DAML: MIT Program Review (2000-2004)

Sandro Hawke (MIT-CSAIL, W3C)

DAML PI Meeting, San Antonio, Texas

1 December 2004

http://www.w3.org/2004/Talks/1201-daml-mit/
$Revision: 1.78 $

DIG

Decentralized Information Group

• Research group at MIT-CSAIL
• Led by Tim Berners-Lee
• overlaps W3C Sem Web Advanced Development (SWAD)

Staff

• Dan Connolly (OWL, cwm, DAWG)
• Danny Weitzner (Tech&Society)
• Eric Miller (Activity Lead)
• Eric Prud'hommeaux (DAWG, algae, Annotea)
• Ralph Swick (Best Practices, Zakim, Annotea)
• Sandro Hawke (Ontaria, LX, OWL, rules)
• Tim Berners-Lee (PI)(cwm, n3)
• (Students)

Affiliated Researchers (for DAML)

• David Karger (Haystack)
• Lynn Andrea Stein (DAML+OIL)

Decentralized Information Systems

• No central bottlenecks (scalable)
• No central point of failure (reliable)
• No central control (social)
• require standards
• Semantic Web Architecture

Tactics

• Stay grounded with needed (real) applications
• Small steps from known solutions (the Web)
• Rough consensus & running code (IETF credo)
• Open source (release early, release often)
• Strive for simplicity
• Trust the market will follow

Get the Semantic Web Rolling

1. Real-World Applications
   (Annotea, Haystack, Ontaria, Meeting Support)
2. Application Platform
   (cwm, N3, delta, algae, LX, semwalker)
3. Transition to Standards
   (JC, RDF Core, OWL, Semantic Web Activity)

Applications

• Provide Use cases
• Close to home
• Demonstrable

Annotea

Post-it notes on web pages

• Demonstrator using sharable RDF for collaboration
• Uses algae for server platform
• Prototypes RDF protocol
• AKA: Shared bookmarks
• Requires browser modifications (alas)

Haystack

Personalized Information Management

• email, news, documents, all in your RDF store
• powerful ui design to make it all navigable
• needs RDF Query, could use rules [?]
• might be the Semantic Web Browser

Ontaria

Ontology Discovery Service

• Help people find and examine ontologies
• Decentralized: all data comes from RDF on the web
• Show provenance and proofs (find the bad data)
• Linked to W3C Ontology Hosting service
• (Semantic Web Browser?)
• 0.8 released 10 weeks ago; ~4k hits/day; 4841 IP Addrs

W3C Document Workflow

Automate W3C Document Production

• Formalization of W3C process
• TR page automation

Calendar Coordination

• iCal and RDF conversion tools
• calendar diff/sync
• travel demos
• bank-statement on calendar

Meeting Support

Semantic Record-Keeping and Tracking

• agenda & minutes ontologies
• action item tracking
• [irc] conversation recording
• audio teleconference bridge tie-in (Zakim)

Application Areas (Summary)

• W3C Core Business:
  document production, meeting support
• Web Users: Annotea
• Information Users: Haystack
• RDF and OWL Users: Ontaria

Application Platform

• open source
• deeply RDF
• mostly rule-based
• mostly perl, python, prolog
• feeding standards work

cwm

general-purpose tools for manipulating semantic web data

• On-ramp for new sem web hackers
• Prototype N3 langauge
• Built-in functions as RDF properties
• Fairly well known, 1.0 release recently

n3

• An easier RDF syntax (has "Turtle" subset)
• N3-full adds nested graphs, variables
• Result: covers rules and query needs with single syntax
• Many implementations other than cwm
• Wide and widening use in RDF community
• Grammar described in RDF

Semantic Web Logic Language

"SWeLL"

The Semantic Web concept is to do for data what HTML did for textual information systems: to provide sufficient flexibility to be able to represent all databases, and logic rules to link them together to great added value. -- original proposal

• (log:includes, log:semantics, etc)
• A vocabulary and syntax for adding rules to RDF
• Horn-equivalent, plus Semantic Web builtins
• Part of N3 syntax and expressable as an RDF vocabulary
• Implemented by cwm

reification and proof

ontologies, processed by cwm

• flatten n3's formula literals into RDF
• trace cwm rules firing
• verify transmitted proofs (rule-traces)

local closed world

• log:definitiveDocument
• log:notIncludes

delta/sigma

ontology, processed by cwm

• format for knowledge differences
• not just RDF-graph-diff, uses owl:fp/ifp
• useful for calendar, RDF testing, update protocol

crypto

• crypto:md5
• crypto:sign
• Used to construct delegated trust demo

LX (2002)

Logic Exchange Ontology

• Straightforward ontology of FOL syntax
• Usability (implies truth) predicate, on RDF
• Tools for syntax conversion (to/from Otter, lbase, n3)
• Axiomatization tested with Otter

algae

• query language / system
• persistent triple store
• SQL database access
• RDF access to other system data (filesystem, finger protocol)
• oldest surviving RDF system?
• backs w3.org ACL system, annotea

XML Integration / Legacy Data

Redeeming the user's investment in XML

• Blindfold: EBNF with annotations, based on btyacc (read n3, WSDL)
• GRDDL: XSLT output RDF/XML
• Ontology-Driven-XML Parsing: letting nice XML be read as RDF (stripe-skipping)

semwalker

Server-side RDF browser with integrated Search

• the code behind Ontaria
• full-featured real-time web harvester
• class-specific view of reasources
• all driven by rules, potentially read from web
• fast 2-level RDF cache for millions of triples
• implemented in SWI Prolog

Pre-Standards

• n3-turtle
• n3-full/LX/SWeLL [cf SWRL, DRS, RuleML, SWRL-FOL, SPARQL/Construct, CL]
• delta ontology [cf Atom, Pub/Sub, DiffFacts]
• proof ontology [cf KSL's PDL]
• local closed world ontology

Transition to Real Standards

• MIT as W3C Host: Unique Position
• W3C is the member organizations, not the Team
• (Team is ~70 people from hosts, members)
• Consensus can't be forced
• ...but it can be supported (examples follow)

Acknowledged Submissions

pre-standardization positioning

• DAML+OIL (now OWL)
• SWRL
• OWL-S

W3C Recommendations

• RDF Core
  • WG participation
  • Many issues raised, clarified
  • prototyping in cwm, algae, ...
• OWL
  • Team Contact
  • Surnia (OWL Full Reasoner), demonstrating OWL-Full Interop
  • OWL Test Results page, making sense of ~300 tests x 10 systems

Current Working Groups

• Data Access Working Group (DAWG)
  • chair, team contact
• Best Practices and Deployment Working Group
  • team contact

Other Consensus Building

• Mailing list participation
• Demonstrating interoperability
• Running wiki (esw.w3.org)

Conclusions

Transition/Handoff

• Influenced RDF Core, OWL, ~SWRL
• Built cwm, ontaria, algae, annotea, etc; (dev.w3.org)
• Papers? SciAm, Haystack, mostly invited talks
• No commercial or DoD pilot/product work
• Changed World? We all helped get the Semantic Web going

Conclusions

(Ongoing Work/Issues)

• Logic on the Web (beyond OWL)
• cwm
• Ontaria (semwalker)
• Policy-Aware Web (trust)

STOP! You're Done! Go Back.

various reports to daml

http://www.daml.org/2001/04/iow/mit/
http://www.daml.org/2003/01/iow/mit/
http://www.darpa.mil/ito/psum2001/K535-0.html
http://www.w3.org/2000/01/sw/DevelopmentProposal
http://www.w3.org/2002/02/iow2
http://www.w3.org/2002/03/DARPA-SOW
http://www.w3.org/2002/11/29-sw-qr
http://www.w3.org/2002/11/DAML-IOW
http://www.w3.org/2003/09/15-damlwp/darpa-wp.html
http://lists.w3.org/Archives/Team/sw-team/2004Apr/att-0048/DAMLQ1report.html
http://www.w3.org/2004/07/DAMLQ2report
http://www.w3.org/2004/10/DAMLQ3report