Wednesday, September 2, 2009

13th International Software Product Line Conference (SPLC)

Now in its 13th year, the International Software Product Line Conference (SPLC) 2009 was held August 24 – 28 in San Francisco, California. In this edition, the presence of industrial and academy people was remarkable, resulting in more than 150 attendees. Practitioners and researchers from different countries had the opportunity to discuss hot topics regarding to Software Product Lines area.


During the first two days, current SPL topics were discussed through different workshops and tutorials. Some workshops were realized at the first time, which indicates the emergence of new hot topics, indicating the need of investigation and study. For example, 1st workshop on consolidating community consensus in product line practice, which counter with the presence of different companies (SEI, BigLever, Yahoo, IBM), and 1st international workshop on Model-Driven Approaches in Software Product Line Engineering. Although the workshop on Dynamic Software Product Lines (DSPL) is in the third edition, this edition was marked by the research and practitioners discussion that did not achieved a consensus regarding to the concept of dynamic SPL, which highlights the lack of maturity in this field. They were discussing the difference between Dynamic SPLs and Single systems, since to dynamically change the code the software needs to be completely developed and integrated. The workshop Scalable Modeling Techniques for Software Product Lines (SCALE 2009 ), have Tomoki Kishi and Kyo Kang, and they start showing the problems when modeling scalable systems, for example, size, complexity, distributed environment, long life cycle and multiple stakeholders. In SPL context the main problems are variability management, configuration management, model based development and complicated development style. Later, the Service-Oriented Architectures and Software Product Lines (SOAPL) which was marked with the presence of the RiSE group represented by Flavio Medeiros, presenting the paper entitled "Towards an Approach for Service-Oriented Product Line Architectures".

On Wednesday, the first keynote speaker, Richard Gabriel from IBM discuss the about "Science is Not Enough: On Creation of Software", which presented a different perspective in the field. His talk was a mix between a talk and lecture including the strong relationship with arts. During this day, papers regarding to configuration, scoping and variability were presented and discussed.

On Thursday, the second keynote speaker, Jacob G. Refstrup - lead architect for the Owen software product line architecture from HP - which is used across multiple inkjet product families. He talked about "Adapting to Change: Architecture, Processes and Tools: a closer look at HP's Experience in Evolving the Owen Software Product Line".

On Friday, the last keynote speaker, Kyo Chul Kang presents the "FODA: Twenty Years ofk for feature analysis and simple but comprehensive way to modeling commonalities and variabilities. He also report the remaining problems of it, for example, addressing other parts of the life cycle (especially application engineering), clear the mapping between Perspective on Feature Models". Kyo shows that FODA has approximately 1300 citations, highlighting the importance of it. He presents as main contributions of FODA: systematic domain analysis, lying the groundworfeatures and software artifacts, standardization of feature model extensions, trade-offs between expressiveness and simplicity, scalability of feature model, managing complexity in view of many inter-dependent features, feature model evaluation and integration with UML model.

During this day the Goldfish Panel "How to Maximize Business Return of Software Product Line Development", had the presence of the most renowned researchers in the area as John D. McGregor, Dirk Murthig, David M. Weiss, Klaus Smith, Jan Bosch, Charles Krueger and Eduardo Almeida, which put in discussion the Brazilians company reality, reporting that they have not a specific domain and asking how is the best way to introduce the SPL approach in these companies. Others issues to think about were discussed as: Is it worthwhile at all?, What are the three top-most value-generating activities?, How to recognize where to focus efforts? (subsystem, functionality,..) and What was your worst ever product line experience?

Finally, the waiting hall of fame was presented by David Weiss showing the results of the evaluation from 2008 and presented the candidates for 2009. Two new companies were trying to have her names in the hall of fame, the TomTom company presents a GPS software product line and Lockeed Martin . In the end, the TomTom was nominated and will be evaluated during this year. The results will be available in the next SPLC 2010 held in Korea.

2 comments:

Eduardo Almeida said...

Moreover, it is important to highlight the session dedicated to the future in the field.

Important topics were identified:

1. Variation (AOP | SOA | End-user programming)

2. Automation (Production Process - MDA, DSL, Generative Programming)

3. Lowering adoption cost (Agile | Tool Support)

4. Distributed Development and Evolution (Open Source Models | Colaborative Environment | Virtual World)

5. Ultra-Large Scale Systems

Eduardo Almeida said...

Some data from the conference:

174 participants (51% from academy and 49% from industry)

54% first time in the conference

83 submissions (30 accepted papers)