Systems Integration Initiative


Notes of a systems integration programme workshop held at
the London School of Economics on 26th July, 2001

Click here for a larger version of this picture (208KB)


The intention of this workshop was to identify activities in which projects might work together to:

Note that the ambition was not to achieve these subsequent objectives at this first, one-day, meeting, but to begin the process of working together to see how they could be achieved.

Since this was intended to be a planning meeting and since it is too early to bring together all project participants, let alone additional participants from outside the projects in the programme, this was a meeting of representatives of most of the projects in the programme. The intention was not to be exclusive, but to get representation of as many projects as possible. In most cases the principal investigators of the projects had been invited to represent their project. There were about 30 attendees representing 28 running projects.


The agenda is attached as an annex. As background, attendees were given a summary of the projects in the programme, abstracts where available, for each project, and a copy of Bob Malcolm's résumé of the technical directions taken in the programme. These are all available on the web site for the Systems Integration Initiative at here)

After introductions by the host, Eve Mitleton Kelly of LSE, and Bob Malcolm, co-ordinator of the Systems Integration programme, four projects each presented a review of their aims, problems and especially their contribution to our understanding of systems integration.

The intention was not to go into great detail on each of these projects, but to characterise the projects so as to indicate the broad span of the programme and to stimulate discussion about common interests given that broad span. To this end two of the presentations were from the IT&CS community and two from the manufacturing engineering community, and in each case one was focused on technical aspects of integration and the other on strategic concerns.

The meeting then split into several smaller groups to compare their various work and to make suggestions on the way forward for the programme - and the subject of systems integration. Each group gave a brief report on their deliberations. (Their contributions have been merged here.)

Conclusions, recommendations, suggestions...

Note: wherever the word 'problem' appears, read 'potential research topic'.

How far should one go?
Many projects, operating in quite different domains, have run into the same problem. To understand better how to integrate their 'systems of concern' they felt it necessary to appreciate better - and perhaps 'model' - the bigger systems in which their systems of concern are embedded. While this is undoubtedly yielding insights, there is then the question of when to stop. Just how much of the environment should one consider? How should one bound that bigger system?

Integration of soft and hard systems and their models
Next, these wider systems are often 'soft' systems - social systems or socio-technical systems. Modelling such 'soft' systems, and integrating models of soft and hard systems is proving difficult.

Integration of soft and hard systems science
Moreover, the study of these softer systems is usually the stuff of the 'soft sciences' - typically the social sciences such as psychology and sociology. While some projects have attempted to draw on the body of social science, and others are working with social scientists on the research team, the very different nature of the research in the technical and social sciences is proving challenging for their co-operation. It is not enough to 'get an ethnographer on board'.

A Systems Integration Science?
Even before one considers where to draw the boundary of the wider system in which one's integrated system is embedded - or is to be embedded - the problems of modelling that wider system in a way which is susceptible to science are intellectually very demanding.

Various approaches to such modelling are being explored in the programme, including the notion of 'power' in a supply chain, and the measurement and management of complexity.

It was suggested that there should be an initiative to establish a new science of systems integration, addressing - and hopefully ameliorating - all these problems and amenable to measurement.

A conundrum
A number of projects reported that their collaborators consider systems integration to be a key business skill. Yet it is not a recognised discipline in its own right and there is no clear body of knowledge on the subject. There is not even a 'home' for study of it in the research councils. Apart from the present programme, research is scattered through many other disciplines - and indeed across research councils, since the study of management issues overlaps with the scope of the ESRC. (But if there were a research 'home', would the study drift off into inapplicable abstraction?)

Definite objectives counter-productive?
One view was that it might be a bad idea to set firm objectives for cross-programme action - that the benefits of mixing different people from different disciplines are often serendipitous. The problem with this is that it can be difficult to justify the resources for such mixing. (Though we should recognise that 'serendipitous' interaction, in the informal margins of otherwise more structured meetings, is well recognised to be valuable.)

Future cross-programme actions
We must beware special pleading out of self-interest. A group working on a common theme (even as broad as systems integration) is likely to recommend self-sustaining actions to support further work. But with that caution in mind, the group felt it would be a great pity to lose the impetus of the both the programme and the exploration of the issues emerging from the workshop. Two specific recommendations were made - for a (first) pan-programme forum, and for a network to maintain cross-programme communication.

A large-scale international conference was first suggested, in order to establish the discipline. But it was felt to be too early for a (successful) event of this kind, and that first some wider community of interest would need to be established.

There was general agreement that it will be useful to hold a full meeting of as many personnel as possible from all projects in the programme, and possibly with other invitees. This should be sooner rather than later - October or November of this year. The group preferred that such an event should not be structured by specific themes. Instead, all present should have the opportunity to hear about all the projects and to consider further the issues raised in discussion (and reported above).

Three days is felt to be too long - especially to attract the industrial collaborators. One day would be too short for either presentations or discussion. The suggested format is therefore two days. On the first day all the projects would give a short (10 min.?) presentation (in the style of an investment forum rather than traditional conference or seminar paper). The remainder of the time should be devoted to discussion - probably in smaller groups, but with a structure which should emerge from the meeting rather than be predefined.

For the longer term, a network was suggested. Ray Paul of Brunel University offered to host (and prepare a proposal for) such a network, while Peter Henderson offered to investigate the possibility of establishing a mailing list at Southampton.

Note: the nature of the host department should not imply that the network will be oriented to the discipline of that department - it should encompass all the disciplines engaged in the programme. Also note that such a network, focused on cross-programme concerns, should not conflict with existing or future networks arising from particular projects in the programme and focused on more specific technical themes.

The network would facilitate all the usual communication and dissemination actions such as workshops, web site (sustaining the present ideo site), mailing list, and possibly a book series. It would draw out and clarify further research themes such as those hinted at above by the common problems emerging across the projects. It would also facilitate sharing of experience (and best practice?) of research methodologies - with particular regard to 'action research' and the integration of 'soft' and 'hard' science.


LSE was thanked again for their hospitality and EPSRC for its contribution.

Annex 1: AGENDA

10.00 Coffee

10.25 Introductions

10.45 Short presentations by a small selection of projects:

12.30 Buffet lunch

13.30 Small group discussion to suggest options for clustering and interworking (No attempt at this time to structure by topic)

14.10 Groups report back with suggestions

14.55 Open discussion, summary & way forward

15.25 Closing words from EPSRC

15.30 Tea

16.00 Close


Back to EPSRC SI programme home

GO TO .... | ideo home | SII home | EPSRC SI programme home | SII call for proposals | SII proposers' guidance | SII contacts