AUTOSAR lecture

I have been asked by prof. Martin Törngren if I could be a guest lecturer in a course about AUTOSAR at the Royal Institute of Technology.
Sounds really interesting, since I hope to also participate in other parts of the course besides my own lectures. But it is also challenging since the first occasion is already on 9-10 February and I have no material prepared. I need to discuss this with my supervisor before I make a commitment.

Architecting Systems with UML 2.0

The best (and shortest) introduction to UML 2.0 from an architect's viewpoint is an article called Architecting Systems with UML 2.0 by Morgan Björkander and Cris Kobryn.
Obviously you need to know the basics of UML 1.0...

Case study of the Architecture Business Cycle

I'm writing a paper with the working title "A Case Study of the Architecture Business Cycle for a Vehicle Software Architecture" together with a co-author.
We have tried to identify how the Architecture Business Cycle (ABC) looks like for the software architecture in a modern vehicle with software deployed to 30-70 electronic control units (ECUs) connected by a number of multiplexed buses, such as CAN, MOST and LIN.

The main results we present are the benefits the Electronic & Electric Systems Engineering (EESE) department at Volvo Car Corporation has gained from participating in identifying the ABC. Of course there are some benefits also from a research viewpoint as well, for example how well we thought the ABC model worked for such a case study.
The data comes from in-depth interviews performed with 20 persons working at the EESE department.

Yet another conference

My supervisor, prof. Thomas Arts, have found another conference he thought might interest me: International Conference Series on the Quality of Software Architectures (QoSA).
Some of the topics listed in the call for papers are directly related to my research, e.g:

• design decisions and their influence on the quality of software architecture
• architectural standards and reference architectures
• coordination of business architecture, business processes, and software architecture
• traceability of software architecture to requirements and implementation

Unfortunately the deadline for submissions are already 8 February 2009 and it will be impossible for me to submit the paper I'm presently working on till then. Hopefully they might add a work-in-progress or a student session. I could go to the conference anyway, but it is always more productive if I have something to present.

Thomas also proposed I should think about giving an AUTOSAR tutorial on the WICSA conference. He thinks it would not be too much of work for me, but I think he is not aware how ambitious I usually am when it comes to teaching...

Quote

The best quote I have found this week about architecting a system, especially true for future car platforms:

The life of a software architect is a long (and sometimes painful) succession of suboptimal decisions made partly in the dark.

From a presentation by Philippe Kruchten 22 June 2007: Who are the Software Architects and What Do They Do? There is an alternative download available here.
I think the slides were quite good as material for introspection. I would have loved to hear the presentation.

Software development takes over

I stumbled upon a news notice about how software development takes over the traditional product development in traditional Swedish industries such as Ericsson and Scania. Unfortunately it is in Swedish.
The article is published on E24, who claim they are the biggest business site in Scandinavia.

Much of what is stated in the article are things that I have discussed with colleagues and other researchers. But I think it is nice to have an "independent" source to refer to when I make a general statement about what type of business we are doing at Volvo Cars as a vehicle developer.

Feedback on my presentation

The two topics at in my presentation at the software architecture workshop that generated most questions (and discussions in the coffee break afterwards) were

1. The vertical versus horizontal development structure. This did not surprise me since I already knew that is an area of interest (see for example the article Software engineering for Automotive Systems: A Roadmap by A. Pretschner et al. in proceedings of ICSE 2007)
2. The handling of S/W variability, especially in an AUTOSAR context. This came as a surprise, and I certainly need to look into both what research and practical implementation have been done in this area. There has been some work on variability and AUTOSAR in the EAST-ADL2, but I am more interested in the technical implementation in the code than how to model variability.