Actions

Ontolog Forum

Ontolog Forum Panel Discussion Preparation Page

Thursday 20 October 2005

(previously scheduled for 2005.10.27, this event has since been rescheduled to 2005.10.20)

  • Topic: "Semantic Web Service Ontology Standard"

Proposed discussion agenda

  • Welcome & introduction: different perspectives on WebServices (10 minutes)
    • (1) Modeling web services (tool support for modeling, editing, describing, visualizing)
    • (2) Doing stuff w/ web services (tool support for composition, coordination, orchestration - ignoring issues of distribution for now)
    • (3) Web services in the semantic web (tool support for handling distribution issues: coordination/orchestration vs. transaction, service registration/discovery)
    • (4) Web services as a business infrastructure (tool support for trust, security, authentification, price/resource/service negotiation)
    • (5) Web services as a formal model of workflow computation (process semantics, AI-style planning, schedulability analysis)
  • An ontology of Web Services technologies (5 minutes)

A brief overview of languages/tools/techniques in terms of how they relate to the above pespectives/concerns. Details on specifics deferred to short presentations.

TODO: use DavidMartin's slide #4 from SWANS presentation.

TODO: Provide URLs.

    • WSDL => (1), (2), (3)
    • WSDL-S => WSDL + (5)
    • WSIF => WSDL + SOAP + => (1)
    • BPEL4WS => (1), (2), (3)
    • FLOWS => (1), (5)
  • 4 x Short presentations (10 minutes overview + 10 minutes Q/A)
    • To provide continuity and allow comparisons across presentations, illustrate

each approach using a common example w/ optional variation points.

    • Points of variation: # of services/participants, location/distribution transparency
    • Complex issues: service mismatch (e.g., side effects that may induce coordination/orchestration problems)
    • TBD: OWL-S

(Ask: David Martin, Grit Denker, Everin Siri, someone at U.Southampton)

Focus on (1), (2), (5)

    • JohnDomingue: WSMO

Focus on (1), (2), (3)

Focus on (1), (2)

    • MikeGrunninger: SWSF

Focus on (5), (2), (5)

    • (a 5 minute note on): BPEL4WS

Legacy,plans,etc... relevant for (1), (3) & (4)

  • Total: 1 hour 40 minutes

Pertinent Discussions so far

  • 2005.06.23 Discussion leading to this program
  • 2005.06.30 Discussion Session: ref. our 2005.06.30 discussion session, Michael Grüninger (who has been working on FLOWS and SWSF) indicated that this will be a session he will like to participate in too.   
  • 2005.07.21 suggest Nicolas Rouquette to be our moderator for this session   
  • 2005.08.04 call - Date and Moderator confirmed:
    • Oct. 27 2005 - Discussion on OWL/S, SWSI - FLOWS, and WSMO ... (need proper title) - Moderator: Nicolas Rouquette   
    • Peter and Nicolas will work offline to setup a new page for preparing this session   
    • Michael: note also WSDL-S, try Amit Sheth (of Univ. of Georgia) or someone from IBM (contact Michael)   
    • candidate panelists: JohnDomingue, Michael Grüninger, ..., Nicolas Rouquette (moderator)   
  • 2005.09.01 call - http://ontolog.cim3.net/forum/ontolog-forum/2005-09/msg00002.html#nid05
    • date of this event re-scheduled to Thursday, Oct. 20, 2005
    • EMichaelMaximilien agreed to join Nicolas Rouquette in helping organize this event
    • 2005.09.05 post from EMichaelMaximilien: hoping to get folks from the OWL-S side of things to join us, e.g., Katia Sycara (CMU) and David Martin (SRI).
  • 2005.09.09 / NicolasRouquette's input:
    • I have looked at John Domingue's presentation. It is certainly helpful to understand what WSMO is conceptually.

From a practical point of view however, there are big differences in the way OWL-S, SWSF and WSMO are described in various documents, presentations, etc... These differences mean that it is difficult for a practioner to understand how these 3 models relate to one another.

      • Ideally, it would be great to have a metamodel mapping similar to the ODM that Elisa presented yesterday for OWL-S, SWSF and WSMO.

In practice, we need to scale back this mapping down to the core concepts that are important across models of web services. So, I propose to focus on 2 groups of concepts:

      • Group1: describing how a web service actually works in terms of a computational model for executing it
        • In SWSF, that would be FLOWS (ontology for web services) and its rule-based execution model, ROWS
        • In OWL-S, that would be the process model
        • In WSMO, that would be the service interfaces specified as Abstract State Machines (ASM)
      • Group2: describing what a web service does for the purpose advertisement, matching, etc...
        • In SWSF, that would still be FLOWS & ROWS
        • In OWL-S, that would be the profile model
        • In WSMO, that would be the capabilities, goals, non-functional properties
      • For WSMO, we could easily get the relevant materials from [1] and [2].
      • For SWSF, I only know of [3] and [4]
      • In many respects, both SWSO and WSMO subsume OWL-S

and as noted in SWSO and WSMO, these two approaches seem to be conflicting views but can be in fact interpreted as complimentary to one another.

a discussion where we use both WSMO and FLOWS/ROWS to emphasize how we could improve that example w.r.t. two criteria:

whereas they are defined in FLOW's process model [8].

        • Criteria 2) Clarifying the semantics of service execution
          • b) WSMO's approach for process representation & execution w/ abstract state machines [9]

& SWSF's approach with FLOWS and ROWS as translated into SWSL-Rules [10]

SWSF's edge over WSMO is in the semantic continuity that FLOWS & ROWS bring across service specifications, descriptions and executions compared to WSMO's multiple languages.

      • So, to better understand the practical relevance of both approaches, it would be

helpful to understand how these two approaches can complement each other.

  • 2005.09.10 / EMichaelMaximilien's Input: (ref: http://ontolog.cim3.net/forum/ontolog-forum/2005-09/msg00017.html)
    • As I suggested in my previous post, we should also include the WSDL-S effort. This has potential for significant traction in the SWS space since it takes a very pragmatic and lightweight approach to adding semantics to Web services. There are various communities already using the specification and it was submitted to the W3C. Here are some references:
    • There are also tools available for the Eclipse platform and others:
    • Others are on the works. I have lined up the two point researchers from both University of Georgia and IBM Research (and teams) to participate in the call.
    • Secondly, who are we inviting for the OWL-S side of the world?
  • 2005.09.29 Ontolog Conference call discussion - Panelists situation:
      • John has also appraoched Dieter Fensel, Dieter cannot be with us this time; but we definitely look forward to having him talk to the Ontolog Forum at some other time. Thanks, John.
      • EMichaelMaximilien: Amit Sheth (University of Georgia and Semagix) will also be joining us
      • EMichaelMaximilien: ...(other sugestions ... please add here) ...
      • Nicolas will write to SRI to try to get a candidate to represent the OWL-S view
      • Monica Martin: in view of the focus of this discussion, maybe we should leave BPEL4WS out, because it calls for a slightly different discussion (more application oriented, and not anchored on ontological standards)
      • Both EMichaelMaximilien & Brand recommends David Martin from SRI
    • JohnDomingue: project SUPER, Director Frank Leymann, at Univ of Stuggart (IBM architecture team), got 12 million Euro funding. John could write and invite him to participate. Maybe Frank can talk a bit about their vision from the floor.
    • Brand also suggested looking over: http://www.brainstorm-group.com
  • 2005.10.13: Official Announcement of the Session relased - http://ontolog.cim3.net/forum//ontolog-forum/2005-10/msg00018.html
  • 2005.10.14: change in one of the panelists (as advised by EMichaelMaximilien) - Amit Sheth will be presenting on the WSDL-S perspective, instead of Rama Akkiraju

Outstanding Issues 2005.09.29

  • is JohnDomingue arranging for some more participants / panelists from Europe (as previously discussed)? - see above 2005.09.29 discussion
  • EMichaelMaximilien: who are we inviting for the OWL-S side of the world? - see above 2005.09.29 discussion
  • tie down the list of panelists - see above 2005.09.29 discussion
  • what are we expecting from each panelist (5/10 minute openning? ... etc.) - see above "Proposed Discussion Agenda" (by Nicolas/2005.09.28)
  • format/plan for the entire session - see above "Proposed Discussion Agenda" (by Nicolas/2005.09.28)
  • put together an abstract for the wiki session page - Nicolas will handle
  • drawn up a list of questions as "issues to be explored" - Nicolas will handle (use Nicolas' framework as a basis and compile from there.)
  • we need the Bio's from Rama Akkiraju, Amit Sheth, probably also David Martin - EMichaelMaximilien & Nicolas Rouquette, please request from your respective invitee(s)

Key items for transfer to the main session page

along with Date, Title, Moderator and List of Panelists

Abstract

(please draft)

Pertinent Questions for Discussion

(please draft and list)



Retrieved from "https://ontologforum.com/index.php?title=ConferenceCall_2005_10_20/Prep&oldid=1118"