meeting notes from SWSL teleconference, 8/12/04

SWSL teleconf meeting notes
8/12/04

taken by Rick Hull 

(see also a note by Benjamin, which summarizes the discussion
 re proposed relationship to RulesML and Joint Committee activities)

Attendees:

  Richard Goodwin
  Sheila McIlraith
  Benjamin Grosof
  Michael Gruninger
  Rick Hull
  Michael Kifer
  David Martin
  Jianwen Su


pre-discussion: that DAML funding is running out, and we
  should start exploring funding avenues, at a minimum to
  get some nominal support for the SWSL effort, and more
  generally for semantic web (services) research

-----------------------------

Topic: 

SWSL position paper to W3C workshop on Constraints and 
Capabilities for Web Services

David: If it is coming from SWSL, then it should represent the full
  view of SWSL committee, or even the SWSI committee

  (Anybody can individually submit a paper, representing their own 
   organzation).

Benjamin: seems like a pretty important workshop; strong semantic web
  connection

David: who else wants to go, anyway?

  Jianwen: it's close for me
  Sheila: I have a strong interest, but teaching commitments may prevent me
  Rick: may want to go

Note: at most 2 people can go per "organization".
  people can separately submit papers, of course, outside of SWSL

Rick: it is important that we come up with *something* for this
  particular venue

Sheila: how about someone creates a skeleton, 
 -- draw on expected skeleton for our "note"
 -- talk about our requirements for SWSL
 -- talk about limitations of previous approaches currently out there
 -- then room for a section on ontology and on rules

Benjamin: in the call, they are looking for some high to medium
  level position paper; don't need much technical content

  we can give flavor of where we are going, and refer to some other documents

  try to maximize overlap of this submission and our note

David: that each paper should include use cases, and should address
   a particular use case that they provide on the web site

David: does someone want to take the lead to make first skeleton?

Benjamin: he may go with UML hat on

Sheila: willing to
  a) do second pass if someone wants to make a first skeleton
  b) produce and/or coordinate the FLOWS section

No one volunteered to take lead on producing the SWSL position paper

----------------------------

David: WSMO/SWSL merge meeting - people are talking about January

Sheila: Jan is a time for submitting to AI conferences, so doesn't
  seem like a great time

  maybe both of these deadlines are late in Jan

MichaelK: 2 day meeting

Sheila: people will be away for 4 to 5 days

MichaelK: first week of Jan people aren't teaching, usually

----------------------------

TECHNICAL DISCUSSIONS:

Rules group:

David: that they have essentially made the decisions needed to
  have an FOL in SWSL

Benjamin: 
  some low-level decisions about some of the syntax (see notes of MichaelK)
  
  that Joint Committee made a note about FOL there; Benjamin has
     been acting a bit as a liaison with that group

    question from them: how much specialized, non-FOL stuff does SWSL want
       to include, e.g., HiLog, reification, F-logic syntactic sugar ...?

    Joint Committee will probably go into hibernation in a couple of months,
       since the FOL was one of last things on their plate.  They will
       produce an FOL spec before hibernating. They have 2 drafts
       now, one minimalist and one not; they will get merged. 
       This includes (a) an abstract syntax and a (b) mark-up syntax.
       There isn't a plan to make an ascii-style syntax.

  Rules group is developing a presentation syntax and a rules FOL,
       which seems

Rick: will the rules FOL from SWSL be compatible with the Joint Committee
    abstract syntax?
Benjamin: they are more minimal; not incorporating things like
    HiLog, reification, and F-logic syntactic sugar

MichaelG: are they incorporating SCL?
Benjamin: no, they are focusing on a pretty stripped down thing
     
David: how much do we care, whether the Joint Committee FOL lang
   is stripped down or not..

MichaelG: seems like the vanilla syntax from Joint Committee is sufficient
   for the process ontology

Benjamin: do we want to request that they tackle more than vanilla FOL?

MichaelK: not sure why we need them to do it.  We can produce it fairly
   quickly.  In fact, we more or less have it.

MichaelG: is this an extension of the language of the Joint Committee?
  Usually it is good to re-use things coming out of W3C-oriented committees
Sheila: and that way we can get others to do work that 

MichaelK: we are already building on RuleML
Benjamin: with respect to SWSL rules, the clear path is to RuleML.
   For the FOL aspect, harder to know what to build on.  
   Initially, we should have a layered approach.
    For the FOL part we would strive for compatibility with the Joint Committee

Benjamin: here's what I hope will happen, and have some reason to believe
     will happen
  Step 1: SWSL produces the SWSL rules + FOL
  Step 2: Joint Committee: vanilla FOL
  Step 3: SWSL rules group checks compability of 
  Step 4: expressive extensions for rules and the FOL language
             (HiLog, reification, F-logic syntax) we pursue as a mark-up
             syntax, in collaboration with RuleML.  Or maybe we
             turn it over to RuleML
  Step5: RuleML is probably involved in step 4, and may be carrying
            it forward in the long run

Benjamin: so, under this scenario, we could issue a note to Joint Committee
  saying that "SWSL does not want you to go beyond vanilla FOL, at least
  between now and November."

MichaelG, Sheila, Rick: that is OK

No one objected.

Action item: Benjamin will send this note, on behalf of SWSL, to the
   Joint Committee.

David: that Benjamin's note should include the larger proposed scenario,
  about how RuleML group would take over various aspects

David: is there any concern that the RuleML group would be less attentive
  to OWL concerns?

Benjamin: to some extent this came up with the SWRL group

---------------------------------

Process subgroup:

MichaelG: 
  August has been declared as "ontology month"

  so, we are postponing the syntactic discussions until we have the ontology
     nailed down

  yesterday we considered the web service architecture document, and 
     discussed the ontological status of messages

  up til now we've been extending OWL-S into FOL, but messages is a
     first place where we seem to need to go beyond the way that
     OWL-S did messages

  Are messages a generalization of I/O or something more

  See 8/11/04 telecon notes.  We didn't really come to a conclusion; we'll
      take another crack at it at next week's telecon.  This seems
      like a very high priority

David: in the latest version of OWL-S we have IOPR's rather than IOPE's
  R - conditional bundling about outputs + effects + optional info about 
         output values.
Sheila: that a set of outputs and effects can be bundled with one condition

David: the Web Service Arch document talks about agents that request or
   provide services.  But in OWL-S we've talked a lot about the need
   for a more broad notion

Rick: that BPEL indicates that a service can be both client or server with
   respect to another service

MichaelG: we can imagine richer forms of interaction - "collaborative",
    ...

David: last week we were discussing whether the SWSL Rules proposal for FOL
  syntax would be useful to the ontology group.  There is an action item
  that Rules group will give written proposal to Process group for evaluation

MichaelK: that MichaelG has an action item - to identify whether the
  process descriptions can be specifying with rules

MichaelG: by next week he'll give a partial answer -- focused on the
  parts of the ontology that we've worked out so far, which is pretty
  much the OWL-S part

MichaelG: we won't represent the concepts completely (which have
  FOL semantics), but we can at least descriptions of certain

---------------------------

Recap:

MichaelK: did we decide to go with RulesML for the markup, or not?

MichaelK: that I am inclined to go with RuleML, and I think Benjamin
   is inclined to do this

Benjamin: pros and cons
  pros: RuleML already includes syntax, general attention, on a very
     large expressive subset, of what is in SWSL rules design.
    Not all of it, but much of it.  SWSL rules was influenced
    a lot by RuleML

  cons: that Joint Committee/SWRL has wider recognition and broader
    appeal within the semantic web community.  A lot of OWL-ees
    are more familiar with Joint Committee/SWRL, and they aren't
    familiar with RuleML.  Industrial types are more familiar with RuleML.
     
  Things RuleML didn't address
    LP as fundamental seemantics
    non-mon
    HiLog, F-logic syntax, ...

  If we depend on RuleML, they may never get to adding this stuff
  And perhaps we don't have sufficient critical mass to take
      on the whole job of making the extension

David: that we don't need a full proposal for XML serialization
  for the first note.
MichaelK: but we need a roadmap
David: right, we can say that we will work with the RuleML group to
   ensure that we have appropriate expressiveness

David: if that is our story, does that imply that Joint Committee
   work on FOL is a waste?
Benjamin: no, it is a useful first phase

Benjamin articulates a resolution (see his notes)

Benjamin will distribute his summarization note for comment


----------

MichaelK: what did we decide about joint SWSL/WSMO meeting?

MichaelG: what about third week of Dec
MichaelK: exam time for me - 18-20 should be OK

David: vague concern from some about being away that close to Christmas

Rick: Chanukah is Dec 8 - 16, 2004

MichaelK: how about very beginning of Dec?

Sheila: still teaching, and near a DAML-PI meeting, I think

David: let's identify some specific dates and proceed by email