PI Meeting Breakout Outbriefs
This page contains DAML summaries of the results of
breakout discussion sessions at DAML PI meetings.
The ontology is available in
N3
and
DAML
formats.
Session scribes are asked to
- download example.n3
- add the results from their session
- send the updated file to
[email protected].
This document is available at
http://www.daml.org/2001/07/breakout/.
It was generated directly from the N3 inputs
using cwm,
breakout.xsl,
test.bat,
and a version of DAML XSLT.
Sessions
DAML-S
Significant Issues
- How can we promote interchange?
- What ought to be the relationship of the
Process Model to Daml+OIL? To DAML+Rules?
- Where do we fit in Web architecture?
-> Declarative vs. procedural
-> How will process model be used?
-> Mimimal requirements for process model
- Can we afford to wait for rules?
- Core language does it need work?
-> Need decidable fragment (we have a proposal)
- Relationship between profile and process
model
- What are our goals with respect to
Industry?
-> Create a viable standard?
-> Influence commercial standards efforts?
-> Don't care; just do state-of-the-art research?
- Tradeoffs involving expressiveness,
simplicity, tractability
Plan of Action
- 2001-09-01
- Connect more with industry - Ralph Swick can
help
- 2001-09-01
- Start publishing marked-up services
- 2001-09-01
- Create a web page for user info,
pointers
- 2001-09-01
- Make discussions public
- 2001-09-01
- 'Tie off' version 0.5, Finalize and document
(provisional) solutions to expressiveness problems; Grounding; Identify and
publish info rdf:about tractable subsets.
- tbd
- Make some tools available, hosted at
daml.org, when ready; Profile Crawler & Registry; Matchmaker; Services
editor
- tbd
- Evaluate alternative language bases for
process modeling
- tbd
- Move forward on identified sub-problems
(based on some assumptions rdf:about Daml-Rules); Ontology of rdf:resources, Time,
execution model; Multi-party communications specification, exceptions,
transactions
- tbd
- Continue to provide input to language/rules
developers, and push for sanctioned solutions
Killer Applications
Significant Issues
- Have to show why DAML provides unique value added
- Killer app is NOT killer demo
-> But, can use domain specific information to illustrate utility
- Use combinations of current DAML tools to provide need functionality
- App needs to satisfy needs of many different users in different roles
- Must provide end-to-end life cycle capability
- Producer ontologies different from user ontologies, need to match information across both
Plan of Action
- 2001-07-27
- Establish daml-app mailing list and bulletin board
- 2001-09-01
- Continue dialog on application and domain for illustration
- 2001-09-15
- Goal for identifying specific killer apps is mid-September
Reasoning and Rules
Significant Issues
- need for some sort of closed world negation?
- need agenda/timeline for DAML-Rules language
- DAML+OIL type languages and rules are both reasonably well understood (theory, algorithms implementations), but the union is not
- how much must be captured in a proof/justification language? Just enough to validate? Chains of rules used?
- clarifying relatioship of rules to some logical semantics.
- rules need builtins for arithmetic, conversions, etc. Shouldn't assume any specific implementation language. Might distinguish light and heavy builtins.
- relationship between query languages and rule antecedents.
Plan of Action
- 2001-08-01
- W3C to create a public [email protected] email list. Use www-rdf-logic in the mean time.
- 2001-08-01
- create a DAML Reasoning web page on www.daml.org analogous to http://www.semanticweb.org/XSB/, to include pointers to implementations, etc.
- 2001-08-01
- create a DAML Reasoning Working Group to include reasoning and query languages. Results could include DAML+OIL (upper level) ontologies (analogous to DAML-S) for queries and justifications.
- 2001-08-01
- BBN to create a [email protected] email list
- 2001-08-15
- Pat Hayes, Ben Grosof, and Ian Horrocks to address how to combine DAML+OIL with logic programming more powerful than Horn: efficient subsets, etc. This may evolve into a rules proposal.
- 2001-09-30
- DAML Reasoning Working Group to propose 'official' DAML+OIL Query Language.