VIM: A Virtual Multicomputer for Symbolic Applications

Annual Report for 1995

The three topics which bind this network together are parallel systems, advanced compilation techniques and artificial intelligence (AI) with a common substrate in the programming language Lisp. The research aim of this project is to demonstrate the advantages accruing from the combination of these three technologies to build a virtual multicomputer for large scale symbolic applications. A virtual multicomputer is an ephemeral, persistent machine of available heterogeneous computing resources (workstation, shared-memory multiprocessor, distributed memory multiprocessor, array processor). The system supports a virtual processor abstraction to distribute data and tasks across the multicomputer, the actual physical composition of which may change dynamically. Our practical objective is to assist the prototyping of dynamic distributed symbolic applications in artificial intelligence and computer algebra using whatever resources are available (probably networked workstations), but so that the developed program can also be run on more exotic hardware without reprogramming.

Examining the six areas of the work plan (work has not started on items which are not mentioned):

Reflective architectures:
Item 1 is now considered inappropriate. Work on item 2 has continued this year with the formalization of multi-language architectures (of which reflective architectures are a particular case) using a type of Dynamic Logic, namely Descriptive Dynamic Logic (DDL). Furthermore, the application of DDL to the formalization of a Reflective Learning architecture, called NOOS, has been completed. Two publications on this work (from IIIA) are currently in press. There has been good progress on item 3 with the development of a new implementation of the EuLisp object system at Bath and the development of a frame-based system and a blackboard architecture on top of this at Warwick. Work continues on items 4 and 5, but with a shift of emphasis towards integration with the World-wide Web.
Knowledge representation and knowledge bases:
Following on from the work on item 3 last year, work is in progress on item 2. An analysis of how to fill in the gap between the Components of Expertise approach and Milord II has been carried out, and a methodology to resolve it has been elaborated. The results of this analysis have been accepted for journal publication. An application of this methodology in taxonomic domains (marine sponges) has been completed. Under item 4, a multi-agent system for Fish Market auctions modelling is being formalized using Concurrent Dynamic Logic (IIIA) and a first prototype has been implemented using PVM. Under this same area a multi-agent systems generator is under development at URV. Work continues on item 6.
Static analysis:
Item 1 is being approached from a novel direction in a new collaboration between GMD.FIRST and URV to apply a rule-based system with fuzzy-matching to process placement. This technique may subsequently by applied to achieve item 2. Work has continued on items 3 and 4 - the latter using a new rule-based specification.
Semantics and verification:
Work on item 2 has been suspended pending further developments on the system front. Item 3 has been completed. A new direction under this heading has arisen in work at Bath on automatic generation of interpreters from operational semantic specifications and in 1996 we will seek to see how lessons from the ProCos project (Kiel) can be applied in this domain.
Parallel models and systems:
Work continues on items 1 and 3. Work is just beginning on item 4 using several small AI applications and another in statistics in the analysis of large and complex datasets. Work is in progress in the development of a real-time garbage collector for EuLisp supporting the goal of item 5. Although there has been some small transfer from EuLisp to the ISO Lisp process (item 6), this thread is at an end, since the ISO definition has now passed its final ballot prior to becoming a draft international standard (DIS).
Parallel applications:
Extensive work on item 1 last summer (by Cristina Moyes) has not proved fruitful in terms of a parallel application, although it has raised worthwhile questions over the effectiveness of the original approach. Work is due to begin shortly on item 6.

Commentary

Development has continued of the distributed programming system reported last year, in particular, the emphasis on interoperability is making it relatively straightforward to construct bindings between youtoo (the new implementation of EuLisp) and existing packages (Tk widget library, the Message Passing Interface (MPI) standard for distributed processing, and the Multipol distributed data structure library). At the same time, work has continued at INRIA on weakly coherent distributed memory systems, distributed garbage collection and the provision of these facilities as an Applications Program Interface (API). We intend to link these complementary lines of research in the coming year.

Under the headings of Reflective Architectures and of Knowledge Representation, there is considerable interest in, and a shift towards, multi-agent systems, which fits well with the developments on language systems noted above. Specifically, there is activity looking at frameworks for parallel and distributed processing, social issues in multi-agent systems (Warwick), the logical formalism of reflective architectures, based on using Dynamic Logic as a common logical framework to describe and identify the most relevant formal characteristics of Multi-Language Logical Architectures (MLA) and then to investigate the expressive power of the Knowledge Bases that can be built upon them (IIIA), a reflective blackboard architecture (Pisa and Warwick), agent applications in administration (funded by the Belgian Government at VUB), and electronic commerce and electronic negotiation and mediation (GEOMED) funded by the EC Telematics Programme, Information Engineering, DGXIII (VUB, GMD).

As noted last year, the static analysis programme has been seriously affected by changes in personnel at GMD-FIRST. However, there is a positive new development as a result of discussions between GMD-FIRST and URV (Tarragona) where work on fuzzy-matching in rule systems is to be applied to process placement in distributed systems. A secondment from URV to GMD-FIRST is scheduled for spring 1996. We do not expect much of the work identified under the Static Analysis heading to be completed now.

The changes at GMD FIT.KI have had more far-reaching effects than foreseen at the time of the last report. As noted in the 1994 annual report, Knowledge Representation, items 4 and 5 have been affected. However, as noted above, URV have continued independently in the direction of multi-agent systems instead of object-oriented modelling. It seems unlikely that item 5 under Knowledge Representation will not now be achieved. On the other hand, items 1 and 3 under Reflective Architectures have proceeded largely unaffected. In a further loss, Juergen Kopp left GMD FIT.KI at the end of 1995. However, GMD FIT-KI is also partner in the GEOMED project mentioned above (includes VUB) and have also acted as the coordinator of the related feasibility study. GMD now provides the technical coordination for the whole project in the person of Tom Gordon. GMD's contribution to the project is primarily in the development of the mediation services. The GEOMED project is a notable instance of the VIM project having lead to closer collaboration between the partners in areas related to topics in the VIM work programme.

Future directions and further information

The research directions remain as listed in the 1994 report, namely: scripting languages, intelligent software agents on the world wide web, reflection, lightweight delivery systems for symbolic computing, computer supported cooperative working and hypermedia.

Fuller and up-to-date details of activities (past and present) are held on the World Wide Web. In particular, an online repository of project software and information is available at Southampton and is now widely used by the partners. This service has been used for project-related research as well as information dissemination. The repository may be accessed directly with the URL:

http://vim.ecs.soton.ac.uk/
or via the VIM WWW home page:
http://www.maths.bath.ac.uk/~jap/VIM

Number of research staff working in each team of the network

Names and nationalities of staff seconded from one partner to another during the reporting period and the duration of their secondments;

Names and nationalities of additional staff in each team (visiting scientists, fellows) financed by the network contract and the duration of their stays;

For each scientific meeting for which mobility was financed by the contract, date, location, purpose and nature of the meeting, and names of the scientists attending with the name of the team they represented;

21-25 January 1995, Barcelona

Purpose
discussion of parallel and distributed systems, parallelisation of classification, input to the ISO process (joint with ISO working group meeting).
Nature
informal discussions
Participants
Ulises Cortes (UPC), David DeRoure (Southampton), Julian Padget (Bath), Christian Queinnec (INRIA),

January 1995, San Remo

Purpose
attend 3rd Euromicro workshop on Parallel and Distributed Processing (sponsored by IEEE) and present paper on VIM-related research.
Nature
refereed conference
Participants
Duncan Batey (Bath)

6-8 February 1995, Paris

Purpose
discussion of distributed information systems and distributed multi-media.
Nature
informal discussions
Participants
Les Carr (Southampton), David DeRoure (Southampton), Hugh Glaser (Southampton), Christian Queinnec (INRIA).

14-24 September 1995, Pisa

Purpose
discussion of reflection and its application in blackboard architectures
Nature
informal discussion
Participants
Iain Craig (Warwick), Giuseppe Attardi (Pisa), Maria Simi (Pisa)

21-23 June 1995, Naples

Purpose
discussion of distributed systems design and multimedia
Nature
informal discussion
Participants
Les Carr (Southampton), David DeRoure (Southampton), Mario Mango Furnari (CNR), Maurizio Giordano (CNR), Claudia di Napoli (CNR)

1-8 July, Naples

Purpose
discussion of multi-agents and MILORD II
Nature
informal discussion
Participants
Mario Mango Furnari (CNR), Maurizio Giordano (CNR), Claudia di Napoli (CNR), Carles Sierra (IIIA)

14-15 September 1995, Paris

Purpose
discussion of first class environments
Nature
informal discussion
Participants
David DeRoure (Southampton), Christian Queinnec (INRIA).

27-28 September, Berlin

Purpose
discussion of the application of fuzzy inference to mapping parallel programs to parallel architectures
Nature
informal discussion
Participants
Angela Sodan (GMD FIRST), Viçenc Torra (URV)

26-29 September, Edinburgh

Purpose
attend International Workshop on Memory Management
Nature
refereed conference
Participants
Luc Moreau (Southampton), Christian Queinnec (INRIA)

2-4 October 1995, Beaune

Purpose
attend workshop on parallel and distributed systems (PSLS'95), discussions on same with network partners also present.
Nature
presentations and informal discussions. Proceedings including refereed papers from those present will be published in 1996.
Participants
David DeRoure (Southampton), Hugh Glaser (Southampton), Luc Moreau (Southampton), Julian Padget (Bath), Christian Queinnec (INRIA)

27-29 November, Southampton

Purpose
discussion of distributed garbage collection and applications program interfaces for distributed systems
Nature
informal discussion
Participants
David DeRoure (Southampton), Hugh Glaser (Southampton), Luc Moreau (Southampton), Christian Queinnec (INRIA)

Joint publications authored by scientists from more than one of the participating teams;

Geographical Mediation Systems,
GEOMED-F Final Report, 273 p.,
Authored by the Consortium Partners:
 German National Research Center for Information Technology (GMD),
 TNO Building and Construction Research (Holland),
 University Joseph Fourier, LIFIA-IMAG (France),
 Vrije Universiteit Brussel (Belgium)

Julian Padget, jap@maths.bath.ac.uk, this version March 04, 1996