09:20 *** Log started on 1999-12-09 UTC-0500 *** 09:21 [DanC] Agenda: new syntax 09:21 [DanC] Agenda: new Metadata Activity charter 09:21 [DanC] (requested by TimBL) 09:23 [DanC] NOTE WELL: this channel is logged (http://www.w3.org/1999/12/09-metadata-irc) 09:23 [RalphS] 9:00 09:24 [RalphS] 11:00 Karger & Stein 09:24 [RalphS] 12:30 lunch 09:24 [RalphS] 5:00 09:24 [RalphS] Tim, DanC, Ora, EricM, Massimo, DanielL, EricP, 09:24 [RalphS] DanBri, Ralph 09:24 [DanC] RRS: background: Danny talked about a "semantic web development initiative" at the AC meeting http://www.w3.org/Member/Meeting/99AC/NovAC/Minutes/19991104PM.html#Technology 09:24 [DanC] RRS: Agenda: let's talk about development opportunities 09:25 [RalphS] +Marja 09:25 [DanC] TBL: let's declare namespaces which are RDF models of existing specifications. 09:26 [ora] *** ora (~ora@w3cdhcp2.w3.org) has joined channel #metadata 09:26 [DanC] TBL: HTTP, ICAL (... IETF apps area stuff) 09:27 [RalphS] EM: want to leave with some priorities 09:28 [RalphS] ... and associated names 09:28 [RalphS] ... and dates and venues 09:28 [RalphS] DanC: venues; e.g. conferences 09:29 [DanC] DanBri: "Cabal policy"; process for collaboration among team, etc. 09:30 [Janet] *** Janet (~janet@18.29.3.27) has joined channel #metadata 09:31 [eric] *** eric (~emiller@w3cdhcp18.w3.org) has joined channel #metadata 09:32 [DanC] "The Grocery List" 8 09:32 [DanC] 8 Dec 1999, Cambridge MA 09:32 [DanC] -- RSS 09:32 [DanC] -- DMOZ 09:32 [DanC] -- Dublin Core 09:32 [DanC] -- INDECS 09:33 [DanC] -- IMS 09:33 [DanC] -- Annotation: What's Related, Amaya developent discussions, XLink services 09:33 [DanC] -- WEBDAV metadata 09:33 [DanC] -- http://www.mozilla.org/rdf 09:33 [DanC] -- mozilla sitemaps 09:34 [DLL-lap] *** DLL-lap (~liberte@18.29.3.27) has joined channel #metadata 09:34 [MM-laptop] *** MM-laptop (~massimo@w3cdhcp10.w3.org) has joined channel #metadata 09:34 [DanC] DanBri: RSS: "Rich Site Summary"; drives Netscape's netcenter. Implemented in python; was going to be RDF, but then Guha and [?] left, and now it's just XML [and not even that, really] 09:35 [RalphS] +Danny 09:35 [MM-laptop] *** MM-laptop (~massimo@w3cdhcp10.w3.org) has joined channel #metadata 09:35 [eric-meet] *** eric-meet (~eric@18.29.0.28) has joined channel #metadata 09:35 [DanB] *** DanB (~danbri@18.29.3.27) has joined channel #metadata 09:36 [DanC] TBL: one cool thing to is to build libRDF; the other is to use it as libwww; i.e. when somebody does an HTTP GET, you stick it into an RDF model and support "is this cacheable" as an RDF query 09:36 [eric-meet] maybe an RDF interface to libwww? 09:37 [DanC] [see also: RRS's stickies on the whiteboard] 09:37 [DanB] When I talked to Dan Libby, RSS dropped RDF for plain XML due to lack of tools, lack of experience with RDF, no datatypes in RDF, etc etc. 09:38 [DanB] Dmoz: see http://dmoz.org/rdf.html for RDF dumps 09:39 [RalphS] http://www.dmoz.org/rdf/content.example.txt 09:41 [DanC] RSS: http://www.oasis-open.org/cover/rss.html --> http://my.netscape.com/publish/help/quickstart.html 09:41 [DanC] "Step 1: Design and Create Your RDF Site 09:41 [DanC] Summary File 09:41 [DanC] " 09:44 [DanC] DanBri: Netscape has 4 divergent RDF apps: RSS, DMOZ, [?], [?] 09:44 [DanC] DannyW: what does the epinions stuff look like? 09:47 [RalphS] r:id should be r:about 09:48 [RalphS] http://dmoz.org/Arts/ 09:49 [DanB] http://chefmoz.dmoz.org/ -- DMoz work in progress on restaurant reviews database. "idea is to use the DMoz model to build and distribute the data..." 09:51 [RalphS] DCMI is in the process of creating a directory service; database of schemas 09:52 [DanC] "Rose Drive Friends Church Home Page" -- http://www.rdf.org/ 09:52 [DanC] EricM: I'm in negotiations with rdf.org folks 09:53 [DanC] EricM: we plan to support endorsement in our directory service 09:54 [DanC] EricM: we need help with tools development; we're better at consensus building. 09:56 [DanC] EricM: we're planning to annotate schemas als "foo:title is an rdfs:subpropertyType of dc:title". 09:56 [DanC] EricM: we're securing social commitments to not changing URIs (and content) 09:56 [DanC] ... and high availability 09:56 [DanC] ............... 09:56 [DanC] -- INDECS 09:57 [DanC] EricM: this is metadata for E-Commerce and IPR 09:57 [RalphS] http://www.indecs.org 09:58 [DanC] EricM: timeline: about another year. EU funded 09:58 [Danny] *** Danny (~djw@18.29.3.27) has joined channel #metadata 09:58 [DanC] EricM: their only application to date is MUSE (http://www.muse.com/) 09:59 [DanC] INDECS="interoperability of data in e-commerce systems" 10:00 [DanC] EricM: Dublin Core and INDECS have common needs: uniquely identifying people, orgs, places, subjects 10:01 [DanC] EricM: INDECS folks have strong ties with the DOI folks 10:01 [DanC] "Digital Object Identifier (DOI)" -- http://www.doi.org/ 10:01 [DanB] Entertainment Merchandise Interchange Standard (EMIS) is Muze's RDF/XML vocabulary that comes out of their INDECS implementation; DATAX is David Megginson's RDF semi-implementation in Java. 10:02 [RalphS] http://www.doi.org/ 10:02 [DanB] (no URI yet for EMIS) 10:03 [Danny] http://www.indecs.org/results/model.htm 10:05 [DanC] Massimo: any relationship between INDECS and <a href="http://www.ebXML.org/">ebXML</a>? 10:06 [DanC] TBL: we often say http://example.com/ora--wrote-->http://example.net/ but really, we mean "the person who's home page is X wrote the document available at Y". This is OK if you're clear in the schema, but... 10:07 [DanC] ... we should provide a schema that helps people get this right. 10:07 [DanC] ... another example: houses 10:07 [RalphS] EM: can talk to organizations who do identify persons 10:08 [DanC] EricM: how about targetting the people who are in the business of identifying houses 10:09 [DLL-lap] TBL: We should develop a list of such orgs, contacts, etc. 10:14 [DanC] DC: I want to show contributions of our own before we go knocking elsewhere too much. 10:15 [DLL-lap] Something I am writing: http://www.w3.org/Team/1999/12/sw-w3c-projects.html 10:16 [DanC] DLL: can we make this public? 10:16 [DanC] DC: yes; c.f. http://lists.w3.org/Archives/Public/sw99/ 10:16 [DanC] and rdf-interest 10:17 [DanC] EricM: making this stuff public helps me, esp when folks are wondering if we're doing the right thing. 10:18 [RalphS] *** RalphS (~swick@18.29.3.27) has joined channel #metadata 10:18 [DanB] Here's a note I started after a meeting with Ralph, DDL, Danny (...) http://www.w3.org/1999/11/SW/ "This document provides an informal overview of the W3C work currently known as the Semantic Web Development 10:18 [DanB] Iniatiative. 10:18 [MM-laptop] (Marja: add annotations to DLL's list) 10:19 [DanC] ........................ 10:20 [RalphS] DanC: make my own records smart; 10:20 [Danny] sounds like haystack overlap 10:20 [RalphS] ... when the phone rings, tell me recent contact with the caller 10:21 [RalphS] ... palm pilot & quicken 10:21 [RalphS] promote jigsaw & ACL stuff 10:21 [DLL-lap] OK now use http://www.w3.org/1999/12/sw-w3c-projects.html 10:21 [DanC] publish http://www.w3.org/Architecture/state as a paper 10:22 [RalphS] Ora: build discovery services 10:22 [DanC] (more on Dan's "make my records smart": http://www.w3.org/People/Connolly/drafts/web-research.html) 10:22 [DanC] (more on what dan wants to publish: larch stuff: http://www.w3.org/XML/9711theory/) 10:24 [RalphS] DLL: annotations, searching 10:24 [eric] eric: self organizing systems 10:24 [DanC] DLL: I have a lot of experience with annotations. I'm interested in improving searching/navigation; self-organizing stuff, fuzzy semantics 10:24 [DanC] DLL wrote http://www.w3.org/Team/1999/12/sw-w3c-projects.html 10:24 [DanC] Massimo: look at RDF development cycle: 10:26 [RalphS] Marja: annotations is low-hanging fruit 10:27 [RalphS] Massimo: "xlogic" 10:27 [RalphS] Marja: annotation tools for education 10:28 [DanB] Own interests - main themes: 10:28 [DanB] internet cataloguing community, annotations, trusted 3rd party 10:28 [DanB] description (eg. librarians, medical experts) , better 10:28 [DanB] (shared) bookmarking, data aggregation in RDF. 10:28 [DanB] low hanging / urgent for RDF: 10:28 [DanB] 1. simple metadata bureau protocol (PICS Bureau++) 10:28 [DanB] 2. development of a deployment-led API in RDF Interest Group 10:28 [DanB] 3. datatypes, language tagging 10:28 [DanB] 4. architectural clarifications (URI/resource, content-neg'n etc) 10:34 [eric-meet] short term: feedback on the current RDF parser/DB/browser suite 10:35 [eric-meet] in http://www.w3.org/1999/02/26-modules 10:39 [eric-meet] store acls and accompanying extra RDF in the persistent RdfDB 10:39 [DanC] source is browsable: http://dev.w3.org/cvsweb/perl/modules/W3C/Rnodes/W3CAclInterface.pm?rev=1.93&content-type=text/x-cvsweb-markup 10:43 [DanC] RalphS: I still want to use RDF for CSCW stuff; better human-to-human communication. 10:43 [DanC] ... need tools for knowledge capture, ... 10:43 [DanC] CSCW = "Computer Supported Collaborative Work" 10:43 [DanC] ... shared bookmarks may be low-hanging fruit. 10:44 [DanC] [Dan wants to talk about: how I do shared bookmarks] 10:45 [DanC] [Dan wants to talk about: how WEBDAV could be used for metadata bureau] 10:46 [Danny] Most interesting aspect of machine-readable semantics to me is that it gives users more control over their interaction with the information space in which they participate. Design challenge magnified in mobile arena. 10:46 [Danny] Social/political-economic challenge: what's the incentive to expose semantics, both of behavior and content? 10:46 [DanC] RalphS: a lot of this overlaps with what the XLink community expects. I think if we don't act on that soon, we'll regret it. 10:48 [Danny] Most interesting aspect of machine-readable semantics to me is that it gives users more control over their interaction with the information space in which they participate. Design challenge magnified in mobile arena. 10:48 [Danny] Social/political-economic challenge: what's the incentive to expose semantics, both of behavior and content? 10:48 [DanC] [[[ 10:48 [DanC] [08:51] <DV_> And while some reacted that I kicked the editors a bit too hard in my last mail Tim Bray now hope to have XLink last Call before the end of the year 10:48 [DanC] ]]] 10:48 [DanC] 8:51 -0600 10:50 [DanC] I need to step out at 11amET for the XML Schema telcon. 10:50 [RalphS] Danny: re: incentives -- PICS would have succeeded if there had been an easy way to charge for labels 10:51 [Danny] 1. Find compelling testbeds that have potential for organic growth. 10:52 [Danny] 2. Use W3C site - ACLs with XML Signature, P3P, content semantics 10:52 [Danny] 3. Integrate with Oxygen 10:52 [Danny] 4. Work with MIT Libraries 10:55 [RalphS] Tim: architectural consistency 10:55 [RalphS] new syntax 10:55 [MM-laptop] [ in 4. talked time ago with Ann Wolpert (director of MIT libraries) about RDF and she was totally interested] 10:56 [Danny] Note the connection between annotation/collaboration tools and other T&S activities like P3P 10:57 [RalphS] Tim: clean up my contacts 10:57 [Danny] To MM> Ann Wolpert continues to be interested. 10:58 [RalphS] Tim: events 10:59 [RalphS] threats: xlink, xml-dsig will produce too elaborate models 11:00 [RalphS] Eric: same for xml query; divergent models 11:00 [RalphS] Tim: leveraging existing specs, making them first-class objects in RDF space 11:03 [DanB] re 'leveraging existing specs', WordNet was a fun example: http://xmlns.com/wordnet/1.6/Car http://xmlns.com/wordnet/1.6/Tree etc -- 1000s of RDF classes (when mapping's done properly) 11:07 [Danny] +David Karger 11:07 [Danny] +Lynn Stein 11:24 [RalphS] http://haystack.lcs.mit.edu/ 11:40 [RalphS] *** RalphS (~swick@18.29.3.27) has joined channel #metadata 11:41 [RalphS] haystack has a storage api -- useful to share this with mailto:www-rdf-interest@w3.org 11:41 [RalphS] produce an RDF serialization 11:41 [RalphS] Tim: want to share my "read flag" [seen bit] 11:50 [Danny] DK: metadata can be created automatically (extracted) and will be correct 80% of the time 11:50 [Danny] DLL: mapping enables the SW to leverage legacy metadata 11:51 [MM-laptop] AND not only legacy..... 11:52 [DLL-lap] ... no matter how much we attempt to standardize formats and schemas. 11:56 [RalphS] DK: separate organization from presentation 11:57 [eric] eric: side note wrt automatic metadata generation: OCLC has been experimenting with this and we've found (roughly) if we're anaylizing a known collection for which we tune our analysis we can achieve around 70-80% accuracy... this accuracy degrades significantly, however, when we analyse arbitrary objects/collections 12:10 [RalphS] RRS: how does haystack handle quoting? 12:10 [RalphS] DK: nothing is a fact 12:11 [RalphS] every tie (statement) is first-class and other ties (statements) say things about it 12:13 [RalphS] TBL: reification is impractical -- every statement needs to be reified 12:14 [RalphS] OL: don't confuse the model with an implementation. It is possible to implement the RDF model in such a way that every statement is first-clas 13:01 [DanC] ............. 13:01 [DanC] resume 13:01 [DanC] rathole on syntax 13:02 [DanC] DC: I'm in favor of developing a new RDF syntax by way of solving application problems (e.g. DSig, P3P) 13:02 [DanC] TBL: the benefit to DSig of a generic RDF syntax is that it makes it clear what you mean. 13:02 [DanC] DC: no, it give you RDF triples, which they don't appreciate. 13:03 [DanC] DC: If we give the DSig folks that checks signatures, they won't mind that it happens to generate triples too. 13:03 [DanC] ... if we give the P3P folks a tool that makes privacy decisions, they won't mind that it happens to generate triples too. 13:24 [DanC] [... gap in scribe notes ...] 13:26 [DanB] RDFS excerpt: 13:26 [DanB] this specification introduces a number of other 13:26 [DanB] *** DanB (~danbri@18.29.3.27) has joined channel #metadata 13:28 [DLL-lap] Discussion on... RDF schema meaning of range constraints: does it validate or generate? 13:38 [RalphS] *** RalphS (~swick@18.29.3.27) has joined channel #metadata 13:38 [eric-meet] TimBL: re schema consistancy checks: inconsistancy in a union implies falsehood 13:42 [DanC] TimBL: pls edit the RDF Schema spec to stop using the word "inconsistency" 13:52 [Danny] http://www.w3.org/Signature/Participants.html 13:53 [eric] Naohiko URAMOTO (IBM), 13:54 [eric] Naohiko was very interested in DSIG and a member of the RDF working groups 13:56 [DanB] NewsFlash: Just heard Stanford are joining W3C and will be reworking the Protege-2000 as opensource RDF editor 13:57 [DanB] http://www.smi.stanford.edu/projects/protege/protege-rdf/protege-rdf.html 13:59 [RalphS] tools: 13:59 [RalphS] API 13:59 [RalphS] persers 14:00 [RalphS] registry management 14:00 [RalphS] schema creation GUI 14:00 [RalphS] data entry/editing GUI 14:00 [RalphS] long-range navigation 14:00 [DanB] *** DanB (~danbri@18.29.3.27) has joined channel #metadata 14:00 [RalphS] long-range navigation (graph browsing) - someday pile 14:01 [RalphS] validation 14:01 [Danny] tools categories: infrastructure and apps 14:02 [RalphS] juicy fruit: 14:02 [RalphS] DSIG 14:02 [RalphS] P3P 14:02 [RalphS] Libraries 14:02 [RalphS] ACL 14:02 [RalphS] W3C Membership 14:08 [RalphS] other tools: 14:08 [RalphS] backend datastore 14:08 [RalphS] robot 14:15 [Danny] w3c membership/acl/dsig app should have generalized backend database strong/API for public use 14:44 [DanB] Local installation of crit-suite for annotation, eg. http://crit.w3.org/http://www.w3.org/Collaboration/ -- is this good basis for SWDI demos? 14:49 [RalphS] *** RalphS (~swick@18.29.3.27) has joined channel #metadata 14:55 [eric] http://www-diglib.stanford.edu/diglib/ginf/ GINF home page 14:56 [DanB] http://www.w3.org/1999/11/SW/ -- DanC argued this should be mutated into briefing package 15:01 [Danny] Action: make clear that the RDF IG is coordinating point for RDF tools open source community 15:02 [RalphS] *** RalphS (~swick@18.29.3.27) has joined channel #metadata 15:03 [Danny] Propose that RDF IG do an interop every 6 months 15:04 [Danny] Briefing package points: 15:04 [Danny] 1. dogfood 15:04 [Danny] 2. rdf ig as community coorditing point & regular interop coordinator 15:05 [MM-laptop] Ora: in Darrpa there are TIE'S (Technology Interaction Experiments), could be useful here too. 15:06 [eric] *** eric (~emiller@w3cdhcp18.w3.org) has joined channel #metadata 15:06 [ora] TIE = technology integration experiment 15:09 [Danny] registry mgmg tool: database, api (w/query), parsing rdf schemas annotation, gui interface 15:09 [Danny] mgmg=management 15:22 [RalphS] *** RalphS (~swick@18.29.3.27) has joined channel #metadata 15:24 [DanB] http://www.smi.stanford.edu/projects/protege/protege-2000/instances_tab.html -- eg of Protege GUI 15:32 [MM-laptop] next meeting: Thu 16 Dec 9.30EST 15:32 [MM-laptop] teleconference for 1 hour 15:33 [MM-laptop] confirmed: eric, danw, dabri, ora, massimo, danl, ?timbl, ?henry 15:33 [MM-laptop] danC will propose something to review at least 12 hrs before the call 15:35 [RalphS] +confirm: ralph 15:36 [RalphS] ACTION: Dan supply a proposal to review before 16 Dec 15:38 [MM-laptop] 3.40pm the meeting is adjourned 15:43 [RalphS] *** RalphS (~swick@18.29.3.27) has joined channel #metadata 15:47 *** Log ended on 1999-12-09 ***