- The development of the base software for the virtual
multicomputer and the incorporation of suitable libraries for
graphical interfaces (via Tcl/Tk) and message-passing (MPI) has now
been completed to the extent that the first full release of the
system, called youtoo, has just been made. During the past summer a
first version of a distributed application modelling the Spanish
fish market was built as a result of a young researcher from IIIA
visiting Bath. In the light of that work, we will be constructing a
second version with the assistance of another young researcher from
IIIA visiting Bath from October to December. Although several aspects
of this line of work are beginning to be turned into publishable
material, the practical work that remains to be done will still take
at least another 6 months and preparation for publication of this
further work, another 3 months.
Active partners: CSIC (IIIA), Bath.
Status: work started, but currently dormant
- Given the state of development of youtoo (see previous paragraph),
we are only now in a position to tackle one of the items in the
project's original programme of work, namely the re-implementation of the
NOOS system. This will enable the deployment of NOOS in a distributed
environment, as well as being a major application for the youtoo system.
It is expected this will take at least six months from our intended
start date in mid-November.
Active partners: CSIC(IIIA), Bath.
Status: implementation of NOOS in youtoo complete,
demonstrators needed
- The network has begun preparing papers for a special number
of Artificial Intelligence Communications, the journal of the European
AI society. Initial preparation of the papers could take as little as 3
months, but type-setting and publication a further 3 to 6 months.
Active partners: Bath, GMD Bonn, CSIC(IIIA), Pisa.
Status: in progress
- As a result of the development of the youtoo system, new work
began this summer, involving Kiel, Southampton and Bath, on the
specification of the operational semantics of language interpreters and
their automatic generation and their subsequent verification. An
especially interesting, and novel, aspect is the specification of
storage management as part of the language's operational semantics.
Consequently, we hope to be able to generate complete correct language
implementations (compiler, run-time system and storage manager) from an
operational specification. It is not expected to be able to finish such
a major item of work within any reasonable extension of this network,
but it seems appropriate to mention this work in the context of a case
for an extension, since without an extension, this collaboration would
probably lapse until further funding could be found.
Active partners: Kiel, Southampton, Bath.
Status: dormant
- A concurrent FPIV project (GEOMED) involving GMD Bonn and the Free
University Brussels makes use of similar technology to that being
developed in this network, namely agents. Now, both projects are in a
position to assess one another and it is important that we take this
opportunity for cross-fertilization that GEOMED and VIM offer us. In
particular, the focus on mediation services for GEOMED presents an
interestingly similar but different scenario to the auction process of
the Spanish fish market. We now need to compare communication protocols
to identify higher level patterns which may have implications for
electronic commerce in general. Again, we foresee this work taking a
further 6-9 months, partly because it involves some of the people
working on the first item.
Active partners: GMD Bonn, CSIC(IIIA), VUB.
Status: unknown
- A young researcher, who will shortly be starting as a research
student at Bath, would greatly benefit from spending three months at
each of VUB and GMD, in order to focus his studies in the area of
robotics. This cannot begin before January next year, due to the
student's present commitments.
Active partners: VUB, GMD Bonn.
Status: student currently at VUB
- A fuzzy inference system was developed jointly by GMD FIRST and
URV (Tarragona) to construct appropriate configurations of a set of
potential strategies for mapping programs to parallel machines. This
is a novel application of fuzzy inference - and has also led to new
techniques in fuzzy inference. This has been made possible by the VIM
project bringing together the know-how of two groups working in
completely different fields. The publication of the preliminary
results is in preparation, and an extension of the network would
ensure that the work and the publication can be finished. (The group
in Tarragona has no budget to continue this research otherwise.)
Furthermore, concrete ideas exist for extending the work in the
direction of distributed systems where configuration is of high
significance, too. The VIM project would not only provide the
necessary funding, but also the important technical framework of
several partners working on distributed systems - and thus also allow
further integration of the partners' work done during the last 3
years.
Active partners: GMD Berlin, URV Tarragona.
Status: unknown
- Further development of youtoo and distributed applications is
planned by way of the primary developer - a young researcher in Bath -
making extended visits to Pisa and to Bellaterra (CSIC/IIIA) in the
first six months of next year. In the first case, work will focus on
transfer of expertise in delivering symbolic applications, portable
storage management tools and experience with an alternative operating
system (Windows NT). In the second case, work will focus on the
continued development of fish market models and the NOOS knowledge
based system.
Active partners: Bath, Pisa, CSIC(IIIA).
Status: Andreas Kind visiting Pisa
- Finally, in order to bring the program of the network to a
satisfactory conclusion, we plan to hold a workshop during the first
half of 1997 to showcase the results of the project, with invited
speakers who are notable workers in relevant fields and scholarships for
young researchers from outside the network to attend. It is intended
this will result in the publication of a proceedings as a book, probably
in the second half of 1997. Further publications should appear in the
same time frame as a result of research currently in progress. Thus, a
final report in October 1997, should be able to identify a significant
number of joint publications which are a result of the existence and
actions of the network.
Active partners: all.
Status: in progress