap> laughter in audience

19:18:16 <dajobe-lap> scares some of our customers

19:18:31 <dajobe-lap> who are not web-centric; such as print, broadcast

19:18:36 <dajobe-lap> "You want me to parse what?"

19:18:51 <dajobe-lap> complexity when have limited resources

19:18:57 <dajobe-lap> perceived lack of tools compared to XML

19:19:06 <dajobe-lap> difficulty in validating RDF vs XML

19:19:16 <AaronSw> timbl and jjc high-five

19:19:17 <dajobe-lap> ... "compared to XML DTD"

19:19:46 <dajobe-lap> example of two RDF/XML mean the same thing; doc with 2 authors

19:19:50 <dajobe-lap> Compatibility

19:19:58 <dajobe-lap> future versions of RDF must remain compatible

19:20:06 <dajobe-lap> "there are millions of docs with XML/RDF"

19:20:10 <dajobe-lap> eric; Woo!

19:20:22 <dajobe-lap> "tens of millions coming"

19:20:36 <dajobe-lap> "don't know upper limit"

19:21:12 <dajobe-lap> The Solution

19:21:18 <dajobe-lap> simplest syntactic profile of RDF

19:21:24 <AaronSw> "The Solution?"

19:21:30 <AaronSw> ^ note question-mark

19:22:04 <dajobe-lap> say to clients; don't worry we are doing this simple thing

19:22:15 <dajobe-lap> this the request I have to this group

19:22:25 <dajobe-lap> I think; would like a simple profile

19:22:34 <dajobe-lap> timbl765: do you use bags? a: no, we do use Seq

19:22:58 <dajobe-lap> timbl - should rdfcore make something like rdf-7? a - yes

19:23:09 <dajobe-lap> .. someplace in the web that I can point to

19:24:28 * jhendler wonders what he means by "metadata subset"

19:24:30 <dajobe-lap> j carroll -have you compared XML with current wds to check changes? a - changes don't seem to impact xmp

19:24:44 <dajobe-lap> .. changes seem to be in our favor

19:25:04 <dajobe-lap> vassilis c - in what way is it possible ot extend the vocab?

19:25:31 <dajobe-lap> a - yes, ajnd you can use the same tool

19:25:44 <dajobe-lap> AaronSw: gonna ask about patents?

19:25:52 <AaronSw> no

19:26:25 <dajobe-lap> timbl - clarifying, if I add more rdf properties, can I edit these new things?

19:26:27 <dajobe-lap> a - no

19:26:32 <dajobe-lap> (not sure I captured timbl)

19:26:54 <dajobe-lap> rob - our tools allow you edit these things

19:27:14 <dajobe-lap> timbl - crate of beer for fetching new schema stuff from web via XMP

19:27:29 <danbri> ...and a shirt (ericm) and a bunch of adobe products (rob)

19:27:32 <danbri> :)

19:27:34 <dajobe-lap> ?x - use a trick to include rdf/xml in a binary file ...

19:27:47 <dajobe-lap> (lost it)

19:28:06 <dajobe-lap> a - embedding it in XML is difficult

19:28:37 <dajobe-lap> aaron sw - use ntriples

19:28:41 * dajobe-lap shakes head sadly

19:28:46 <dajobe-lap> a - "looked at what?"

19:28:59 <dajobe-lap> rob - not familiar with it

19:29:08 <dajobe-lap> ... will take it back to tech group

19:29:34 * AaronSw waves ntriples banner

19:29:34 <dajobe-lap> talk #2 - IsaViz, a Visual Environment for Browsing and Authoring, Emmanuel Pietriga,

19:29:57 <AaronSw>http://www.w3.org/2001/11/IsaViz/

19:29:57 <dc_rdfig> D: http://www.w3.org/2001/11/IsaViz/ from AaronSw

19:30:12 <AaronSw> Tim: Just speak clearly. That's what they tell me...

19:30:14 <AaronSw> [laughs]

19:31:31 <dajobe-lap> D:|IsaViz: A Visual Authoring Tool for RDF

19:31:32 <dc_rdfig> titled item D

19:31:44 <dajobe-lap> D:Emmanuel Pietriga

19:31:45 <dc_rdfig> added comment D1

19:32:11 <dajobe-lap> java app using a pile of other java apis

19:32:25 <dajobe-lap> major graphing part is based on AT&T graphviz

19:32:38 <dajobe-lap> and xvtm zoomable UI from xerox

19:33:07 <dajobe-lap> isaviz demo...

19:33:19 <grrrr> if i were building a log of ntriples scanned randomly from irc conversation, what would a single record look like?

19:33:42 <dajobe-lap> look at the rdf for this log

19:33:50 <dajobe-lap> logger, here?

19:33:50 <dajobe-lap> See http://ilrt.org/discovery/chatlogs/rdfig/2002-05-11#T19-33-50

19:34:00 <dajobe-lap> so try http://ilrt.org/discovery/chatlogs/rdfig/2002-05-11.rdf

19:34:13 <dajobe-lap> emmanual demos an M&S example

19:34:46 <dajobe-lap> (he must have a very fast PC, java going very fast)

19:35:26 * maxf thought similarly, but it's probably the Windows VM that's faster than on his unix box

19:35:44 <AaronSw> grrrr,

19:35:45 <AaronSw> <online:#T19-34-13> <http://purl.org/dc/elements/1.1/date> "2002-05-11T19:34:13Z" .

19:35:45 <AaronSw> <online:#T19-34-13> <http://purl.org/dc/elements/1.1/description> "emmanual demos an M&S example" .

19:35:45 <AaronSw> _:j685 <http://xmlns.com/foaf/0.1/nick> "dajobe-lap" .

19:35:46 <AaronSw> _:j685 <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://xmlns.com/wordnet/1.6/Person> .

19:35:46 <AaronSw> <online:#T19-34-13> <http://purl.org/dc/elements/1.1/creator> _:j685 .

19:35:47 <AaronSw> <online:#T19-34-13> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://xmlns.com/foaf/0.1/chatEvent> .

19:36:06 <AaronSw> --

19:36:23 <AaronSw> Ooh, loads a big graph

19:36:30 <grrrr> that would be ntriples?

19:36:38 <AaronSw> yes

19:37:01 <dajobe-lap> which is not a format user apps should use :)

19:37:53 <grrrr> is that n3 syntax?

19:38:03 <dajobe-lap> a semi-subset

19:38:12 <dajobe-lap> see the ntriples definiton in the rdf test cases wd

19:38:24 <dajobe-lap> which I co-edit

19:38:32 <grrrr> googling for that term now

19:38:57 <dajobe-lap> see http://www.w3.org/TR/rdf-testcases/#ntriples

19:39:28 <dajobe-lap> "NOTE: N-Triples is not an user RDF syntax - it is intended for RDF Core WG testing purposes and checking RDF applications for conformance with the specifications."

19:40:27 <danbri> the last bit suggests that it is itself conformant with the specs. I'd be happy having .nt views of .rdf files available by content negotiation

19:40:32 <danbri> ie. amongst consenting parties

19:40:54 <dajobe-lap> please, no

19:41:06 <danbri> no content negotiation?

19:41:20 <dajobe-lap> don't promote ntriples

19:41:22 <danbri> same goes for SOAP Encoding view of graphs

19:41:31 <danbri> I wasn't promoting it!

19:41:42 <grrrr> i might use jdom to build ntriples and serialize them into and out of a given arbitrary format.

19:41:44 <AaronSw> don't promote rdf/xml you mean, surely

19:41:45 <danbri> (Aaron was...)

19:42:05 <AaronSw> n-triples is awesome! it's great!

19:42:10 <AaronSw> s-triples is even better...

19:42:23 <dajobe-lap> incomplete, not internationalized, ...

19:42:28 <grrrr> ntriples don't seem to quite match jdom as easily as logger's xml

19:42:34 <AaronSw> heh, that's what jeremy said

19:42:34 <DanC> ok, I wanna play with photo metadata tools. anybody got a tool where I can say that http://dm93.org/2000/sledding2000-12-17/MVC-006X.JPG depicts my dog Bo?

19:42:48 <danbri> In that case, why do we say it's any good for checking conformance to the spec?

19:42:57 <AaronSw> DanC, a tool that writes the data or that reads it?

19:43:02 <dajobe-lap> since it isn't a user format

19:43:10 <DanC> tool that helps me write it.

19:43:15 <maxf> DanC, rdfpic

19:43:17 * dajobe-lap inserts months of rdf-core traffic about i18n

19:43:20 <AaronSw> swordfish...

19:43:30 <grrrr> Danc netpbm might offer tools to crack open the metadata tags

19:43:36 <dajobe-lap> for the logs; emmanual continuing his demo

19:43:47 <AaronSw> DanC, http://swordfish.rdfweb.org/rweb/imgmeta

19:44:10 <maxf> DanC, http://www.w3.org/TR/photo-rdf/

19:44:31 <AaronSw> photo-rdf supports depiction data?

19:44:37 <libby> rdfauthor? (in java now, but a bit rough: http://swordfish.rdfweb.org/~pldms/)

19:45:11 <DanC> does rdfauthor know about photos? I'm happy with N3 as a generic RDF authoring tool.

19:45:14 <AaronSw> I found http://swordfish.rdfweb.org/rweb/imgmeta really helpful

19:45:14 <danbri> <wordnet:Dog foaf:name="Bo"><foaf:bestfriend><foaf:Person><foaf:mbox rdf:resource="mailto:connolly@w3.org"/></foaf:Person></foaf:bestFriend><foaf:depiction rdf:resource="http://dm93.org/2000/sledding2000-12-17/MVC-006X.JPG"/></wordnet:Dog>

19:45:22 <danbri> or something like that. no good gui for it

19:45:26 <dajobe-lap> next version of isaviz - 1.1

19:45:29 <danbri> ie for the wordnet thing

19:45:45 <dajobe-lap> ... enhanced performance for large models, bug fixes

19:46:03 <danbri> jim ley's svg gadget helps if you wanted to outline the doggy part of the picture. see under svg on http://www.jibbering.com/

19:46:06 * DanC waits for rdfweb.org to answer

19:46:07 <dajobe-lap> ... "svg export bug fixed"

19:46:18 <tav`> tav` is now known as tav

19:46:26 <dajobe-lap> suggestions for improvement - directe editing using schema

19:46:38 <dajobe-lap> - stylesheets for shape and colour, alternative layout such assa a table; templates

19:47:21 <libby> damian's site on rdfweb is doen. tmp loactions: http://swordfish.rdfweb.org/~pldms/

19:47:24 <DanC> another photo for annotating: http://dm93.org/2000/0008painting/MVC-014X.JPG

19:47:57 <DanC> I wonder if I could do a quick-n-dirty Zope photo annotation UI

19:47:58 <dajobe-lap> q - when is next release? a - two weeks

19:48:29 <dajobe-lap> q - (jim hendler) todo list, lots of literature on why graphs & graphics don't scale ...

19:48:51 <dajobe-lap> ... what hooks in code to build other authoring methods, layout, style?

19:49:00 <dajobe-lap> a - is possible, we just have to do it

19:49:13 <danbri> danc, see http://www.jibbering.com/svg/AnnotateImage.1.svg?img=http%3A%2F%2Fdm93.org%2F2000%2Fsledding2000-12-17%2FMVC-006X.JPG&h=768&w=1024&title=DanC%27s+Dog+Bo&a1=Bo&w1=http%3A%2F%2Fxmlns.com%2Fwordnet%2F1.6%2FDog&a2=&w2=&a3=&w3=&a4=&w4=&nt=&operahack=.svg

19:49:19 <danbri> ..ouch, long url.

19:49:21 <danbri> works in svg plugin

19:49:21 <dajobe-lap> q - (jeremy carroll) re use of jena. We are considering giving xml qname access

19:49:40 <dajobe-lap> a - useful

19:50:00 <dajobe-lap> q (jc again) - which serialiser?

19:50:20 <dajobe-lap> a - rdf/xml or ntriples, both work fine

19:50:38 <dajobe-lap> q (?x) - fish eye view useful

19:50:57 <dajobe-lap> ... survey on graph vis in transactions on comp graphics

19:51:05 <dajobe-lap> a - FE, yes for browsing

19:51:46 <dajobe-lap> mutters about xerox patent on something related

19:52:16 <dajobe-lap> q (?y) - nice interface and shows how larg graphs can be unreadable

19:52:50 <DanC> ?y=Pasqualino Assini, NESSTAR, Ltd

19:52:59 <dajobe-lap> ... treating all properties the same, could style

19:53:02 <dajobe-lap> a - yes, on todo list

19:53:21 <dajobe-lap> q (timbl) - qname alternative is to make parser generate namespace pbinding as a hint

19:53:26 * DanC suffers galeon crash!?!?!

19:53:28 <dajobe-lap> ... (to jeremy carroll)

19:53:38 * dajobe-lap agres

19:54:18 <dajobe-lap> ... comment: we have no in-face-uris rule for user interfaces ...

19:54:50 <DanC> the jibbering...svg thingy crashes galeon repeatably

19:54:52 <dajobe-lap> ... but now in editing we have uris. In nextstep world we have icon(lost it)?

19:55:09 <dajobe-lap> ... .url files or drag resources from clipboard and use styling techniques on icons

19:55:13 <dajobe-lap> ... property labels

19:55:35 <dajobe-lap> ... rdfs:Label and subProperty of them which can be displayed. URI on mouseover

19:55:44 <DanC> hmm... rdfs:label ... is foaf:name a subproperty? sometimes I think EricM uses rdf:value as the generic name property

19:55:46 <dajobe-lap> ... good for uis to move to removing in-face-URIs

19:56:12 <dajobe-lap> talk #3 - RDFAuthor, Libby Miller presenting work by Damian Steer, Libby Miller and Dan Brickley

19:56:30 <dean_> danc, do you have the adobe svg viewer installed?

19:56:39 <danbri> (most of the work by Damian -- he wrote RDFAuthor)

19:56:40 <DanC> umm... I think so

19:56:50 <danbri> we did some serverside stuff

19:57:04 <dajobe-lap> BLURB:RDFAuthor: Enabling everyone to author RDF

19:57:04 <dc_rdfig> E: RDFAuthor: Enabling everyone to author RDF from dajobe-lap

19:57:24 <dajobe-lap> E:presented by Libby Miller, work mostly done b y Damian with bits by Libby and Dan Brickley

19:57:24 <dc_rdfig> added comment E1

19:57:29 <dajobe-lap> (please edit that E1: dan)

19:57:37 <DanC> ok, time me: I'm gonna try installing rdfauthor

19:57:49 <timbl765> (Ivan wonders what the french for andgle bracktes is)

19:57:58 <DanC> phpht. 404 at http://rdfweb.org/people/damian/2001/10/RDFAuthor/

19:58:01 <dajobe-lap> "on OSX recently ported to swing in pure java"

19:58:10 <dajobe-lap> DanC: yeah site down due to scsi disk problem

19:58:24 <dajobe-lap> libby - motivation

19:58:43 <dajobe-lap> ... getting info from people's heads to rdf for non-technies, educative role and enableigng use of rdf model

19:58:47 <maxf> timbl, no french equivalent AFAIK

19:58:59 <DanC> indeed, the RDFAuthor tutorial is a great RDF primer

19:59:14 <dajobe-lap> libby demos (prettier) rdfauthor version

19:59:14 <shellac_unstable> thanks :-)

19:59:20 <dajobe-lap> hey damian

19:59:28 <dajobe-lap> libby does demo

19:59:41 <shellac_unstable> so I hear

20:00:19 <DanC> 'educative' (tm) libby miller, 2002.

20:00:43 <dajobe-lap> E:temporary rdf author: [http://swordfish.rdfweb.org/~pldms/|Damian's home page]

20:00:43 <dc_rdfig> added comment E2

20:01:12 * shellac_unstable curses scsi

20:01:13 <DanC> hmm... no, she didn't make up that word. it just wasn't in my vocabulary.

20:01:29 <dajobe-lap> shellac_unstable: should we point people to MEG CVS?

20:01:44 <timbl765> http://rdfweb.org/people/damian/2001/10/RDFAuthor/ Error 404

20:01:45 <dc_rdfig> F: http://rdfweb.org/people/damian/2001/10/RDFAuthor/ from timbl765

20:02:00 <dajobe-lap> F:|RDF Author real home page

20:02:00 <dc_rdfig> titled item F

20:02:01 <DanC> everybody together, re visual query: ooh! aaah!

20:02:02 <dajobe-lap> F:404 just now

20:02:02 <dc_rdfig> added comment F1

20:02:03 <shellac_unstable> I wouldn't yet - maybe for the brave

20:02:31 <DanC> I'm feeling brave.

20:02:33 <dajobe-lap> demoing of very complex graph called simple-2002-... rdf

20:03:03 <jhendler> danbri - rdfauthor looks a lot like PIQ - have you hooked it up to query yet? Could then author and query w/same interface like we did for SHOE.

20:03:20 <DanC> she just demoed the query, jim

20:03:31 <dajobe-lap> cvs root is cvs.ilrt.org:/cvsroot repository meg/software/client/RDFAuthor-jfc

20:03:40 <jhendler> DanC - problem with lack of temporality on chat - I hit return before she did...

20:03:49 <dajobe-lap> that's an anon pserver login, no passwd

20:03:58 <dajobe-lap> libby - codepcition demo

20:04:03 <danbri> jim, yes, when damian announced the query functionality that i'd badgered him for, he pointed to binpiq as 'dans inspiriation'

20:04:05 <dajobe-lap> from danbri to rohit kohare

20:04:12 <dajobe-lap> khare

20:04:21 <dajobe-lap> linked from data in this conf

20:04:31 <dajobe-lap> and the data was created by this tool

20:04:57 <danbri> nodes are people, edges are photos... (rdfesque...)

20:05:06 * jhendler - danbri - don't get me wrong wasn't looking for a plug - this is very impressive - good stuff!

20:05:26 <dajobe-lap> libby shows the bookmarks feature

20:05:32 <danbri> I wasn't, but I'm always keen to keep track of the history...

20:05:48 <danbri> like most rdfapis derrive from the mozilla one...

20:06:27 <dajobe-lap> ... bookmakrs useful for repetive tasks

20:06:37 * shellac_unstable has just remembered problem earlier what prevented cvs commit

20:06:53 * shellac_unstable scurries away to commit all changes

20:06:56 <dajobe-lap> hehe

20:07:09 <dajobe-lap> OK, I've got the working CVS checkout info

20:07:12 <jhendler> This will be great for paper writing about RDF -- been looking for a nice way to generate graphs that show off DAML stuff...

20:07:21 <dajobe-lap> export CVSROOT=:pserver:anonymous@cvs.ilrt.org:/cvsroot

20:07:23 <dajobe-lap> cvs login

20:07:27 <dajobe-lap> (return no passwd)

20:07:34 <dajobe-lap> cvs checkout meg/software/client/RDFAuthor-jfc

20:07:37 <dajobe-lap> --

20:07:43 <dajobe-lap> you will need ANT, java 1.3

20:07:47 <dajobe-lap> and some messing about with classpaths

20:08:07 <shellac_unstable> IIRC just the path?

20:08:10 <dajobe-lap> libby - links to jimleys tool sof rvbisual markup

20:08:28 <DanC> " must be an absolute pathname

20:08:28 <DanC> cvs [login aborted]: Bad CVSROOT.

20:09:10 <dajobe-lap> ... q&a

20:09:45 <dajobe-lap> q (chaals) & to emmanual - svg format that you use. WHy use a different native format; why not put the rdf in the svg?

20:10:17 <dajobe-lap> a - wanted to do this; ther eis an issue

20:10:43 <dajobe-lap> DanC: works for me?! unset CVS_RSH ?

20:11:12 <dajobe-lap> q (timbl) nice to click on urls and load files on demand for properties nentioned but not loaded

20:11:24 <dajobe-lap> a - if there is a seealso link in a file to itself, it will load that graph

20:11:48 <danbri> jim, see http://www.google.com/search?hl=en&q=inspiration+rdfauthor+query

20:11:52 <dajobe-lap> q - when a thing is a resource, don't say it! everything ios a resource

20:11:59 <shellac_unstable> ahem - missing in JFC version (currently)

20:12:03 <dajobe-lap> can someone paste in libbys presentation url

20:12:16 <dajobe-lap> talk #4 Mike Dean - DAML Tools

20:12:26 <dajobe-lap>http://www.daml.org/2002/05/www2002-devday/Overview.html

20:12:26 <dc_rdfig> G: http://www.daml.org/2002/05/www2002-devday/Overview.html from dajobe-lap

20:12:31 <danbri> libby's: http://rdfweb.org/2002/05/rdfauthor/Overview.html

20:12:32 <dajobe-lap> G:|DAML Demos, Mike Dean

20:12:33 <dc_rdfig> titled item G

20:13:16 <dajobe-lap> E:[http://rdfweb.org/2002/05/rdfauthor/Overview.html|libbys presentation]

20:13:17 <dc_rdfig> added comment E3

20:13:34 <jhendler> G: some of Mike Dean's contributions to DAML and some links to some other people's stuff

20:13:34 <dc_rdfig> added comment G1

20:13:41 <dajobe-lap> E:demoed the [http://swordfish.rdfweb.org/discovery/2002/02/paths/|codepiction demo]

20:13:42 <dc_rdfig> added comment E4

20:14:06 <dajobe-lap> E:[http://rdfweb.org/2002/05/rdfauthor/Overview.html|libbys presentation - RDFAuthor: enabling everyone to author RDF]

20:14:06 <dc_rdfig> added comment E5

20:14:09 <dajobe-lap> ugh!

20:14:11 <dajobe-lap> E5:""

20:14:11 <dc_rdfig> deleted comment E5

20:14:19 <dajobe-lap> E3:[http://rdfweb.org/2002/05/rdfauthor/Overview.html|libbys presentation - RDFAuthor: enabling everyone to author RDF]

20:14:19 <dc_rdfig> replaced comment E3

20:14:56 <dajobe-lap> mike shows the daml dynamic viewer

20:15:19 <libby> jim, sorry I didn;t mention shoe

20:15:28 <libby> I panicked!

20:15:40 <DanC> er... gerald? captivate just exits when called with no args; perhaps it could emit a clue?

20:16:01 * dajobe-lap forgot to capture libby doing the talk

20:19:15 <jhendler> libby - why? You did great, don't worry.

20:19:15 <DanC> hmm... I think I'll try photo annotation with swBot

20:19:25 <jhendler> No one scribing Mike - I'll try to take it ...

20:19:29 <libby> I love mike's demo

20:19:50 <jhendler> Mike Dean - presents demo of his dynamic viewer - talking fast and mumbling - hard to follow.

20:20:19 * AaronSw works on his swBot-chump hybrid demo

20:20:29 <jhendler> Shows dynamic model can show triples as well as links - can add extra tags

20:20:34 <libby> yeah, do taht aaron

20:21:01 <jhendler> Shows DAML-map, where all the participants in a workshop were from displayed on a map.

20:21:25 <danbri> Since seeing dean's demos, I'm convinced to really learn SVG, know it like I know HTML.

20:21:32 <danbri> ie. not very well, but enough to be fluent

20:21:43 <danbri> this map stuff is so much fun...

20:22:02 <libby> Mike has a cool expandable graph tool, based on graphviz, that pulls in info dynamically

20:22:44 <libby> 'key challege is getting enough daata to bootstrap the sweb'

20:22:50 <jhendler> Mike shows links to some other tools

20:22:51 <libby> (mike's personal opinion

20:22:53 <libby> )

20:23:12 <jhendler> Shows links to databases, sucking in the daml - hyperdaml tool also shown

20:23:22 <libby> he is so right. this stuff doesn't have a great dael of point until there's loads of it

20:23:32 <libby> or at leat hard to demo the point

20:23:36 <jhendler> shows automatic scraper tools from HTML (see also the RDF scraper tool from UMCP :->)

20:23:45 <danbri> I also like Mike's (and TAP/Guha's) focus on getting concrete factual data out there, and not just generalisations (rules/ontologies).

20:24:06 <libby> right, it's fantastic

20:24:25 <libby> miked has generated millions of triples in various ways so he canb stress-test he database stuff

20:24:29 <jhendler> Mike shows grabbing of rdf data from web dynamically

20:24:46 <jhendler> Mike is building a page for instances - lots of RDF content

20:24:55 <libby> ooh, lovely

20:25:09 <jhendler> Mike showing semantic search - disambiguation of "Altair"

20:25:51 <libby> Mike has done high performance in-memory mapping rdf database, with token ised literals :))

20:26:16 <libby> the subject gateways comminty won't use our stuff because of lack of good string searching capability

20:26:17 * danbri hopes to to literal tokenizing with postgres' freetext addon

20:26:21 <danbri> yeah, quite

20:26:21 <jhendler> Mike showing daml search for words, using RDF it comes up with different contexts

20:26:45 <danbri> other approach is to do data merging in an rdfdb, then dump out derrived stuff and index w/ z39.50 tools like zebra

20:26:54 <jhendler> Mike has database search, generate subgraphs and dump to his graphic broswer -- Wow!

20:27:10 <danbri> yes, very cool :)

20:27:48 <jhendler> Mike provides links to other stuff.

20:28:06 <jhendler> Includes a DAML+OIL for application developers introduction.

20:28:12 <AaronSw> Why do DAML people call all their stuff DAML tools instead of RDF? is it for funding reasons?

20:28:39 <jhendler> DAML is RDF - but it includes extensions -- he'll mention this later

20:29:16 <danbri> a certain element of branding, but that's not such a terrible thing. It helped get some very nice tools built!

20:29:17 <jhendler> So it is like RDFS which is separated at the appropriate time -- so DAML should be called DAML (becoming OWL) - but it is like Xquery or Zpath or etc is to XML -- i.e. both

20:29:40 <AaronSw> but nobody talks about rdfs dbs

20:29:46 <jhendler> I think we need to work on this as a group - lots of good stuff in DAML/OWL which is all RDF, but w/extension for ontologies - so how do we handle this.

20:29:46 <libby> yay, dev day agenda in RDF :)

20:30:17 <libby> mike talks about the owl requirements document

20:30:24 <libby> - wants comments

20:30:36 <jhendler> yes, but they talk about XQuery as separate from XML and as part of it without problem

20:31:21 * danbri proposes http://jibbering.com/rdf/foafwhoss2.1 as semweb screensaver during coffeebreak

20:31:33 <libby> heheh

20:31:50 <AaronSw> jhendler: backchannel chat about relation between daml and rdf

20:32:00 <AaronSw> ... what's the relation?

20:32:33 <jhendler> Mike - DAML layered on RDF - allows more modeling/expressivity in the graph

20:36:15 <AaronSw> [break]

20:38:19 * DanC updates swBot code w.r.t. cwm/llyn break-up

20:39:00 <xower> What time is it over there now? (or, "How long until AaronSw's talk?") :-)

20:39:11 <AaronSw> .time gmt-10

20:39:15 <AaronSw> it's 10:39AM

20:39:24 <AaronSw> my talk is at 11:30, i think

20:39:37 <xower> Ah. Thanks.

20:40:10 <AaronSw> ok, off to get my stuff

20:47:54 <sbp> good luck

21:02:40 <AaronSw> thanks

21:03:12 <jhendler> Brian McBride presents Jena

21:03:23 * jhendler I'll scribe if folks want

21:03:45 <jhendler> Brian plays with microphone

21:04:03 <jhendler> Brian - Jena semantic web framework -

21:04:22 <jhendler>http://www.hpl.hp.com/semweb

21:04:22 <dc_rdfig> H: http://www.hpl.hp.com/semweb from jhendler

21:04:31 <jhendler> H:|Jena page

21:04:32 <dc_rdfig> titled item H

21:04:49 <jhendler> H:| Jena tutorial

21:04:49 <dc_rdfig> titled item H

21:05:05 <jhendler> H: described by Brian Mcbride in SW developers day

21:05:05 <dc_rdfig> added comment H1

21:05:26 <jhendler> Jena is a framework for constructing an integrated set of sem web tools

21:05:35 <jhendler> please help us extend/build - going open source

21:06:05 <jhendler> audience finally says "we know what RDF is"

21:06:22 * danbri declares victory: nobody in this (pretty full) room claims not to know what RDF is!

21:06:25 <DanC> well, like there was no social pressure against saying "no, I don't know what RDF is"

21:06:57 <jhendler> Brian shows overview - grows from RDF API, now readers, writers, stores, and a couple of APIs

21:07:01 <danbri> fair comment, though its just same with XML... "knowing what it is" is a bit underdetermined too.

21:07:19 <DanC> timbl, what happened to moveContext?

21:09:26 <jhendler> Core API - make it easy for application programmer

21:09:41 <jhendler> extensibility by plugin interfaves

21:09:51 <jhendler> s/interfaves/interfaces

21:11:47 <jhendler> RDF API - new inmplementation better than old thanks to Jeremy - graph ops added

21:12:15 <jhendler> HP seems suddenly interested in merging from diff companies - errr, data sources (crowd laughs)

21:12:35 * danbri wonders about foafcorp...

21:12:52 * edd had idea this morning that he'd like to have an xml.com article on FOAF

21:13:19 <danbri> like we discussed two years ago, <blush/>...

21:13:59 <danbri> that'd be really fun! I'd like to have some writeup on rdfweb.org too that wasn't outdated and gappy... not sure how the two ambitions would fit.

21:14:13 <jhendler> uses Berkely DB store, from sleepyCat software

21:14:25 <danbri> BerkeleyDB seems to be a pretty common approach to RDF storage, wonder how the various mappings of RDF into BDB compare.

21:14:35 <AaronSw> heh, bwm's slides are almost identical to the ones i deleted from my talk. phew

21:15:09 <jhendler> RDB store - can be fast and scalable

21:16:26 <jhendler> RDB compared to memory - report is in the doc section of the download

21:17:02 <DanC> "it's in the download"... why do we have to download and unpack the distribution to read the performance report? why not just point right to it?

21:17:06 <grrrr> im having a problem locating the foaf IRC namespace inclusions

21:17:32 <grrrr> the blogger seems to be referencing elements of a namespace privately updated from the published 0.1 spec

21:17:59 <AaronSw> interesting -- he found bdb faster than rdbs

21:17:59 <jhendler> gives examples of RDQL - based on squish based on rdfdb

21:18:11 <danbri> that's why we called in friend of a friend: the schema updates are circulated by word of mouth

21:18:16 <jhendler> Shows sql like query for RDF - select/where/and/using

21:18:44 <jhendler> returns a table

21:18:58 <jhendler> Joseki - an RDF server - andy seaborne - remore access to RDF DV

21:19:21 <DanC> er... do we need to transcribe the slides? they're not available by HTTP?

21:19:36 <grrrr> danbri you might add that small note to the schema that google points to

21:19:58 * jhendler if danbri can declare victory, so can I :->

21:19:58 <libby> prob w bdb is its a bit hard compared w sqll

21:20:18 <danbri> yeah, I've heard all the bdb rdf hackers swear at it copiously

21:20:21 <jhendler> DAML API - rdf meets daml+oil - simple inferencing

21:20:24 * DanC gets swBot working...

21:21:06 <jhendler> we are hoping parka-db will be easier than bdb, but we're still having some code rot problems - hope to release in a few weeks if we can figure out these solaris dependencies

21:21:43 <jhendler> Jena Users - agents/semweb; Boeing; Cerif SOAP WS; DAMLDB

21:22:29 <jhendler> Deli, Dublin Core Metadata registry, Genie (Fipa agent based sys arch)

21:22:40 * DanC starts up swBot in #rdf-bot

21:22:42 <jhendler> Isaviz, Jema, Agent search, Nesstar

21:22:56 <jhendler> Ontorama, Python-Jena, RDFAuthor, Sesame

21:23:22 <jhendler> Tamino, Triple, Tucana, W3C validator

21:23:29 <RM_39> danbri: so is there a newer version of foaf schema than 0.1?

21:23:44 <RM_39> (except for places like in your head ;)

21:23:56 <RM_39> hi every1 btw

21:23:57 <danbri> no, we decided to just add stuff to that namespace uri (and eventually to the document found there).

21:24:09 * danbri wonders who RM_39 is the foaf:nick of

21:24:25 <jhendler> proposed - tutorial, support for new rdf, thread safetym inference for new inf engines, layered composite models, RDF Schema, OWL API (reasoner, validation, integration w/Jena core)

21:24:40 <jhendler> (I add - test set and validation for OWL spec)

21:25:05 <RM_39> danbri: not sure you've heard from me

21:25:28 <danbri> well hi, anyway! sorry the foaf docs are so scrappy...

21:25:31 <RM_39> although I wrote libby some questions about Palm addressbook and RDF some time ago

21:26:02 <jhendler> opensource - BSD license, jena-dev@yahhogroups.com, jena.sourceforge.net - jena-devel@sourceforge.net --- please come help us!

21:26:16 <RM_39> anyway - regular foaf:nick of me is CaptSolo, this one was temporary

21:26:19 <RM_39> RM_39 is now known as CaptSolo

21:27:26 <CaptSolo> danbri: so http://xmlns.com/foaf/0.1 always contains up-to-date schema?

21:27:31 <jhendler> question - ?X - the kind of queries you're doing won't test the database - complicated ones will do much more

21:28:16 <jhendler> brian - your group did a comparison - we're getting different results - we should try to get together and figure this out.

21:28:23 <danbri> it's always the latest we've added to the web. sometimes we rather casually invent properties and try them out first...

21:28:57 <jhendler> ?x second question - suggestion - time came to set up working group around this issues (sic) and standardize RDF query engines.

21:29:02 <jhendler> we could explore this

21:29:14 <CaptSolo> danbri: what RDF schemas are out there that can be used to describe person and person's data?

21:29:47 <CaptSolo> one I know is FOAF - very interesting idea, although very e-mail address concentrated

21:29:54 <CaptSolo> another is vCard

21:30:00 <CaptSolo> any else?

21:30:05 <jhendler> carole - you said you'd do a DL reasoner - they're hard to build, why won't use a free open source one.

21:30:13 <libby> danc's got one

21:30:23 <jhendler> Brian hands to Jeremy - licensing issues and Java solution.

21:30:37 <DanC> I've got one what?

21:30:47 <danbri> vcardy schema?

21:30:51 <libby> yeah

21:30:55 <jhendler> GPL license is not as broad as ours

21:31:31 <jhendler> carole - fact is now going to go broader - did you ask? Carole says "ask Ian's boss" (which is me)

21:31:36 <DanC> oh... I use this vcardy thing: http://www.w3.org/2000/10/swap/pim/contact

21:32:03 <AaronSw> tim: bridging gap betw py and java is really good

21:32:07 <jhendler> Tim - encouraged, encouraged, encouraged - would be nice if someone could bridge python and java -- wasn't sure about interface separate from store

21:32:26 <jhendler> may be very hard to separate -

21:32:57 <jhendler> Jeremy - we need better way to go to query store more directly - improves efficiency

21:33:20 <CaptSolo> DanC: thanks

21:33:37 <jhendler> jeremy - we're starting to build inf engines, starting to encounter similar problems

21:34:22 <libby> timbl: qls should get results back as RDF, why not?

21:34:36 <libby> andys: kina does that already

21:34:48 <jhendler> answer - working on it - but queriers seem to prefer the bindings

21:35:17 <danbri> If you've got the original query and the bindings, you can flip perspective and turn the answers into the matched sub-graph, no?

21:35:34 <jhendler> JH thinks a tool someone should build is to take a query and a binding list and return a triple store

21:35:44 <jhendler> ErciP - both are useful

21:35:53 * danbri nods

21:36:28 <jhendler> Aaron - Plesh Project (Slides in omnigraffle)

21:36:29 <libby> ---aaron is talking----

21:36:53 <danbri> ---timbl is photographing ---

21:36:59 <jhendler> plesh project - make the grid at the user level!

21:37:13 <jhendler> scribe missed the URL

21:37:18 <jhendler> s/url/uri

21:37:36 <jhendler> distribute data, semantics, "power"

21:37:42 <libby>http://plesh.net

21:37:43 <dc_rdfig> I: http://plesh.net from libby

21:37:48 <jhendler> powerplesh - instructions task etc

21:38:05 * edd wishes for vnc feed from projecting laptop

21:38:34 <jhendler> semplesh - internet for distributing semantic interfaces

21:39:04 <jhendler> web services - problem, need to know where the server is

21:39:11 <jhendler> crawler - too centralized

21:40:16 <jhendler> semplesh - the "emergent networks" - a distributed p2p sort of approach - nodes, clients, readers and writers

21:40:25 * xower also wishes for vnc feed; as well as video feed... :-)

21:40:27 * danbri wonders if Aaron has seen the ~95/6ish work of the Harvest project

21:40:47 <jhendler> example - send a letter to a non-existent person.

21:41:04 <jhendler> Aaron - query could sit around until the rule gets bound some day

21:41:34 <jhendler> based on distributed hash tables

21:41:59 * jhendler imagines aaron's eventual grandkid getting hundreds of messages from everyone at this meeting --

21:42:14 <jhendler> distributed storage

21:42:31 <jhendler> other part - tool kit "Infogami" -- Python RDF toolkit

21:42:35 <danbri> hi synack

21:42:40 <jhendler> parse - rdf/xml, n3, ntriples

21:42:49 <jhendler> sotre - flat files, BDB

21:42:59 <jhendler> query - Ntriples w/free variable

21:43:06 <jhendler> infer N3 rules

21:43:34 <jhendler> Jim H. hopes aaron will share his slides with us.

21:43:43 <jhendler> aaron shows a demo

21:43:54 <jhendler> demos rdfbot

21:44:01 * DanC considers dueling bot...

21:44:38 <danbri> 'chumpette'

21:45:48 <jhendler> aaron wrestles with chatbot

21:46:35 <jhendler> shows chumpette - took triples from chat, and created appropriate web site.

21:47:22 <jhendler> aaron demos adding some triples that cannot do anything until n3 rules come along.

21:48:09 <jhendler> n2 rules get typed into snak (as soon as it reopens)

21:48:46 <jhendler> aaron not phased when his computer has problems - the resilience of the young :->

21:49:10 <jhendler> aaron shows using a n3 rule to change what shows up on the web - v. cool

21:50:22 <jhendler> question - what were these rules, answer - N3 pointer

21:50:34 <jhendler> q [missed]

21:50:38 <jhendler> a - I did that

21:51:09 <jhendler> q - jeremy carroll - relation to RSS

21:51:35 <jhendler> Aaron - just a demo, can do RSS and etc. this was just a little python program to show the idea

21:52:07 <jhendler> EM - can you drive an entie website this way? aaron - working on that

21:52:56 <jhendler> liam - trust question - how do you deal w/other people doing things to you and etc.

21:53:18 <jhendler> aaron - can't really do easily on irc, but for othert things - signatures and PGP and the like

21:53:36 <danbri> CaptSolo, I just added a ptr to FOAF Wiki page from http://xmlns.com/foaf/0.1

21:53:56 <jhendler> em - tells folks about rdfig and the chump, etc -- so much for our exclusivity :->

21:54:09 <DanC> seems to me IRC might actually be used to address trust issues: if 4 people I know/trust are in a channel, I'm quite likely to believe stuff said in that channel

21:54:11 <jhendler>http://just.a.example

21:54:11 <dc_rdfig> J: http://just.a.example from jhendler

21:54:20 <jhendler> J: as em is talking, we are adding

21:54:20 <dc_rdfig> added comment J1

21:54:33 <jhendler> J: please don't click on it

21:54:33 <dc_rdfig> added comment J2

21:54:45 <libby> I:|The Plesh

21:54:45 <dc_rdfig> titled item I

21:55:08 <jhendler> J:| this was a demo, but aaron didn't refresh his screen - feel free to ignore it

21:55:08 <dc_rdfig> titled item J

21:55:11 <xower> Peer Review controls trust on IRC? What if they're /away?

21:55:34 <danbri> the chanserv and nickserv bots are a step towards trustable irc...

21:55:35 <jhendler> ===========

21:55:41 <jhendler> Guha to show TAP

21:56:18 <jhendler> guha - stuff I did w/McCool

21:56:36 <jhendler> context - islands of XML from disparate web services - shows Tori Amos example

21:56:58 <jhendler> lots of unrelated pieces, user has to link them together

21:57:27 <jhendler> sort of like pre-web hypertext and rdbms today

21:57:49 <DanC> Aaron, pls ask EricM if we need to transcribe all the slides, or if he's got the presenters to give him softcopy.

21:58:06 <dajobe-lap>http://tap.stanford.edu/

21:58:06 <dc_rdfig> K: http://tap.stanford.edu/ from dajobe-lap

21:58:11 <jhendler> Goal of tap - create a coherent SW from disparate chunks

21:58:19 <dajobe-lap> K:|TAP - Guha and McCool

21:58:20 <dc_rdfig> titled item K

21:58:31 <CaptSolo> danbri: thanks - traveled the link to wiki

21:58:38 * jhendler turn web into giant realted web of semantic - wish I'd thought of that!

21:58:44 <timbl765> TAP sldies are great general SemWeb explanation slides

21:59:02 <AaronSw> DanC, k

21:59:16 <jhendler> tap - what we do -- inspired by DNS - protocols to publish/navigate

21:59:31 <jhendler> bootstrapping - chunks of data in specific places

21:59:36 <jhendler> applications - some applications

21:59:52 <AaronSw> Too bad omnigraffle doesn't support SVG

21:59:55 <em-hi> DanC, i've got many of the presenters slides... MikeD is updating this on the semweb dev day page chumped earlier today

22:00:04 <AaronSw> should i export as PNG?

22:00:23 <em-hi> for the ppt's that aren't online yet, i'll provide a full page that links the rest

22:00:29 <danbri> aaron, that'd be good (other things'd be better, but something better than nothing :)

22:00:53 * em-hi wonders how many rdfig's are in the semweb dd audience now ....

22:01:20 <jhendler> TAP aiming at simple protocols - example getData

22:02:05 <jhendler> contract - you expose your data by allowing GetData to hook to your backend stuff

22:02:53 <jhendler> key tech issues - cahcing, directories, names - going to talk about names

22:03:22 <jhendler> each DB has an internal name for people, places, things, etc.

22:03:31 <jhendler> cannot have a standard for this.

22:04:13 <jhendler> names are crucial in info exchange. no URN for "<newton, NC>" or "<tori amos>" - too many real names in the world

22:04:36 <jhendler> traditional approach - name mapping tables, won't scale - needs calling program with name maps

22:05:20 <jhendler> that mapping won't scale to the millions or zillions of things you need

22:05:50 <jhendler> TAP naming - reference by description - "Musician whose firstname is "tori" and lastname is "amos""

22:06:01 <jhendler> "description-based name negotiation"

22:06:39 <jhendler> If you can come up with - a starting vocabulary, a description lanaguge (standard) and a protocol

22:07:22 <jhendler> jhendler thinks this is a great intro to "ontology" as part of sem web, even if he doesn't use the word

22:07:57 <jhendler> idea - expose your data as a graph - map incoming descripotions to nodes in your graph

22:08:19 <jhendler> loose coupling - needs a coupling mechanism

22:08:34 <CaptSolo> if foaf:mbox is unambiguos property, does that mean that every foaf instance should have a mbox?

22:09:35 <jhendler> protocol needs to allow "negotiation" for handling ambiguity, failure, multiple answers, etc.

22:09:55 <libby> CaptSolo, no, means that every foaf mbox value has exactly one matching starting node

22:10:04 <libby> (I thnk)

22:10:11 <jhendler> kernel vocabulary - music vocabulary - it's a "specification"

22:10:56 <jhendler> jh - someone should have a WG for how to write these vocabularies in a machine readable way

22:11:11 <jhendler> guha - semantic search - add semantics to search

22:11:24 <jhendler> hard to use crawl, grab, index model for dynamic sites

22:11:35 <jhendler> use semantic based augmentation to do better

22:12:03 <AaronSw> I:[slides|http://plesh.net/2002/www/] are now up in [OmniGraffle|http://www.omnigroup.com/applications/omnigraffle/] and PNG

22:12:03 <dc_rdfig> added comment I1

22:12:04 <jhendler> KB used to extend query (jh - glad there's no AI in here :->)

22:12:40 <jhendler> shows demo on Yo Yo Ma

22:13:47 <jhendler> idea - since these guys on right hand side (advertiseds) show up on the search - then the advertisers are motivated to make their stuff available

22:14:07 <jhendler> jh - wonders where all these KBs come from - need a distributed way to design, build, etc.

22:14:42 <jhendler> TAP KBs - write a large KB, scrape important sites, use to bootstrap sem web

22:14:49 <CaptSolo> liby: ok, but in order to uniquely identify a person (=record) in rdf database, we need that person to have at least one unambiguous property?

22:15:00 <libby> right, yep

22:15:13 <jhendler> scribal error - couldn't hear his example - something about some good place to get info on something

22:15:22 <jhendler> simple noun phrase analyzer

22:15:37 <jhendler> uses various low end bootstrapping mechanisms

22:16:05 <jhendler> KB coverage music, movies, authors, sports, autos, companies, appliances, toys, baby products, places, electronics, health

22:16:06 <em-hi> TAP includes some capabilities of listening to various feeds, doing some simple noun phrase analysis and helps build the knowledge base

22:16:16 <CaptSolo> then what to do in case if a person has no mailbox?

22:16:34 <jhendler> semantic site search - example - w3c page

22:17:14 <jhendler> special purpose KB written for a specific site

22:17:27 <jhendler> pulls in the realted links

22:18:01 <dajobe-lap> start at http://tap.stanford.edu/w3c.html and search for: eric miller in the right hand search box

22:18:05 <jhendler> Guha has idea of different people linking, copying using

22:18:14 <jhendler> jh - I call that exactly the ontology stuff

22:18:41 <grrrr> CaptSolo perhaps subclass mailbox from URI and at least offer/infer to folks a URI as thier unique ID

22:18:51 <AaronSw> so cute: http://tap.stanford.edu/cgi-bin/w3csearch.pl?q=eric+miller&sitesearch=w3.org

22:18:58 <libby> heheh

22:19:19 <AaronSw> heh! " Eric Miller is the Activity Lead for the W3C World Wide Web Consortiums Semantic Web Initiative. Erics responsibilities include the architectural and technical leadership in the design and evolution of S"

22:19:23 <AaronSw> the evolution of S, eh?

22:19:27 * em-hi will get guha for this... :)

22:19:47 * AaronSw notes that alpiri has no data on Guha

22:19:50 * em-hi thinks maybe AaronSw as well....

22:20:02 <AaronSw> hey!

22:20:03 <danbri> rdfweb does...

22:20:09 <em-hi> :)

22:20:28 <sbp> Aaron: was the crash IRC, bot, or API related?

22:20:42 <timbl765> I thought the idea was it shouldn't really matter who does ....

22:20:46 <timbl765> .. have the data

22:20:47 <AaronSw> my irc client crashed, sbp

22:21:06 <xower> Switch to AthenaIRC... :-)

22:21:21 * sbp goes through the slides...

22:21:32 <AaronSw> xower, fate would have found a way to get me another way...

22:21:33 * dajobe-lap notes source code is NOT available, or at least not that I can find it

22:21:48 <AaronSw> ask during the qs

22:21:51 <xower> It always does... :-)

22:22:01 <dajobe-lap> ah, he says code not available

22:22:04 <jhendler> jh tries TAP on lots of w3 and non-w3 people who've appeared in w3 archives. Latter does nothing unless someone scrapes those archives -- this could be a great demo - and we have a scraper that could do it -- em, we should make this so...

22:22:05 <MarkB_HI> MarkB_HI is now known as df

22:22:36 <jhendler> guha - want this to be used all over the place - need applications, lots of research as well - please join

22:22:56 <jhendler> Guha calls "semantic search" - JH thinks that a terrible name (SHOE "semantic search" page scared people away)

22:23:27 <jhendler> someone please scribe

22:23:53 <AaronSw> Q: how does this relate to libraries and publishers who have tons of data

22:23:54 <dc_rdfig> Label Q not found.

22:24:20 <AaronSw> ... you should plug in to that kind of work

22:24:54 <CaptSolo> are there ontologies that describe person from "official information" point of view - as somebody who has a passport (number), etc.?

22:25:05 <CaptSolo> like national registry?

22:25:08 <AaronSw> A: we tried to create a KB free of all IP constraints. took the hard way so we could give the data away

22:25:21 <AaronSw> ... but now we can start in on others

22:25:57 <AaronSw> jhendler: [plug for ontologies written in standard format]

22:26:07 <AaronSw> ... standardize content expression, not content

22:26:13 <AaronSw> A: yes. yes.

22:26:20 <AaronSw> ... really large ontology in our DB

22:27:25 <AaronSw> Q: find a better name than semantic search

22:27:32 <AaronSw> s/Q/jhendler/

22:27:40 <AaronSw> eric, be glad you're not rdf:type http://tap.stanford.edu/cgi-bin/kb.pl?node=Terrorist&op=show&syn=10&browse=all

22:28:24 * DanC concludes, after playing with swBot in #rdf-bot, that swBot isn't a cost-effective image annotation tool

22:28:44 <jhendler> JH - how about "semantic web portals"

22:29:01 <jhendler> (with no hyphens, to leave the noun phrase ambiguous)

22:29:04 <AaronSw> I found swBot more usable when i plugged in n3

22:29:06 <AaronSw> err n3s

22:29:15 * em-hi likes semantic web portals

22:29:35 <DanC> n3s?

22:29:54 <AaronSw> the AN3->N3 converter

22:30:01 <AaronSw> code at http://infomesh.net/2002/n3s/

22:30:13 <AaronSw> import n3s; n3 = n3s.process(input)

22:32:39 <xower> AaronSw: What happened to Slide 10?

22:33:08 <jhendler> tim - it's not really the N^2 thing - because for legacy systems, we could have third parties provide maps -- newcomers link to somewhere (jim H calles this the "united nations" phenonmenon)

22:33:11 <AaronSw> i took it out :/

22:33:22 <AaronSw> bwm had a nearly identical one

22:33:31 <AaronSw> it simply described the rdf->bdb/dht mapping

22:33:32 <xower> Ah.

22:34:26 <jhendler> guha - at many places we scrape things

22:34:49 <jhendler> there are places where they have vocabularies

22:35:50 <jhendler> Tim - we should allow a dscouple between documents and data (right hand - docs, left hand data)

22:36:15 * jhendler semantic web data portals (any other buzzwords we need to add?)

22:36:20 <DanC> tim was suggesting "data search" instead of "semantic search"

22:36:35 <libby> 'portals' is a bit overused...

22:37:08 <DanC> I don't think timbl said anything about 'decouple'; he was just making a marketing/terminology suggestion: that folks are familiar with 'data'.

22:37:46 <jhendler> jh doesn't like "data search" - the DB guys claim to have that already (Xquery infrastructure) and it sux.

22:38:43 <AaronSw> i think we should rename semweb to dataweb

22:39:32 * AaronSw -> lunch

22:40:46 <shellac> E: Oops - broke import in JFC RDFAuthor just before uploading

22:40:46 <dc_rdfig> added comment E5

22:41:03 <shellac> E: Works now

22:41:03 <dc_rdfig> added comment E6

22:41:43 * shellac was a fool to mess with the forces ranged against this demo

22:42:10 * danbri grins

22:42:16 <danbri> was a good demo though

22:42:48 <shellac> Any piccies?

22:43:20 <danbri> dunno, not yet.

22:43:49 <shellac> I just meant were people taking any - though little geeks can move quickly

22:43:58 <shellac> why you not at lunch?

22:44:03 * dajobe-lap has pics

22:44:08 <shellac> cool

22:44:12 <dajobe-lap> although I missed libby

22:44:33 <dajobe-lap> lunch

22:44:34 <shellac> see - weird forces I tells you :-)

22:49:48 <CaptSolo> libby, danbri: can you answer about ontology for describing persons which have passport number, birth certificate, etc. official propreties?

22:50:29 <CaptSolo> probably one can take FOAF or vCard (or other) as a basis and create schema to describe those additional properties missing in chosn schema?

22:50:31 <grrrr> CaptSolo those are a bit heavywieght for a research concept

22:50:43 <shellac> CaptSolo: they're at lunch, I think

22:51:20 <shellac> CaptSolo: Back in maybe 1, 1 1/2 hours

22:51:28 <CaptSolo> shell: yep, lunch is on of best passtime, better than sitting in irc

22:51:44 <CaptSolo> shellac: the question is open to everyone, just if i get the answer

22:52:36 <CaptSolo> in 1 hour it'll be 3am... which could be just the time to go to sleep :)

22:52:48 <shellac> I don't know of any such ontologies, but I don't know much

22:53:12 <CaptSolo> but if they com back and anwer, it'd be well anyway - will check logs l8r

22:53:45 <shellac> Ouch. I though I was out of sync with hawaii here in the uk

22:54:31 <CaptSolo> shell: in uk you are 2 hours back in time

22:54:37 <CaptSolo> :>

22:54:41 <shellac> no were not!

22:54:54 <shellac> grrr :-)

22:55:06 <grrrr> ?

22:55:17 <shellac> sorry grrr

22:55:26 <CaptSolo> shell: yeah, yeah - you're right on time and i'm 2 hours ahead - are you saying that? :>

22:55:34 <shellac> meant as sound effect

22:55:45 <shellac> what is it like in the future?

22:55:50 <CaptSolo> shell: tiger do emit such sound effects, i think

22:56:20 <CaptSolo> well - not bad, just a bit sleepy...

22:56:25 <shellac> On the ontology question, there's nothing stopping you making up you own schema

22:56:27 * CaptSolo looks out of the window... dark...

22:56:30 <CaptSolo> dark future :)

22:56:34 <shellac> :-)

22:57:05 <CaptSolo> shell: i know, but i'd better explore existing work not to duplicate...

22:57:15 <shellac> you can always tie it in to other schemas later, if they look more official

22:57:21 <CaptSolo> there is not sense having many very similar and duplicating schemas

22:58:04 <CaptSolo> and i think that somebody should have thought about schema to describe i.e. registry of citizens before me

22:58:14 <shellac> actually it often doesn't matter - if your x is the same as someone else's y

22:58:35 <shellac> then you can say x is y (essentially)

22:58:46 <CaptSolo> grrr: what do you mean they are heavyweight?

22:59:58 <CaptSolo> i am exploring modeling of person's data using RDF

23:00:10 <CaptSolo> that's why I have questions like the one above

23:00:11 * AaronSw waves from "Talking to TimBL"

23:00:13 <AaronSw> TimBL: documents are what messages leave behind

23:00:28 <grrrr> CaptSolo that's really a relative pursuit in most instances

23:01:13 <AaronSw> "I got this job because i could ramble on incoherently faster than anyone else... except for Rohit"

23:01:16 <AaronSw> - TimBL

23:01:29 <CaptSolo> AS: :)

23:01:30 <dc_rdfig> Label AS not found.

23:02:14 <niq> AaronSw after TimBL's job?

23:02:23 <AaronSw> heh

23:02:33 <AaronSw> what made the web work? minimal design, the people who got it (like the people at this conference... at least some of them), simplicity underneath, decentralization

23:02:43 <AaronSw> DanC: DNS helped a lot...

23:02:47 <CaptSolo> grrr: well, persons data take many shapes and actually different kinds can have very little in common except for nam

23:02:49 <CaptSolo> name

23:03:02 <niq> DNS much older than WWW

23:03:27 <AaronSw> Q: why do xml and rdf express the same thing? [boos]

23:03:27 <AaronSw> Tim: Bzzt! Syntax error line one. Please ask a well-formed question.

23:04:10 <niq> afs: I wouldn't dare, in this forum

23:04:14 <grrrr> CaptSolo URI can provide that, if offered

23:04:19 <AaronSw> Tim: RDF addresses real and abstract things. XML talks about types of elements, not people in cars. XML SCHEMA DOES NOT TELL YOU ABOUT CARS. It talks about a car registration form, its dirivng license and other docs about a car

23:04:34 <niq> (bugger, A: ==> afs: due to stupid software)

23:04:52 <AaronSw> ... it tells you that this document has one car plate field. rdf says that a _CAR_ has only one licnese plate

23:05:56 <AaronSw> Q: your opinion on rand vs. rf [laughs]

23:05:56 <AaronSw> Tim: current draft has RAND and RF almost exactly the same. they're both fine as long as you constrain rand to have afee of zero

23:06:04 <AaronSw> Rohit: "you're one block away from the firing line"

23:06:13 <AaronSw> tim ducks behind a wall

23:06:20 <AaronSw> Tim: i gave some bits of keynote about that.

23:07:02 <CaptSolo> grrr: i am still baginner in sw field. how would you form URI for person?

23:07:14 <CaptSolo> if it does not have mbox, for example...

23:07:35 <CaptSolo> and if persons can be of different countries...

23:07:59 <AaronSw> Americans go back to a house on the prarie where railroads push their ways across the land. people staked claims showing that they lived off the land, etc. all patent claims have to do is sketch the land and you're done. worse, somkeone can make claims that contain their claims and claims inside other people's claims.

23:08:10 <AaronSw> ... railroad tracks have to be a straight line

23:08:12 <grrrr> <protocol>://[[pass:user@]hostname[:port]/[filespec[#elem][?args...]]

23:08:52 <niq> What is AaronSw on? Can I have some?

23:09:10 <AaronSw> i'm just scribing what timbl says

23:09:12 * sbp thought it was a kick-ass analogy, personally...

23:09:33 <AaronSw> Q: why not more virtual web conferences?

23:09:33 <AaronSw> A: raise your hands if you think this could only be done f2f [lots of hands]. we'll ask the folks on irc later.

23:10:00 <niq> twaddle. b****r f2f

23:10:48 <sbp> if f2f were the only way of holding conferences, people like me would be permanently stuffed, and many more would be unable to attend events as regularly as they may like

23:11:33 <niq> anyone who believes f2f still matters hasn't grasped the 'net

23:11:34 <sbp> but OTOH, virtual conferences tend to break down, and it's always fun too meet the people that you've been working with remotely

23:12:04 <sbp> but yeah, in general I agree with niq

23:12:15 <CaptSolo> grrr: thacknically yes, but what [hostname]/filespec#elem to choose?

23:12:36 <niq> sbp, I agree "it's always fun", but it's a luxury, not a necessity

23:12:37 <AaronSw> Rohit: has Wifi improved the conference? [lots of hands] put them down if youi're a speaker

23:13:00 <AaronSw> Tim: who thought it made it worse [jim puts hand up] jim?! you were on irc the whole conference! JIm: yeah, and i didn't get anything done!

23:13:05 <grrrr> CaptSolo URI's don't imply a working URL, just the format

23:13:10 <sbp> lol

23:14:04 <AaronSw> [they beat up tbray]

23:14:05 <CaptSolo> :))))

23:14:05 <niq> (sbp, will you be at #er f2f in June?)

23:14:22 <AaronSw> Tbray: is this a valuable con?

23:14:22 <AaronSw> Rohit: it's and important and valuable conference and you left us for vacation?!

23:14:28 <sbp> niq: very doubtful. there's probably less than a chance in a thousand

23:14:35 <niq> :-(

23:14:40 <CaptSolo> grrrr: ok ok

23:14:57 <CaptSolo> actually to go further about persons data:

23:15:02 <AaronSw> DanC: i did a lot of good catching up. i caught buzz on irc that globalization was the more interesting part

23:15:20 <CaptSolo> is it a crazy thought to describe a CV using in RDF [database]?

23:15:26 <AaronSw> Rohit: how many people are attending the semweb sessions? [1/2 of hands] No eric, you only count once.

23:16:00 <CaptSolo> i see a very good use of such decription of cv in RDF (or uing web services, maybe not limiting to RD)

23:16:25 <grrrr> CaptSolo I think it is one of the founding scenario premises in the DAML overview

23:16:52 <CaptSolo> exchange of cv data between companies and creating distributed cv databases where every1 has to publish only his cv

23:18:04 <CaptSolo> grrr: can you elaborate? /me does not know DAML in details no it's possibilities

23:18:16 <CaptSolo> although it's on the list to be explored in detail

23:19:18 <grrrr> DAML is the DARPA brand name for RDF. it makes an RDF repository available to agents which enter queries

23:20:03 <grrrr> Im not grounded firmly to say much more than that

23:20:23 <CaptSolo> grrr: As I understant DAML extends RDF a bit.

23:20:36 <CaptSolo> (if someone could clarify)

23:20:43 <grrrr> RDF was "chosen" i think

23:20:50 <deltab> yes, that's my understanding too

23:20:56 <sbp> DAML is a set of terms that enable people to make ontologies

23:21:10 <sbp> so, it's not an extension of the RDF syntax or model per se...

23:21:17 <sbp> but it is an extension of the basic RDF vocabulary

23:21:39 <Eliezer> Hi, I'm looking for a tool which I'm almost certain already exists, and I'm poking around in IRC channels where there might be someone who knows about it. I'm looking for a simple structured long document editing tool that lets me (a) edit complex network structures, where nodes can appear in more than one hierarchy and there can be links between nodes, and which (b) lets me word process extended documents (preferably with styling) within

23:21:49 <AaronSw> Rohit: the IETF has an answer on how to accelerate the standards process. print it out and drop it, you'll get 98.5m^2

23:22:01 <grrrr> Eliezer TIBCO TurboXML

23:22:41 <Eliezer> looking it up now...

23:22:49 <AaronSw> Q: what are the worst web barnacles?

23:22:49 <AaronSw> A: drop off at the conference desk a thing with the outside saying barnacle and in the inside your favorite barnacle. [repeats about the simplicity/debarnacle wg]

23:22:52 <shellac_unstable> hasn't gravity increased a bit?

23:23:29 <AaronSw> Rohit: tell them where the // comes from?

23:23:40 <AaronSw> A: from the apollo filesystem

23:23:48 <AaronSw> anyone got a question for tim? rohit's beating me up

23:23:58 <CaptSolo> grrr: ok ... RDF/DAML should make possible scenario with making CV described in RDF available to agents

23:24:10 <AaronSw> the // was his example of a url barnacle

23:24:24 <CaptSolo> but describing CV in RDF <- that's what seems complex to me

23:24:45 <CaptSolo> it's not just record about person, it's quite complex set of relations

23:24:53 <CaptSolo> can be described but quite a mess

23:24:56 <AaronSw> anyone got a question for tim? rohit's beating me up

23:25:01 <AaronSw> sbp? niq? shellac?

23:25:03 <grrrr> CaptSolo would require some hand-holding either to enter or retag portions of a cv

23:25:46 <shellac_unstable> CaptSolo: well, it could get hairy

23:25:48 <sbp> erm... what next for SemWeb APIs? we've seen the crypto stuff, and the RDF diff rant. What's he going to pull out of the hat next?

23:25:53 <CaptSolo> hand-holding?

23:26:11 <shellac_unstable> CaptSolo: but RDF can make like easier

23:26:49 <CaptSolo> grrr - if entering a new cv in a web form (i.e.), it could be transformed directly to RDF.

23:26:57 <grrrr> CaptSolo Resumes from Monster.com or dice.com are entered by forms, plain old documents are effectively random by comparision

23:27:01 <shellac_unstable> CaptSolo: eg you could just put 'attended school with homepage: <URL>'

23:27:14 <CaptSolo> IF all the information in CV can be described in RDF

23:27:35 <grrrr> takes me a few minutes to cut & paste my resume into a form

23:27:59 <shellac_unstable> and any other info about that that institution (not added by you) is merged 'free'

23:28:16 <CaptSolo> grrr: more than a few minutes if the web form is too "sophisticated".

23:28:30 <CaptSolo> did a couple of entries last week...

23:28:40 <CaptSolo> takes .5 - 2 hours

23:29:12 <grrrr> CaptSolo would be nice to do it in one place, per one standard, yes.

23:29:21 * CaptSolo will be looking for a job agter completing my master's paper

23:29:43 <CaptSolo> the current employer is doing downsizng because our project was canceled :(

23:29:49 <grrrr> how is a school tagged/described, how is a curiculum, job, duty, etc. quantified?

23:30:18 <CaptSolo> grrr: exactly, one standard could help a lot.

23:30:31 <CaptSolo> but - maybe it is more easily done in XML?

23:30:55 <CaptSolo> it's stupid, but you don;t have to worry about URIs, etc..

23:31:09 <grrrr> namespaces are URI's

23:31:14 <CaptSolo> grrr: you are asking precizely the same questions that bother me

23:31:55 <CaptSolo> grr: I mean in XML you just define namespace for data, describe them in plain text and that's it

23:31:57 <sbp> the problem with XML is that it doesn't necessarily say anything about the data structure

23:31:58 <deltab> I think there are some directories of ontologies

23:31:59 <grrrr> the flip side of mechanial entry processes is NLP, finding assertions in text that are rarely mistaken expression.

23:32:15 <Eliezer> Tibco TurboXML looks like it will let me edit arbitrary network structures, but it doesn't seem designed to let me word process documents attached to a given node.

23:32:29 <CaptSolo> deltab: can you give a pointer?

23:32:42 <CaptSolo> delt: any ontologies ofr info like a CV?

23:32:43 <grrrr> Eliezer xmlns:xhtml does

23:32:44 <AaronSw> daml.org has a bunch of ontologies

23:33:02 <deltab>http://www.daml.org/ontologies/

23:33:02 <dc_rdfig> L: http://www.daml.org/ontologies/ from deltab

23:33:07 <CaptSolo> how to tag school / company?

23:33:30 <deltab> L:|DAML Ontology Library

23:33:31 <dc_rdfig> titled item L

23:33:37 <CaptSolo> not all companies / schools (majority of them here) have home pages

23:33:49 <Eliezer> xmlns:xhtml will let me embed documents in XML, but does Tibco TurboXML actually have an HTML editing pane inside the IDE?

23:34:15 <CaptSolo> they could be either tagged using some nation-/world-wide registry URI reference

23:34:17 <grrrr> Eliezer worst:case, use MS-word to export XML/html and paste it in

23:35:10 <CaptSolo> or every user of this CV ontology will have to invent his own tagging, which negates usefullness of this ontology if data can not be portable

23:35:30 <Eliezer> The problem I'm trying to solve is that I'm writing an extended design document and I'm trying not to spaghetti while writing it. Outlining tools aren't enough; I need network editing tools. It's not a problem where I'm trying to single-source the document; the problem is writing it to begin with.

23:35:38 <AaronSw> Q: why is rdf better than xml which is more extensible and explicit and can use urls? why is rdf model so limit for knowledge sharing, etc.? what's the point?

23:36:12 <AaronSw> A: it uses xml because a large number of people use xml. you can make a better language that doesn't. it won't do a lot of damage as long as they conver tto the same triples, you're interoperable.

23:36:13 <CaptSolo> shellac: if schoolhas no home page, such tagging won't work

23:36:16 <grrrr> Eliezer Together/J

23:36:25 <Eliezer> I need to be able to jump around in the network, and view nodes and relations between them, while I'm writing the design document.

23:36:30 <Eliezer> checking it out now...

23:37:19 <shellac_unstable> CaptSolo: well, that was just an example, any identifier would do

23:37:29 <AaronSw> ... the rdf model is "limited" like a link is "limited". you can use them to express a verty large amount of stuff. a lot of stuff is data, not rules, not math expressions. all that data in xml is in fact expressible as triples. you can map it into binary relations. and that's very powerful, you map all this stuff into rdf and you end up with lots of triples you can query.

23:37:52 <AaronSw> ... if you want something more expressive you need variables and quotation and quanitifers for the variables. i do that personally.

23:38:21 <AaronSw> ... to sell it to the world getting a data std is first priority

23:38:43 <AaronSw> ... the rules language should be a superset of the data lang

23:39:08 <Eliezer> I'm working with a high-level design - the nodes aren't necessarily going to correspond to any particular classes or modules inside a program. Most of the UML modelers I've looked at so far are built to correspond to a class diagram somewhere, and not to let me word process documents for the nodes and links.

23:39:44 <AaronSw> Q: when will we see stds for rdf query langs?

23:39:52 <sbp> ooh, that's a good Q

23:40:16 <shellac_unstable> who asked that?

23:40:25 <AaronSw> don't know his name

23:40:41 <sbp> -- Anon.

23:40:49 <CaptSolo> Shell: yes. and here, i'm afraid, we need some centralized body assigning identifiers OR some technique of creating unique id from description of school / work company

23:40:52 <AaronSw> A: who would join today? [5-10 hands] hard to judge how long things take. there are sql-like langs.

23:40:55 <grrrr> Eliezer a case tool as big as rose or together will have collabortion diagrams which are themselves choc full of linkage and html editing capabilities. you'll probably have to go heavy on UML or on office productivity tools, but not get one that does both

23:40:56 <CaptSolo> OR - maybe not

23:41:03 <AaronSw> ... seems time for stds, maybe track webont by 1-2 yrs

23:41:20 <CaptSolo> I was thinking in terms of relational db

23:41:40 <Eliezer> grrrr: I'm sorry, I didn't quite get your last answer - you don't know of any UML or office tool that does all of what I want?

23:42:34 <CaptSolo> if information about the name, location, etc. of school is in the same RDF database, then any URI will do as long as i identifies the rest of info about this school

23:42:42 <grrrr> Eliezer that is a correct statement, but UML tools are handy for network diagrams and can attach simple documentation to nodes, and dump html tree-browsers of everything at once

23:43:08 <shellac_unstable> CaptSolo: Well, the thing you're interested in doesn't have to have a URI

23:43:24 <AaronSw> T.V.: too many namespaces?

23:43:24 <AaronSw> Tim: i was going to define &tim; to be my favorite namespace. but you can write a schema that combines namespaces with equivalentTo into one.

23:43:57 <CaptSolo> if i do not use uris, is there big difference between using rdf and using xml?

23:44:08 <CaptSolo> thought uri is one of cornerstonesof rdf

23:44:11 <AaronSw> ... w3c web space for webpages for svg, html, math, WAP, etc. and allowed to evolve then you end up with a nonmoduilar programming and clashes.

23:44:12 <shellac_unstable> CaptSolo: You indirect

23:44:21 <AaronSw> ... find cite stolen by svg

23:44:21 <Eliezer> The problem is that I need to attach complex documentation to nodes. That's why I'm looking for a structured, long document editing tool that handles network structures.

23:44:39 <Eliezer> I'm sure someone, somewhere has encountered this problem before.

23:44:49 <Eliezer> That's why I'm hunting around.

23:45:06 <shellac_unstable> CaptSolo: Eg <school> has <principal> with <mbox> (a uri)

23:45:52 <AaronSw> my job to be unhappy. happy we're making progress. unhappy yhey

23:45:55 <AaronSw> 're too big

23:46:08 <shellac_unstable> CaptSolo: school doesn't have a uri, but no other school will have a principal with that mbox

23:46:20 <grrrr> Eliezer Together/J and Rose both offer hyperlinks to documentation outside the system. the demand for one tool is unrealistic, either go with embedded word/visio or go with a network diagramming tool that is not a wordprocessor

23:46:33 <shellac_unstable> CaptSolo: That's how foaf does things

23:46:38 <AaronSw> happy at test suites, formal models. hopeful at debarncling old specs. hopeful we can close a few WGs. so many things people are excited about. not enough time in the day to handle it all.

23:46:45 <AaronSw> rohit: so what has you excited is prospect of doing less?

23:46:46 <CaptSolo> shell: i see the direction

23:46:59 <AaronSw> tim: expanding the spec but then refocusing it. very concerned about com'plexity of spec.

23:47:07 <AaronSw> ... needed complexity, not arbitrary

23:47:13 <Eliezer> embedded word/visio?

23:47:23 <Eliezer> is that one tool or two?

23:47:47 <CaptSolo> but if all you have is a line from CV, then the most information you can have is <school> has <name> and is located in <location (city, country)>

23:47:52 <AaronSw> ... overall i'm pretty happy. especially at the mental level for process for changing the process. need these questions. that's the path to greater and greater happinges

23:47:59 <grrrr> Eliezer that is microsoft

23:48:10 <AaronSw> the question was mine: are you overall happy or unhappy with the w3c?

23:48:19 <AaronSw> Rohit: you're dismissed. go back to making more barnicles.

23:48:38 <CaptSolo> Aaron :))))

23:49:49 <niq> happy with w3c ... ahem ... without prejudice?

23:50:03 <CaptSolo> shell: URI can be http://cvdb.com/schools#school_name

23:50:37 <shellac_unstable> CaptSolo: you could use literals as identifiers, though

23:51:14 <shellac_unstable> CaptSolo: (warning: bit dodgy until datatyping)

23:51:45 <CaptSolo> shell: not use URI scheme at all?

23:51:57 <shellac_unstable> CaptSolo: yep

23:52:48 <shellac_unstable> CaptSolo: URIs are ideal, but sometimes things aren't ideal

23:53:22 <grrrr> shellac_unstable that's as elegant as defining structs with void* though

23:53:28 <grrrr> even a URI is a step up

23:53:45 <shellac_unstable> oh - undoubtedly

23:54:25 <shellac_unstable> I wouldn't recommend _not_ using URIs

23:54:40 <shellac_unstable> hehe - subvert the semweb

23:54:43 <CaptSolo> grr: :)

23:54:56 <grrrr> a URI might survive untainted from one generation of repository to the next where an unconstrained string may arbitrarily revamp conventions or commercialize a namespace service (hmm, dns?)

23:55:34 <Eliezer> grrrr: It's not actually unreasonable to ask for a structured authoring tool that allows network structures and not outline structures - it seems like the sort of thing that would be relatively easy to whip up in Mozilla+XUL - but it seems almost certain that someone, somewhere has done it already. Anyway, thanks for trying! Anyone else have any suggestions?

23:56:05 <grrrr> Eliezer I'd use my case tools and type <B><P><TABLE> keywords in my docs myself.

23:56:17 <grrrr> in fact, i do

23:56:46 <grrrr> my javadoc comments are elgant html, or one liners, or absent.

23:57:15 <Eliezer> Is that really sustainable for 10K doc / node? This is not Javadoc, more like a research paper describing something with multiple levels of organization.

23:57:21 <CaptSolo> grrr: i write java or html in notepad quite often

23:57:37 <CaptSolo> easy :)

23:57:47 <CaptSolo> although some case tools do help

23:58:00 <niq> aaargh - adobe presentation overloads XMP:-(

23:58:10 <CaptSolo> grrr: what case tools?

23:58:25 <Eliezer> Hey, up until now, I was using a set of HTML documents linked together by a custom Python tool into a long document with footnotes, glossary, autogenerated TOC, and so on

23:58:48 <grrrr> Eliezer I'm comfortable with it. Together/J will dump a document tree a mile long and render the html it finds.

23:59:21 <Eliezer> The problem is that this approach is now inadequate to support the complexity of what I need to describe - I can't *write* the design document without being able to jump around in the network structure. I lose my place.