ACTIVITY REPORT
Prof. B. Cohen visiting University of Lisbon

The purpose of this visit was to enable an exchange of ideas on categorical approaches to enterprise modelling.

May I first of all express my appreciation for the hospitality extended by the University of Lisbon to me and my family. I thoroughly enjoyed the meetings I had with staff and students and the opportunities to discuss the issues raised by my somewhat eccentric approach to the modelling of enterprises, which is based on the empirical techniques and clinical praxis developed by my colleague, Philip Boxer [BC98].

My original objectives were to present these ideas for the critical appraisal of University of Lisbon staff, and to gain a deeper appreciation of how their work might influence them.
 
The first objective was achieved by the two seminars I delivered (‘Services and Agents’ and ‘Agents and Enterprises’ — copies of slides available on request) and by subsequent discussions, as follows:
•  with Antonia Lopes, I discussed ‘feature interaction’ as a particularly concrete and recognisable manifestation of certain enterprise problems that relate to, but are wider in scope than, those to which the ‘superposition’ technique of COMMUNITY has been applied;
•  with Helder Coelho, I discussed relationships between the articulations (existential, referential and deontic) of Boxer’s approach and the values (intentions, beliefs, desires and expectations) of the ‘mental states’ approach;
•  with Luis Fiadeiro, I discussed the significance of the deontic articulation to a category theoretic setting of Boxer’s model;
•  I also had interesting discussions with several of Helder’s and Luis’s PhD students and, thanks to Helder’s introduction, with A. G. Portela;
•  Felix Costa provided me with a copy of his paper on compositional reification in response to my comment on the failure of traditional ‘refinement’ to address both horizontal and vertical composition.
I regret that I was unable, in the time available, to provide more detailed case study material from Boxer’s field work but hope to rectify that omission in future.

As to the second objective, I am now particularly attracted by the possibility of a category theoretic treatment of the ‘articulations’. I took the opportunity to refresh my understanding of category theory (to which I was introduced nearly 20 years ago but have never applied), using the excellent introductory text that Luis supplied. The models that I had been constructing (so far, only of the lowest levels of the existential articulation) are compositions of relations expressed in set theory and constrained by invariants ­ a structural framework not far removed from, but much less powerful and elegant than, category theory. Although I have yet to recast them categorically, the benefits of doing so, and the insights and challenges that might provide, are becoming clearer.
Two papers currently in preparation [C99, BC99] have already benefited from these insights and I look forward to further correspondence with Lisbon as this work progresses.

References
[BC98] Boxer P, Cohen B. Analysing the lack of Demand Organisation. Proc. Computing Anticipatory Systems, Daniel M. Dubois, ed. Conference Proceedings 437, pp157-181, American Inst. Phys., Woodbury, NY, 1998.
[C99] Cohen B. Being Served, the Purposes, Strengths and Limitations of Formal Service Modelling. Based on a presentation at ECOOP 98 and to appear in a forthcoming Kluwer book, edited by H Kilov.
[BC99]  Boxer P, Cohen B. Doing Time: The Emergence of Irreversibility. Based on a presentation at the Gent Closure workshop in 1999 and to appear in the published proceedings.