The 1st Conference on Model-Driven Development and Product Lines will be held in Leipzig, Germany from 19th-20th October 2006. 

See the full programme (in German) and register at http://www.software-families.org/.

If time permits I’ll hopefully be able to put together a short report on the conference afterwards as our German partners, pure-systems, will be participating.

A recent McKinsey Quarterly article (registration required) “The next generation of in-house software development” describes how some pioneering banks, pharmaceutical and media companies have taken a lead from software product developers and turned elements of their custom-application support into packaged products that can be reused in multiple applications.

The headline benefits they’ve reported include:

• A 30% reduction in the cost of managing applications
• A 60-80% reduction in application development times (from conception to deployment)

Although it doesn’t explicitly use the term, the article describes how these companies are using a Software Product Line approach, where a core set of reusable assets - code, documentation, processes, are developed once and reused many times in multiple similar applications. An example is given, where around 60% of a company’s applications accounting for 80% of its application budget are addressed by only 5 product lines (which they term archetypes) with an increase in these percentages predicted as new applications are required.

One final point the authors make is that this approach is best suited to in-house development organisations which have a rapidly-changing portfolio of applications such as those in banks and media companies, rather than those in businesses where change is slower.

In an article for BCS, Nick Rozanski describes the job profile for a typical Technical Architect role and gives some pointers if you’re interested in moving into such a role.

Nick is an experienced architect and one of the authors of Software Systems Architecture: Working With Stakeholders Using Viewpoints and Perspectives and a nice bloke too.

We had our first SPA Cambridge pub trip last night and I think it’s fair to say we had a quality rather than a quantity turnout

Aside from being a chance to sample the Fort St. George’s beer offerings and update each other on what we were up to we did do some work on future sessions for SPA Cambridge and had some general discussion on software. We’ve invited a number of people to speak over the next few months but here we were looking a bit longer term and thinking a bit more creatively - possibly because of the surroundings.

Subjects that came up included:

• Program verification - probably too heavy a topic for SPA
• Robot Lawnmowers - one of the group has acquired such a device and was keen to (have a) talk about it. Sadly these things are not programmable which severely limits their desirability in my view.
• Embedded systems development - a session on this for non-embedded developers or conversely a session on non-embedded development for embedded developers
• Microsoft research - some sort of overview of the activities at MS Research in Cambridge, or some specific piece of research e.g. something on F#

Here is a flavour of the general software topics that were discussed:

• Why does government keep planning massive IT projects that seem doomed to failure. Is it because of the need to be seen to be doing something? To generate headlines?
• What is an enterprise architecture? Is it sensible to have a single such architecture for a federated organisation where each sub-organisation does things differently?
• In the US, companies seem to hire-in contract process improvement specialists but the trend in the UK is to employ them directly. Why should this be the case?

The next SPA Cambridge meeting will be on Wednesday September 13th. Eoin Woods will run a session entitled The Past, Present and Future of Software Architecture. Visit the SPA Cambridge website for more details and to preregister.

Back in April I noted that I was thinking of adapting Enthiosys’s Buy a Feature game to be used in the context of a Software Product Line scoping tutorial I’m running at SPLC 10.

I’ve been busy the past couple of weeks completing my slides for this session and refining the game concept into something that should be enjoyable as well as educational.

Although the idea of playing a game as part of learning may be alien to some people it does seem to be a valuable learning technique and also seems to be becoming increasingly popular in some parts of the software field.

In case anyone is thinking of participating, but is unsure about something so radical, I can assure you that much of the rest of the session will be pretty conventionally-based on slides and very short exercises. Most people should find the game itself pretty straightforward to grasp and game-play shouldn’t take up more than around 30% of the session.

Hope to see some of you there. It should be fun.

A couple of things I heard recently got me thinking about the question of the life span of a programming language.

I think it was Andrew Watson, at his recent talk on Modelling for Maintainability, who mentioned that programming languages have historically become obsolete about every 15 years or so. The second comment was in a Software Engineering Radio podcast where the speaker noted that some classes of developer had already moved away from Java to Ruby.

Now Java is still immensly popular, but this anecdotal evidence, coupled with long-term trend figures from the TIOBE Programming Community Index suggest that it is in decline, along with other languages such as C, C++, Perl and Delphi.

I know at least one person who’s made the move from C++ and is now planning to program purely in C# until he retires in 20 years or so but I wonder whether this is a realistic possibility.

Will the pace of change accelerate such that programming languages become obsolete significantly faster? Or, as program sizes increase, will it simply become too expensive to contemplate moving to another language and we’ll be locked into current languages in some sort of QWERTY phenomenon?

Alternatively, will a move to model-driven software development take place that will enable faster obsolescence of programming languages and indeed make programming languages less relevant?

Answers on a postcard…

The call for session proposals for next year’s Software Practice Advancement conference - SPA 2007 - is available. 

The organisers say, “We are seeking sessions that explore emerging practices that software teams can leverage in their project work. We particularly value sessions that create an environment for participation and exchange of ideas.”

Topics and themes are pretty broad including novel systems structures, what really works, evolving systems, technology, people and process. Session types include the familiar workshops, tutorials, case studies, think tanks and working groups.

I’ll also bring some paper copies of the call to the SPA Cambridge pub trip for anyone who’s interested. The 2007 conference sees a new venue in Homerton College, Cambridge and it would be nice to see more participation from SPA Cambridge attendees.

As a presenter at the past two SPA conferences I can vouch that leading a session is a rewarding experience, SPA delegates are well-informed, will ask hard questions and will help advance your understanding of your session topic.

The closing date for submissions is 11th September.

After the success of last year’s miniSPA, the British Computer Society’s Software Practice advancement group is again running this free one-day event to showcase the best sessions from its annual conference SPA2006.

miniSPA 2006 will be held on Friday 11th August at the BCS Davidson Building in central London. I’ll be going along and plan to participate in James Dobson, Richard Mitchell and Duncan MacGregor’s sessions all of which I missed at SPA2006. I can also recommend Chris Cooper-Bland and Rob Machin’s session on Requirements and Architecture Traps and Pitfalls which I did participate in at SPA2006.

Hope to see some of you there.

I just noticed this quote from Paul Beardow of Superscape in an article last year on the Sony Ericsson Developer World site:

“if you want your game to run on the widest possible range of devices the design has to accommodate different assets and mechanisms to enable optimal builds for each handset. One of the ways that we plan for this is by building for ‘families’ of handsets that have similar capabilities and performance.”

Last year I wrote a whitepaper arguing that adopting Software Product Line engineering could be beneficial for some mobile games companies. Paul’s quote seems to back this up.

Our stand at the Embedded Systems Show 2006 has been confirmed (number 741).

This is a bit of an experiment for us as it’s a bigger, more general event than we’d usually attend. We’ll probably also have a few more personnel on hand as a result so it’s a good opportunity to say hello and, if you like, to discuss whether we can work together. We promise not to give you a hard sell and we’re very friendly, honest.

If you plan to come along and have somehow managed to avoid being bombarded with complimentary show invitations then drop us a line and we’ll send one along. Most of the big names in the embedded world will be there.