Software Process Modelling in Practice Kensington Town Hall Conference Centre, London, UK 22–23 April 1993

Software Process Modelling in Practice Kensington Town Hall Conference Centre, London, UK 22–23 April 1993

Conferencereport Software Process Modelling in Practice Kensington Town Hall Conference Centre, London, UK 22-23 April 1993 The Software Process Model...

145KB Sizes 0 Downloads 28 Views

Conferencereport Software Process Modelling in Practice Kensington Town Hall Conference Centre, London, UK 22-23 April 1993 The Software Process Modelling in Practice Conference, sponsored by Information and Software Technology, attracted approximately 65 delegates to Kensington Town Hall Conference Centre. My own mental image of a Town Hall is that of a glorified hut with a few rickety chairs, but thankfully this preconception turned out to be wildly inaccurate. In fact many of those attending expressed their satisfaction with the venue, particularly with the Council Chamber and its impressive audio conferencing system. The make-up of the two days' events was to have two long, wide-ranging sessions on the first day, with shorter, more specialized sessions, on the second. The first morning was given over to a tutorial on the technical aspects of process modelling, by Bob Snowdon of ICL. This gave some historical context, attempted to clarify definitions and terms, and examined some notations used for process modelling. This tutorial would be particularly useful for those new to such concepts, though it also sparked some healthy debate about the merits of the definitions of process terms. The afternoon session of the first day was taken by Bill Curtis, who concentrated on the Carnegie-Mellon Software Engineering Institute (SEI) Capability Maturity Model. This was an extremely impressive presentation, and as one who has had reservations about the maturity model, I ended the afternoon feeling much more positive about the work at the SEI. The second day consisted of eight 45-minute sessions. The topics covered were: (1) The superior software organization: Bill Curtis This talk examined some of the evidence for the successes of process maturity, and also discussed some of the insights gained from the behavioural/cognitive work by MCC and others. Key points raised were the idea of design as an 'opportunistic process', and the importance of application domain knowledge. (2) Software process maturity and the Information Systems developer: Ken Thompson

Vol 35 No 5 May 1993

(3)

(4)

(5)

(6)

(7)

(8)

Ken Thompson looked at the differences between Information Systems (IS) developers and the target users of the Capability Maturity Model. In particular he noted that Information Systems development may only be a small part of a larger, and not software oriented organizational culture, and that both the development and the nature of the actual projects undertaken may differ considerably. He then detailed experience in modifying and applying process maturity to IS developers, taking into account these differences. Process assessment standards: Alec Dorling This talk introduced SPICE (Software Process Improvement and Capability dEtermination). This is an International Standards initiative to produce a consensus utilizing the best features of existing methods. Process modelling and metrics: Dieter Rombach Rombach's main theme was 'Measurement based guidance through explicit project plans'. An important aspect of this talk was his explanation of the need for measurement and modelling to be recognized as complementary techniques. Process assessment--The 'Bootstrap' approach: Gunther Koch Bootstrap is a European Commission project which has produced an alternative maturity scheme with an emphasis upon organizational issues. Process centred environments: Christer Fernstr6m Fernstrrm discussed the idea of a generalized process support system, illustrating these ideas by reference to a specific example: Process Weaver. A comparison was made between database, process centred and service oriented environments, with the latter kind being described as 'the future' direction for process support. Process engineering within British Telecom: Alan Pengelly Pengelly described a pragmatic and successful approach to process modelling undertaken by British Telecom. Software measurement was cited as the motivation for the work, and was regarded as an important complementary activity. Process modelling prospects: Manny Lehman Lehman gave a controversial talk in which he examined the weaknesses of purely graphical descriptions, and went on to explain why process programs cannot be used to control the software process.

317

Conferencereport The highlight of the conference for myself was undoubtedly Dieter Rombach's illuminating session. Aside from detailing the complementary nature of modelling and measurement, Rombach also gave a number of valuable lessons learned (from process modelling experience). These 'lessons learned' included the need for measures to be attached to models, starting with descriptive modelling, automated support for data collection, guiding not controlling, dynamism in models and involving process owners in data interpretation. However, I was also not alone in being impressed by the clarity and sense apparent in the talk by Ken Thompson, or the strength of the arguments put forward by Bill Curtis. The conference was an excellent opportunity for academic and industrial modelling practitioners to discuss their views. On the face of it, there seem to be a number of distinct camps: those advocating process maturity as the basis for improvement; those who see automated process support as the way to go; and those who wish to use relatively simple techniques to better understand their process. Nonetheless, there were some areas of

318

general agreement that were raised by a number of speakers. For example, it was recognized that we have enough tools and languages, already, and that it is the understanding of process which we lack. My only real criticism of the two days would be that there was too much emphasis given to capability maturity and evaluation, which took at least half of the allotted time slots. My expectation from the title was that there would be far more on practical experiences of using process modelling methods and notations. However, the quality of the speakers was the real attraction of this conference, and in that respect it did not disappoint with the standard of presentation and content being very high.

Keith Phalp Bournemouth University, UK Note: A Special Issue on Software process modelling in practice, including most of the papers given at the conference as well as papers submitted in response to a Call for Papers will be appearing as a double June/July issue of this Journal.

Information and Software Technology