I’ve added some annotations to the slides from an introductory talk on Software Product Lines that I gave to the British Computer Society Software Practice Advancement group in London yesterday and put them on the Software Acumen site. They’re available as PDF (1649 Kb) or Powerpoint slide show (1218 Kb).

Immo Hüneke hosts these events for SPA and I hope he’ll make his copious notes available as some good questions came up but, being the speaker, I didn’t have time to record them all.

This talk was itself an updated version of a talk I gave at the SPA 2005 conference.

I recently met someone who had been a software development manager in his past career but had now changed role. He thought that being a software development manager was the worst thing you could be in the software industry due to the pressure exerted on you from management above, team members below, and other stakeholders elsewhere in the organisation.

I’ve been a software development manager before and I certainly recognised those pressures. However, I couldn’t agree that it was the worst job in the software industry.

Sure, it’s not always great being stuck in the middle like this or dealing with the legacy of those who’ve gone before, but I feel that the position can offer a great upside, particularly when everything comes together. As the overall manager you also have a significant voice in how the team goes about its daily development work.

Also, other jobs have their own pros and cons…

I’ve been busy these past few weeks promoting the Code Generation 2007 event that we’re holding in Cambridge, UK from May 18th - 20th next year.

In the past week we’ve had Object Management Group, IET and IASA all sign up as supporting partners for the event. Support like this is critical in getting word out about the event - particularly when getting people interested as speakers. I’m expecting more sponsors to sign up in the next few weeks and this will help publicise the event further.

Objective View magazine, a developer magazine published by Mark Collins-Cope, has also joined us as a media partner, if you’re an Objective View subscriber (and if not, why not?) then you should see CG2007 advertised in issue 10 of the magazine.

Andy Moorley of Truedata is also doing a fine job with the website and we’re starting to see good visitor statistics for the site. We’re expecting further updates in the next couple of weeks including pricing information for non-session leaders and details on invited speakers.

I’ve also added a note on the site about suggested sessions that currently don’t have a session leader. So far we have:

• Code Generation is not the ’simplest thing possible’ - discussing the use and development of Code Generation tools in Agile teams
• Adoption of Model-Driven approaches - where to start?
• Evolution issues when using Domain-Specific Languages - e.g. what happens when your language changes

Please get in touch if you’d like to lead or participate in a session along these lines.

I recently asked a question over on cgn-talk (a Yahoo group associated with The Code Generation Network (note membership required)):

“Agile methods, such as eXtreme Programming, advocate doing the simplest thing that could possibly work when it comes to designing software. Does anyone have any thoughts on whether this means agile developers should avoid building their own code generators?”

Some interesting answers came up that I’ll try and paraphrase here:

• Agile developers would advocate code generators as targets of refactorings much like patterns
• Duplicate code in C++ can sometimes be refactored to C++ templates - a form of in-built code generator.
• Code generation & reflection-based tools can enable more Agile development
• Database access … is almost always better started with generators
• Agile modelling and Code Generation are a perfect fit
• I think the answer is no, or at least it should not be. A logical approach to code generation is to create the first prototype by hand. When it comes to the second, you refactor (as someone else pointed out, code generation can be perceived as refactoring). The third, etc are free.
• Some agile shops prefer to do OR/M runtime mappers instead of code generation. I think the benefit could be achieved equally well with better long term consequences (for many projects) with code generation.
• Where I think code generation really shines in agile is in refactoring / overcoming fear of bad code / architecture.
• Code generators are part of the tools that we use every day for java development, tools like castor, hibernate, wsdl2java, XDoclet and other tools are or use code generators. The only difference is that we don’t think of them as code generation tools.

So, the answer seems to be that agile developers will use code generation tools if appropriate and will also consider creating a code generator, or refactoring to a generator, if that is appopriate.

According to Nigel Shadbolt, President of the British Computer Society, the British computer industry faces a crisis due to a mismatch between demand and supply of computing graduates.

There has been a steep decline in the numbers of people studying CS in recent years and of those counted as studying CS many are only taking CS as part of their degree.

Nigel seems to atribute this to poor schooling and the ‘grey’ image of Computer Scientists. I can’t say much about current schooling as it was some time ago when I was schooled and I know that the ‘grey’ image is a bit of a myth, probably put about by jealous media types.

I also wonder whether the trend to offshoring jobs is deterring people from CS study. I wouldn’t want to take on tens of thousands of pounds worth of debt to enter an industry where jobs are being offshored at the rate of thousands per year.

BCS itself seems broadly supportive of offshoring. See for example the BCS position on offshore outsourcing and the report of the BCS Working Group on Offshoring. Perhaps they should investigate further whether prospective CS students are being turned off because of offshoring.

Nigel Shadbolt’s comments were reported on the BBC News site here: http://news.bbc.co.uk/1/hi/technology/6155998.stm

Thanks to Manning, The Code Generation Network’s November / December competition prize is a copy of “Code Generation in Action” by Jack Herrington, former editor of CGN.

See http://www.codegeneration.net/tiki-read_article.php?articleId=85 for full details.

The competition will run until December 31st 2006.

It’s just struck me that we’ve publicised the CG 2007 call for participation everywhere except here!

The conference runs from May 18th-20th at Homerton College, Cambridge, UK. We’re expecting around 120 participants so it’s a good opportunity to demonstrate your  knowledge in front of a sizeable audience although we’re aiming to run several parallel sessions so we can’t guarantee the full 120 at your session.

Session leaders whose proposals are accepted won’t have to pay conference fees. Visit our page on running a session to find out what’s involved. The call for participation closes on Friday January 12th 2007.

Some thinking has been applied and we’ve now come up with a name for the previously untitled Code Generation Network event taking place in Cambridge, UK in May 2007.

The event will be (imaginatively) titled Code Generation 2007 and now has an associated website at http://www.codegeneration.net/cg2007/.

We’re now starting to publicise the event more widely, so please pass details of the event along to friends and colleagues and please consider submitting a session proposal.