FOUNDATION FOR INTELLIGENT PHYSICAL AGENTS

Document title

Resolutions from the Barcelona, Spain Meeting, March 2004

Document number

f-out-00156

Document source

Board of Directors

Document status

Preliminary

Date of this status

Date as: 2004/4/5

Change history

2004/4/5

Initial draft, Mike Kerstetter

 

 

 

Resolutions of the 31th FIPA meeting in Barcelona, Spain
March 17-19, 2004.

 

1.      FIPA activities:

 

Meeting

Dates

Location

Country

Host

32

14-16 July, 2004

Washington DC (Rockville, MD)

USA

Renato Levy, Intelligent Automation

33

27 Sep.-1 Oct., 2004

Erfurt

Germany

In conjunction with MATES 2004 conference, Dr. Rainer Unland, Organizing Chair.  Jim Odell will coordinate on behalf of FIPA.

2.      The board approves the FAB recommendation to merge TC IP into TC Modeling.

3.      The board approves the FAB recommendation to merge TC Ontologies into TC Semantics.

4.      FIPA requests that the Secretariat update Board-related mailing lists as appropriate and remind all committee chairs to update their relevant sections of the FIPA website with current information.

3.      The board will continue to work on the strategic positioning of FIPA and its activities with respect to other relevant standardization bodies.

4.   Resolutions and Statements of Technical Committees, Working Groups, and Special Interest Groups:

TC Ad-hoc

Discussion about current specification draft. The following decisions were taken:

·        Changes to specification 00023 (Agent Management) and new specification 00095 (Agent Discovery Service) approved as experimental by TC ad hoc

·        Work on white paper will not be continued, workplan is fulfilled

·        TC Ad Hoc remains formally constituted during the call for implementations and follow-on activity.

·        TC ad hoc thanks the co-chair Heikki Helin for his work, Michael Pirker appointed as new Co-Chair

Next steps:

·        Final formatting of the FIPA Agent Management specification

·        Deliver experimental specs 00023 and 00095 for approval to FAB

·        Specification number 00096 not used, give back to FAB

·        Call for Implementations

·        Call for Information (to the experimental spec)

TC Ad hoc thanks to: Joan Ametller, Fabio Bellifemine, Michael Berger, Santiago Fraga Castro, Monique Calisti, José Manuel Juarez Herrero, and Michael Pirker.

TC Methodology

TC Methodology proposed that FIPA adopts the following resolutions:

·        Existing Fragments evaluation and abstraction level comparison.

o       We looked at the already defined fragments and argued that some levels of abstractions could be identified and an hierarchical composition of such fragments is possible. This possibility should be pursued during the identification and definition of future fragments

o       We identified the need of analysing the different level of abstractions that could be useful in fragments description (work definitions, activities and steps) and produce some guidelines about their use in fragments extracted from different methodologies

·        Method base structure

o       An initial specification for the method base structure (proposed by A. Garro and L. Sabatucci) has been discussed and several modifications introduced in it. Further work will be done offline interacting in the mailing list with the aim of consolidating an initial model for the next meeting.

·        MAS meta-model

o       We analysed the three meta-models of ADELFE, GAIA and PASSI and concluded that they can be seen as an extension of the Modeling TC MAS meta-model.

o       In order to avoid confusion, specific methodology meta-models will be labelled using the syntax:

<methodology> MAS meta-model

o       Further studies will be carried on in order to identify commonalities among meta-models that could help in composing a new meta-model reusing the existing ones or eventually unify a core for them

·        Fragment definition

o       We analyzed comments from some members and accepted few of them

o       As a consequence, members are discouraged from using entry 9 of fragment definition (dependency relationships among different fragments) since this information will be more properly located in the method base and therefore further work will be done to express it in the method base structure

·        Methodology scope and consequences

o       We initiated the discussion about the possible approaches in constructing a methodology and specifically dealt with requirements driven and unification driven points of view

o       We started a discussion about the identification of a number of disciplines that should be covered by a FIPA methodology

TC Methodologies gratefully thanks all people attending the meeting and the others contributing offline

TC Modeling

TC Modeling proposes the following resolutions:

·        The meta-model created at this TC represents the link of the agent models with their counterparts in software engineering and should clearly indicate/define what the paradigm contributes to software systems.  (i.e. what agents can do better then OO and when?)

·        Methodology specific meta-models should extend/inherit the general meta-model created at the modeling TC

o       The general meta-model has to able to support all current methodology meta-models. Desire to k

·        eep the general meta-model a minimum set.

·        The general meta-model will be the basis used to define the new modeling elements required to describe agent systems.

·        Current meta-model core is starting to consolidate, but there are other parts that will be needed, and we can start this in parallel with the remaining efforts in the core

·        Start the definition of missing parts and how the combine/compose with core meta-model

TC Security

The main achievements at the meeting were discussions related to the FIPA f-in-00095 document. This is close to being a completed preliminary specification proposal. Informal advice has been sought from the FAB regarding dealing with the general issue of extending existing current standard specifications, in particular the abstract and concrete envelope specifications) to support the new envelope security extension. Currently, it is proposed that a new extension field be added to the current envelope standard specifications that points to a set of extension headers, one of which is the new security header (another, in the future, could be a possible QoS header). At this time, it is left to the FAB to state more exactly the process for dealing with enhancements to current specifications.

The FIPA Security TC proposes the following resolutions:

1.      To complete a preliminary specification of the security envelope specification in time for the FAB to consider for the next meeting. To propose the changes needed to the related standard envelope specifications.

2.      To propose a draft of the abstract security architecture specification

3.      That the following be thanked for their active input at this meeting: Sergi Robles (Univ. AB), Alois Reitbauer (Profactor), Fabio Bellifemine(TiLab) and Stefan Poslad (QMUL). Apologies are offered to anyone whose name was omitted.

TC Semantics

TC Semantics had excellent technical discussions and expanded from two meeting sessions to three.

·        Resolved:  To adopt and execute the initial plan.

·        Initial Plan:

o       DRAFT & Circulate:

§         Goal / proof of success

§         White Paper Outline

§         White Paper

o       Establish Awareness of other standards / languages

§         query languages, conferences, organizations, human contacts

                        (e.g., W3C semantics, AgentCities, AgentLink…)

o       Call for inputs

o       Discussion

o       Create Workplan

·        The chair thanks the many people who participated in the sessions.

FIPA Architecture Board

FAB adopts the following resolutions:

·        Approval of the specifications for Ad-Hoc Networks to be integrated as slight extensions of FIPA Agent Management Specifications

·        Recommends the merging of TC Modeling into TC Methodologies

·        Recommends the merging of TC Ontologies into TC Semantics

·        Specification number 96 is no more used by TC ad-hoc. The SECRETARIAT is requested to take note and take the necessary actions.

·        Approval of the requirement of the Security TC to represent security-related information into the Envelope even if this is expected to break compatibility  of the codecs. FAB recommends to find a solution with as minimum effect as possible on the existing implementations, but ensuring a clean solution for future extensions of FIPA.

·        Appoints Fabio Bellifemine and Stefan Poslad to review the 3 responses from the following individuals/groups that have been received for the call for an infrastructure for exploiting FIPA URN, with Steven Willmott as consultant:

o       Ákos Hajnal and László Varga, MTA SZTAKI, System Development Department, Budapest (Hungary)

o       Stephen Cranefield, Department of Information Science, University of Otago, New Zealand

o       S. Rullo, P. Iritano, A. Garro, Exeura Knowledge Management Solutions & Univ. of Calabria, Rende (Italy)

Image Committee

The Image committee resolves the following:

·        AAMAS04

o       Distribution of FIPA Inform! Newsletter, vol 5 issue 1

o       Distribution of the FIPA slides as annex to the newsletter

§         MCA coordinates with Jacquie and AAMAS OC

o       Ask AAMAS committee if any spot for FIPA is available:

§         Ceremony for best papers’ award (Mike or Stefan) – FIPA prize

§         Plenary announcement about the next FIPA meeting

o       Call for papers for the AAMAS04 workshops: reminder on the chat mailing list about deadline

·        Liaisons

o       Besides the usual established connections, strengthen FIPA liaisons with

§         Agentlink III

§         Global Grid Forum

§         Semantic Web

o       Issued before next FIPA meeting

o       Content:

§         Meeting Report (M. Kerstetter)

§         News in brief (J. Kelly)

§         Barcelona workshop report (Monique and Sergi)

§         FIPA membership and participation (BoD)

§         Technology overview: Semantic Web (Terry Payne)

§         TC methodology report (possible addition)

§         Corner of announcements

·        Action Points

o       Collect input about relevant events: (usual business)

§         Agents and non agents events

o       Collect input/feedback about main liaisons (usual business)

o       Collect requirements on the web-site image (usual business)

o       MCA asks presenters of Barcelona workshop to send slides to Jacquie for FIPA web site publication

o       FIPA Inform! Newsletter Vol. 5 Issue 1

§         Call for contributions (MCA and Jacquie)

§         Deadline: 31 May 2004

The Image Committee would like to thank Jacquie Kelly for her great work, Stefan Poslad, Mike Kerstetter and all others contributing to the success of the FIPA Image.

TC Interaction Protocols

TC Interaction Protocols did not meet.

TC Ontology

TC Ontology did not meet.

TC Services

TC Services did not meet.

5.      FIPA thanks Sergi Robles and Joan Ametller of Universitat Autonoma de Barcelona for their gracious and generous work in hosting the 31th meeting of FIPA.

6.      FIPA thanks the FIPA Secretariat, Jacquie Kelly for her efforts in coordinating the 31st FIPA meeting.

7.      FIPA invites its membership and contributors to the FIPA process to its 32th meeting, in Rockville, MD.