2011-01-12 6:48 pm
Ensuring architectural stability when software gets older (and constructs are adjusted, elaborated...) is still a hard task. At the University of Antwerp, Belgium, a team is conducting research on how to avoid the system-wide propagation of adaptation effects in software, or in other words, how to develop 'evolvable software architectures'. They import the concept of systems theory into software development, in search of a theoretical framework that insures stability.
Professor Mannaert explained the problem at the 2009 IARIA conference on Future Computing. You can find the slides here.
Stable software development examples are not easy to find, but we are able to recognise certain good practices when we see them. I wrote a concise paper on the open source social network framework Elgg in Dutch.