SPLC 2008 and an invited talk at one of the workshops

Peter Dolog's picture
SPLC 2008 dinner

A week before (8 - 12 September 2008) I attended the Software Product Line Conference 2008 (SPLC 2008). Eventhough I was looking into the software product lines engineering and domain engineering in particular already before (my dissertation was about domain engineering for adaptive web application after all), It was actually for the first time. The main reason for my participation this year was an invited talk on feature based design of web service transaction compensation at the workshop on Service-Oriented Architectures and Software Product Lines. As a part of my talk I also presented shortly some prospects in the direction of variability and SOA in the context of the idSpace project. The workshop was very interesting and interactive. The most interesting thing was especially the participation of industrial partners with their opinion as well as an interactive brainstorming afternoon session where several issues such as design, dynamic adaptation vs. static design time binding, and others have been discussed from various perspectives. It seems to me that SPL community is interested in what the SOA brings to them, i.e. SOA as a new architectural style, while to me as more web engineering oriented person, it was interesting to see how various variability modeling approaches can deal especially with run time adaptation as well as how open architecture and autonomy fits to the product lines.

I also attended the main conference, where I enjoyed both keynotes: basic research oriented from Dave Parnas on program families and their relation to software product lines as an rather industrial and consulting concept according to David as well as Luc Koch on experience of Philips with software product lines concept. Interestingly, Luc mentioned that Philips is looking into SOA in the context of their product lines as well.

I also enjoyed very much the highly interactive panel on relation of agile software development and software product lines with group work and brainstorming. I also participated in one of the work groups in the panel and it seems to me that while some of the techniques from agile approaches can work in some stages of software product lines engineering, generally, the approaches are quite different eventhough targeting the same issue (changes, management, knowledge). I contributed to the discussion in the workgroup with a comparison of SPL to one of the iterative software development methods: EVO. The method is described in the book of Craig Larman I use in my software engineering course. The EVO method interestingly features similar distinction as software product line engineering between in activities which are carried out with customer and activities which are carried out as development for later use at a customer side. In EVO the two groups of activities are called "front room activities" (with customer) and "back room activities" (development for later use). In product lines engineering, they are called "application engineering" (building an application as a member of a family for a customer) and "domain engineering" (building reusable assets for later use). Eventhough the organization and management of the activities seems to be different in these two approaches, there is an interesting relation perhaps worth of studying. Another thing, which caught my attention was that the large companies which are here for already number of years and are successful have sooner or later adopted an engineering approach which is in fact software product lines engineering approach and they report on success especially due to the fact of having implemented the software product lines concept. I do not know whether we can say the same about agile approaches, but perhaps more longitudinal studies are necessary here. The SPL conference series actually supports the researchers of software product lines with an interesting concept of hall of fame where those successful cases from industry are reported, archived and available for future research purposes.

Last but not least, the conference had a number of interesting contributed papers presented in various sessions, so I think it was definitively a good invested time from professional perspectives. Besides that, a number of discussions, interactions I had, and social events (see picture from the main dinner) complemented the program and helped to have the SPLC 2008 as a very pleasant and enjoyable event eventhough it was raining almost all the time.