Yesterday was the first official day of this year’s SPA conference.

08:45 saw me arrive at Homerton College to continue helping with the conference set up - a rather early start for me, particularly given the advance of the hour last night.

Andy’s check lists came to the fore again but by 10:30 we’d done most of the set-up and had processed our first delegate - me. A steady trickle of people dropped by to register and by 13:00 we’d probably processed around 70 participants. Most of these were headed for the afternoon’s workshops which traditionally start the conference.

This was also an opportunity to meet Marina Haase in person. I shepherded Marina’s ‘Creativity in Practice’ session for this year’s conference - a first-time experience for both of us - and consequently have been in touch with Marina for the past few months guiding her through the SPA conference quality control process and helping her pitch her session for the SPA audience.

Throughout the day a few people spoke to me about the Code Generation 2007 event I am organising. However, enthusiasts for model-driven methods are a little thin on the ground at SPA 2007 although many people use code generation to some extent.

A frequent concern was that several of them had grown sceptical due to hype in this space in the past. However, my own impression is that Code Generation tools and technologies have matured significantly in the past two or three years and as you can see from the CG2007 event programme most of the sessions are led by people who have applied the technology successfuly for real benefit.

Predictions are dangerous but from my own perspective we’re on the verge of of something significant in this area. Microsoft’s work with Domain-Specific Languages has raised the profile of this area - although Metacase have been working in the area for much longer - and I get the impression that there is a real buzz growing around some of the open source tools such as openArchitectureWare.

A quiet afternoon at the computer saw me making the finishing touches* to my ‘Scoping Game’ session which will run on Wednesday and chatting with Peter Marks about the state of play w.r.t. code generation tools and technologies - Peter and David Harvey had run a workshop on Intentional Programming at SPA 2005 which I enjoyed immensely.

The evening began with dinner and drinks courtesy of Google. Apparently they have a big recruitment drive at the moment going on all across Europe and see SPA as fertile ground. Homerton had laid on a hog roast, which had been cooking most of the day, and apparently this was pretty good - although I stuck to the veggie chilli and chocolate crepes.

I spent the evening chatting variously about:

• Why a one-man VB project in Excel was significantly more productive than a team of 35 C programmers and several orders of magnitude cheaper.
• How people had migrated to Software Engineering from Classics, Music and Physics.
• and of course, why Lisp (and Smalltalk) are languages of choice

Although I headed for bed early the discussions went on into the small hours so there were a few sleepy heads for Day 2…

* I will be making more finishing touches on the session tomorrow.

Yesterday saw me spending a day behind the scenes helping set up for the SPA 2007 conference.

Andy Moorley, who looks after the event administration, runs the operation with military precision and has an impressive number of checklists to support him in this task. After unloading the van and opening all the boxes Andy, Louisa, Pete and I started the (lengthy) job of filling the delegate packs.

After this, while Louisa dealt with assembling the shared conference folders, Pete and I assembled the conference ‘Wall’. This is the place for random stuff left by participants and it consists of a large sticky surface to which various bits of paper, card, elastic bands etc. can be stuck. (Photo to follow).

Then we set up the various rooms and had a tour of the site. SPA 2007 is on a larger scale than Code Generation 2007 - the conference I’m organising with Andy’s help in May at the same venue Homerton College, Cambridge - and takes in pretty much every building at the college.

Later we met up for dinner with Giovanni Asproni at CB2 in Cambridge. The food here’s usually pretty good and we spent a couple of hours dissecting (and solving ;-)) the software industry’s problems - as one does at these types of event.

According to our server logs some lost soul arrived at the Code Generation 2007 web site the other day after typing ‘Eurovision 2007′ into a search engine.

Just to be clear, one of these events will bring together groups of experienced and talented individuals from around Europe to impress and entertain their audience. The other is a singing contest.

Steven Kelly of Metacase has been blogging about May’s Code Generation 2007 event, saying that it “is looking like possibly the best DSM event to date”. Steven should know as he’s been intimately involved in the DSM scene for years, but we’re too modest to agree.

Steve Cook of Microsoft has also blogged about his panel participation at CG2007. Steve will take part in the UML vs. Domain-Specific Languages - a false dichotomy?panel organised by Andrew Watson of OMG and expects controversy.

Finally, Jason Gorman has also mentioned the event on his Agile Software Process Improvement blog - although he seems to suggest that some people might want to keep their interest in model-driven development to themselves - no worries Jason, we won’t tell your mates if we see you at the event.

Visit the CG2007 Community page to see who else is talking about the event.

The end of the Code Generation 2007 early-bird booking period last Friday saw a flurry of bookings come in - which was very pleasing However, some people did report problems paying with credit cards through paypal. This seems to be caused by corporate credit cards blocking payments through paypal. We can issue invoices if people can’t pay in this way but this has meant a little more administration than we’d planned for but we don’t mind really. If you’re one of the people who tried to pay but had problems then please get in touch.

I’ve also been collating suggestions for Heffers who will be running the book stall. Most of the session leaders have now come back to me on this so we should have a good selection of the best code generation books around. With a fair wind we should also have copies of the Microsoft DSL team’s new book (Domain Specific Development with Visual Studio DSL Tools) hot off the presses. You get 10% off any book bought at the event.

I also met with the Gonville Hotel conference manager again last week and things are coming together w.r.t. the conference dinner. Most importantly the menu and wine have now been selected. A vegetarian option will be available for the non-carnivorous and a non-alcoholic option will be available for the non-alcoholics.

Finally, we’ve had a look at where people coming to the event web site are visiting from. The picture below shows where the bulk of the traffic came from on Sunday. Click on the image for a more complete view.

The March 2007 issue of Embedded Systems Engineering magazine contains a short article I wrote on “Developing Mobile Applications with Variant Management”.

50% of the development budget for a mobile game can be spent on porting and testing the game across multiple devices. In the article I explain how variant management with Feature Modelling works and how it can significantly reduce the costs, give better performance and reduce executable size for new variants of mobile applications.

The article is available here (on the Software Acumen web site).

Juha-Pekka Tolvanen has pointed me at some market-research into awareness of model-based code generation techniques among developers and the answer seems to be that most developers are aware of (some) different model-based approaches.

The OMG’s MDA approach was best known with over 70% of developers claiming to have heard of it. Microsoft’s Software Factories came in at 65% awareness and Domain-Specific Modelling at 46%.

Encouragingly, 47% of developers reported having a positive attitude to model-based approaches with 39% already using it or planning to use it.

I hope that a more in-depth analysis of the research emerges soon and if it does we’ll publish it on the Code Generation Network web site.

The past week has thankfully been a little quieter in terms of event planning.

We’ve arranged a book stall for the event (courtesy of Heffers, Cambridge) and participants will be able to get 10% off purchases made at the conference. We’ve asked session leaders for their book suggestions to help Heffers bring the right stock but are of course open to your suggestions. (Don’t be shy of recommending your own book if you’re a published author.)

We have a growing number of people blogging about the event listed at the CG2007 Community page. If you’ve talked about the event in your blog or on your web site then please let us know and we will link to your page.

Our event accommodation page has also been updated and now includes a list of hotels and guest houses near the venue. This should make much it easier for you find somewhere suitable to stay for the duration of the event.

Finally, another reminder that the early-bird booking period for the event closes on Friday 16th March. If you’re planning on coming to the event and have your credit card handy then you can head over to the booking page right now

The SPA 2007 conference organisers got in touch with me the other day to ask whether I’d like to run a session at the conference at the end of this month. I’d previously been notified that my session The Scoping Game, which was the reserve session, wasn’t going to be needed, but apparently they’ve now had someone drop out - just when I was feeling relaxed about not having to lead a session.

I previously ran a longer version of The Scoping Game at SPLC 2006 and it was well-received by the participants there. I’m adapting the session for SPA 2007 in order to make it more accessible for SPA participants and to account for the reduced time available. We’ll still have a chance to play two rounds of the game but won’t look in too much detail at the practicalities of Scoping in a Software Product Line context.

Details of my SPA 2007 Scoping Game session will appear shortly on the SPA 2007 site here.

An article on the BBC notes that “Rising development costs and small user bases mean games developers are unlikely to make a profit on new titles until 2008…”

Development costs are increasing largely due to increased complexity in the games being developed - this has led many companies to increase the size of their development teams and so increase their production costs. Time-to-market is also increasing as a result.

To try and recoup money “many games publishers were choosing to release the same game across as many different hardware platforms as possible to access the widest possible user base”. This is a strategy I’ve seen mobile games developers use - although in the absence of a Software Product Line strategy porting costs can be as high as 50% of the total development budget.

However, some companies are taking a single-platform approach. For example, Epic’s Gears of War is noted as being exclusively developed for the Xbox 360. The author of the original article (Next Generation Consoles: Games publishing, hardware analysis and forecasts to 2010)  suggests that this approach is only viable for developers if they’re given additional payment to mitigate the (market) risk of single platform development.