The OO software development process (abstract)

Abstract

A 92 Vancouver, British Columbia, Canada 5 - 10 October 1992 Addendum to the Proceedings Panel Report- The 00 Software Development Process Report by: Dennis de Champeaux HP-Labs Members of the panel Dermis de Champeaux, HP-Lab, moderator Bob Balzer, Information Sciences Institute Dave Bulman, Pragmatics Kathleen Culver-Lozo, AT&T Bell Laboratories Ivar Jacobson, Objectory Stephen J. Mellor, Project Technology Pankaj Garg, HP-Lab, co-organizer of the panel Aim of the panel Improving development processes is a key issue for the software community in general. The literature gives us some guidance with respect to major activities of the development process. The waterfall, spiral and fountain model all agree upon distinctions such as: requirements capture, analysis, design, implementation, testing, maintenance. They also agree in that they do not provide a finer granularity of the inside activities of analysis and design. We postulate that this omission is a key inhibitor to a more widespread acceptance of the 00 paradigm by project managers. They do not know how to plan a non-trivial 00 software development project. The purpose of this panel was to take the first steps towards developing such process models. More 5-10 October 1992 -61- specifically the panelists have been asked to address the following questions (among others): . Can you envision a generic OOA/ OOD development process? If so what are the major steps and deliverables of these steps. If not, why not? l If a software development organization had to choose between either doing a paradigm shift to 00, or improving its rating on the SE1 maturity scale what would you recommend? What are the relationships between the two? Qualify, if necessary, the answer with respect to an extant maturity level. It turned out that these, and the other questions, were quite hard for the panellists. Hence, we envision a follow on panel in the future where these topics will be raised again. In addition to this panel, I organized in 1992, workshops on 00 process: one at ECOOP and one also at OOPSLA. All the events had in common that the presenterslpanelists had hardly anything to say about process. The most tangible contributions were modest extensions of the standard macro process: requirements capture, analysis, design, implementation, test, maintenance, enhanced, of course, by the recommendation to iterate. Addendum to the Proceedings Hence, although the panel had lively exchanges, I refrain from giving the details because I do not feel that the state of the art was advanced. Instead I would like to point the reader to a report that proposes some details on the 00 analysis and design process: Dennis de Champeaux, Towards an 00 SW Development Process, HP-Labs Tech Report HPL-92-149, November 1992. [The contributors Doug Lea, Penelope Faure, and Mohammed Fayad could not be listed as authors for legal reasons.] For a copy contact: HP-Labs/ Technical Publications Dept./ 1501 Page Mill Rd Palo Alto, CA 94304 (415) 857 4573 OOPSLA92 -62 - Vancouver, British Columbia