I’ve mentioned before that the International Lisp Conference will be taking place here in Cambridge from 1st - 4th April, 2007. The Call for Papers is now available.

Deadline for abstract submissions: December 15, 2006
Notification of acceptance or rejection: February 5, 2007
Deadline for final paper submissions: March 2, 2007

We’re taking part in next week’s Working Meeting on “Models and Aspects for Product Families: A Roadmap for Research” at the University of Lancaster. Danilo Beuche of pure-systems will also be there, and between us we’ve put together a presentation titled: “Aspects and Models - the tool vendor’s perspective”.

Abstract: “As tool vendors we are interested in providing solutions that benefit our customers in some way. Such benefits can potentially be realized in the combination of Aspects and Models. Since our core product is a Feature Modelling toolset for Software Product Line development, our main area of interest is naturally the combination of Aspects with Feature Modelling.

Here we have identified a number of potential benefits and a number of problem areas. However, even if these benefits can be realized and these problems be solved, there still remains the issue of building an economic case for tool adoption. The adoption process should ideally be reasonably straightforward for organisations, and the benefits of using the tool need to be predictable, and measurable, in order to justify the customer’s purchase and on-going usage.”

The purpose of the meeting is to prepare a roadmap for future research in this area. Hopefully this will include further work on combining Aspect-Oriented Development with Feature Modelling…

We’ve decided to exhibit at the Telelogic UK Conference next month (13th-15th November).

Requirements Management of Software Product Lines is a big issue due to the difficulty of capturing the relationships between different requirements for different variants. For example, you may find it difficult to specify that some requirements are common to all or some variants, or it may be impossible to say that some requirements are optional, alternative or mutually exclusive. We’re hoping for a lot of interest in pure::variants and our pure::variants Synchronizer for DOORS® add-on which addresses these problems.

If you’re at the conference, come along to our stand and say hello to myself and Danilo Beuche of pure-systems who’ll also be participating.

(DOORS is a registered trademark of Telelogic AB)

Today seemed busier than yesterday, perhaps because of the more attractive conference themes, or maybe the better weather. However, I am getting ahead of myself.

First off I dropped by the Phaedrus stand to say hello, Chris Hills seemed in fine spirits but I passed on his offer of coffee and doughnuts - a regular feature of their stand I’m told.

Then it was down to work, or at least posing for a photo before the exhibition started… You can click on the stand picture for a larger version if you are that way inclined.

This morning was definitely busier than yesterday morning, with several people coming to the stand to find out more about pure::variants and to see whether it could address their variant management problems. There was probably also a similar number of people who were just curious about what variant management is.

During lunch, Andrew Watson of OMG dropped by. He told me that the model-based development conference stream he was chairing has been particularly successful this year, and this can, at least partly, be credited to Andrew’s efforts over the years to generate interest in this approach in the UK embedded community.

The afternoon was quieter so it gave us a chance to reflect on how we could do things better at ESS next year.

So, it’s all over for another year, but I feel that our experiment in having a stand here this year has been successful. However, one cannot tell until we see whether our higher visibility is converted into paying customers…

Thanks to the organisers for making things pretty smooth for us newbie exhibitors at this event.

Regular readers will know that we’re exhibiting at ESS 2006 for the first time this year. We arrived yesterday and set up the stand - with a few minor glitches - and were ready to go at 09:30 this morning.

Surprisingly, for me at least, people were in the door at this time and gradually filtered through the bigger companies to make their way to our stand at the back right of the exhibition, and, it turned out, it wasn’t until around 15:30 that the number of visitors tailed off.

We had a diverse range of people come up to us and we had some fairly in-depth chats on variant management and how pure::variants addresses this problem. We were also able to introduce some people to the concept and the fact that there is a tool that now solves this problem.

I expect tomorrow to be slightly busier than today as there is stream on model-driven development and attendees at this stream may have a stronger software interest and may also be more receptive to the idea of using feature modelling for variant management.

One day down, one to go.

A recent ZDNet Asia article has caught my attention:

The article begins by noting that “[SOA] is a design approach where individual business services, such as authenticating networks or pulling up customer information, are used in different applications”. The aims of this approach are to make the organisation’s systems more cost-effective and responsive.

One architect is quoted as saying that they measure “the success of its SOA project by how much software reuse is happening and how quickly the company can offer services to customers”.

Ever since I had a discussion with Ashley McNeile at SPA 2005 after I’d participated in his SOA session I’ve hypothesised that the Software Product Line approach could contribute to SOA. This was further reinforced after reading a couple of CBDI Forum articles on “Applying Product Line Techniques to SOA” earlier this year (membership required to access articles), and now in reading this ZDNet article.

However, there is a lot of hype around SOA, and I do find it very difficult to sort the wheat from the chaff. I’d be interested to hear from anyone working in this area to try and validate the hypothesis that Software Product Line engineering techniques can help realize Service Oriented Architectures.

Software Product Line Engineering doesn’t get a lot of mainstream coverage so it’s nice to see that the Fall 2006 issue of Methods and Tools has a couple of articles on this very topic.

The first article is a good overall introduction to the Software Product Line approach. The second covers incremental adoption of Software Product Lines at Engenio Storage Group and shows that Software Product Lines can be adopted in months not years.

If this arouses your interest then watch out for Winter’s Methods and Tools that will contain an article on managing your Software Product Line with Feature Modelling written by Dr. Danilo Beuche of pure-systems GmbH and yours truly.

The Embedded Systems Show 2006 takes place on Wednesday and Thursday next week at the NEC in Birmingham. As I noted earlier, we’re experimenting with having a stand there this year. Needless to say this has meant a lot of work in terms of letting people know we’re going to be there, designing the stand, printing leaflets etc. as well as the usual administration involved in getting people to the show and getting them a bed for the night.

Looking at the latest exhibition plan I see that the exhibition space is now full, and also that some of the big(ger) names in embedded systems development are now on board which is great.

However, none of them will be demonstrating pure::variants, the leading tool for embedded software variant management, so do come along to stand 741 for a demo or a chat!

Admittance to the exhibition is free and at the time of posting you could still register here or you can always turn up on the day.

There’s also a concurrent conference with streams on dependable systems, DSP development, Model-Based Development (I blogged on this stream here) etc. but at this stage you’d have to contact the organisers if you want to register for that as I don’t know whether it’s sold out or not.

Thanks to Wiley, The Code Generation Network’s October competition prize is a copy of “Model-Driven Software Development: Technology, Engineering, Management” by Markus Völter, Thomas Stahl, Jorn Bettin, Arno Haase, Simon Helsen, Krzysztof Czarnecki (Foreword by), Bettina von Stockfleth (Translated by)

See http://www.codegeneration.net/tiki-read_article.php?articleId=84 for full details.

The competition will run until October 31st 2006.

SPLC 2006’s opening keynote was given by Carliss Y. Baldwin of Harvard Business School and looked specifically at the Option Value of Software Product Lines. This was a talk that touched upon Software Product Line economics and so it is worth being familiar with the topic matter when considering the business case for Software Product Line adoption or evolution. (NB I am not an economist so check out Carliss’s slides here or her book with Kim Clark, Design Rules: The Power of Modularity v. 1, for a definitive version.)

Carliss began by noting an important historical distinction between the computer industry and the automotive industry. Between 1980 and 2004 the computer industry switched from a market dominated by vertical organisations that produced everything from chips to systems to software to a market dominated by horizontal organisations that produce a subset of solutions that need to be combined to produce a total solution. In contrast, the same 20 year period saw little change in the organisation of the automotive industry with most of the major players being (nearly) vertical organisations, at the start and finish of the period.

Carliss and her fellow researchers attribute this in part to the modularity that is inherent in the computer industry i.e. the degree to which the architectures of computer systems (in the largest sense) are partitioned into components that are (mainly) independent.

The differences between the two industries is explained by noting that design activities (such as modularisation) create option value (money or the promise of money) and such value works as a force in the wider economy. Given sufficient option value, such designs can change an industry as they did in the computer industry by making it easy and profitable for new entrants to come to market and capture value.

To further explain this Carliss introduced the terms manageable designs and unmanageable designs. Unmanageable designs are modular but, importantly, have very high option potential i.e. have a high option value of redesign. Manageable designs have low option potential and so value stays within the supply chain; with unmanageable designs however, value escapes from the supply chain into other organisations. Furthermore, the situation is unstable, since it is easy and profitable for other new entrants to come along and change the market further.

So, what’s all this got to do with Software Product Lines? Well, the answer is that a move to Software Product Lines may create an unmanageable design architecture. If you modularize a previously integrated system some of the modules may end up having high option potential and there will then be an incentive for new players outside your supply-chain to capture value by redesigning those modules. On the other hand, if you’re already in a domain that behaves like the computer industry then you should note that the economics of the situation mean that there will be a shift from integrated solutions to platform-module systems and so you should plan for that shift.

So, how do you make money when unmanageable designs are the name of the game. Carliss identified several strategies including:

1. Own the platform, not the modules – e.g. Ebay.

2. Use Mergers and Acquisitions to be the Lead Firm – buy or merge with the winners after they’ve emerged.

3. Use Design Architecture to reduce your footprint – create a lean organisation and product set.

4. Use open source to replace platforms you don’t own – e.g. promote Linux, Eclipse.

One final thought – The AUTOSAR initiative aims to create an open standard for automotive electronics. Given the increasing value of electronics in modern vehicles will the automotive industry go the way of the computer industry if this initiative is successful?

A full report, including this tutorial report, will appear on the Software Acumen website shortly.