Semantic Web
Services (and S'm'other Semantic Sundries)
Bijan Parsia Maryland Information and Network Dynamics (MIND)
laboratory, Semantic Web Agents Project (SWAP)
University of Maryland,
College Park
I dedicate this talk to all the people who give me grief
about writing complete sentences, paragraphs, and other sensible narrative
units in my presentations, or for putting in too many interesting side
issues, or even just developing certain topics with all the abandonment
those rich and complex subjects deserve in all but the most shallow and
inane discursive contexts.
This one's for you.
Semantic Web
Services
- WS tasks
- Discovery and Selection
-
Negotiation and Contracting
- Composition
- Execution
- Monitoring, Simulation, Verification
- SWS: Greater
and better automation
- Analyzable descriptions: "Rich, but not
too rich"
- Fundamental Principle: There Are Never Too Many Web
Services Groups (TANTMWSA)
- Corollary: The probability the Bijan
will end up on one of them is directly proportional to its existence
Note on the
talk
- These slides...
- ... cover too much
-
shallowly
- glide over deep issues
- relevant to WSDL
- BOF material, IMHO
- Feel free
- to
buttonhole me
- solicit demos
- come visit
-
admire my selfless dedication
WSDL
- WSDL has a Requirement: "Mapping to Semantic Web/RDF"
- I
joined the group to provide relevant expertise
- Group has some
interested members, but...
- Group intends to produce rec status
mapping
- Work has been waiting on stability
- WDs to
start at/after this F2F
- Aiming for May LC
- Y'all
demanded it; please support it (esp. in CR)
- Two Big Issues:
- Using RDFS/OWL "types" (and the Decker problems)
-
Modeling XML Schema
WS-Architecture
- Quite a lot of Enthusiasm for Ontologies
- CMU
participants produced Enthused-for Ontologies
- Part of the
published document
- Could be good ground for other specs (WSDL?)
- Don't orphan them!
- SWSA may make use of them
WS-Choreography
- Some wanting of semantics, some not wanting
- !McCabe: Chore is the
"how", not the "why"
- pi-calculus -- a whole other formalism
family
- SWSL is interested, may coordinate
- BPEL's
deafening silence
Semantic Web
Services Interest Group
- Brand spanking new (or was)
- OWL-S and SWSI learning
publicity there
- Perhaps transfer the WSDL nagging?
-
That's where I discuss SWS related stuff
UDDI
- Longtime interest
- Recently made moves to incorporate
some RDF
- Lots of low hanging fruit (we hope)
- CMU folks
joined TC
- Lots of
non-triviality
- Do first, nag later
OWL-S (nee
DAML-S)
- Reached 1.0, fair bit of use
- Including commercial!
- Grid too
- Covers (in intention) all the tasks
- Still plenty of open issues
- Precondition and
effect language
- Data flow
- Semantics!
- OWL
Full vs. DL (The Protege experience)
- More
- Some
sort of closure by May
- The rest of the DAML program
-
The DAML successor?
SWSL
- Half of SWSI
- Successor to OWL-S
- Covers (in
intention) all the tasks
- Slowish start
- Hope for a toy
proposal by May
- May is an Exciting Month
- May may
be rescheduled for Dec.
- Lots of inputs
- OWL-S
- NIST PSL
- Concurrent Transaction Logic
- AI
Planning
- Rick Hull's work
- More interaction with
Standards Groups
Some tools,
software, and systems
Challenges
- Composition as Planning
- Reasoning about State
-
The Decker problems
- Matchmaking in and open environment
- Expressiveness
- vs. Usability
- vs. Analyzability
- vs.
Standards (Rules, please. More, actually.)
- Policies!
- Dublin Core
- Crowded space; fight reinvention
-
Buyer's fatigue
- Less magical thinking, please
- Even
less magical speaking, please