Software Evolution and the Future for Flexible Software

Speaker: Keith Bennett

29th April 2004 , 2:30pm , Room 518; Claremont Tower

Abstract

Most successful software evolves over its lifetime, often in ways in which the original designers cannot conceive. I’ll start by presenting a brief overview of the state of the art, and summarise how we currently cope with such evolution, drawing on industrial case studies.

The “holy grail” for evolution is to build software which is intrinsically flexible – it is easy to change without compromising dependability (in broad terms, the cost of making a change is proportional to the size of the change, not the system). Software engineers have not been very successful in understanding how to do this, yet the demands to change software in internet time are increasing, particularly in so-called “emergent” organisations. This problem is being taken up as part of a proposed grand challenge in computer science.

I’ll present the results of research being undertaken by the Pennine Group on a project called IBHIS which is exploring radical solutions, with a more demand-centric view leading to software which will be delivered as a service within the framework of an open marketplace. Based on this foundation, I'll then describe recent work, which has resulted in an innovative demand-side model for the future of software evolution. Components may be bound just at the time they are needed, executed, and then the binding may be discarded.

----------

Prof. Keith Bennett
University of Durham

Pennine Research Group **

** The Pennine Research Group is the name of a group of software engineering researchers at Keele, UMIST, Durham and Leeds who have been collaborating for the past 8 years on service-based software engineering. See www.service-oriented.com.