I’m not responsible for managing the development of software and code for KB+ (apart from in the sense that I have overall responsibility for the project), so I am going to keep my comments short to minimise the chances of getting things wrong.
Over the weekend I was interested by John Naughton’s column in the Observer on Why big IT projects always go wrong.
The column discussed the work of Fred Brooks who worked at IBM in the 1960s. Brooks wrote a series of essays on managing software projects, called The Mythical Man-Month, one of the main lessons he learned in his time at IBM was that the more programmers he added to a project the more likely it was to fall behind.
Now, I’m not suggesting that KB+ is on the same scale as IBM’s developments in the 60s, but something we have strongly resisted the urge to do is throw more programmers at the software development side of things.
To do so would not only have increased the management burden on the project as a whole, but it was judged, it would make it more likely we would spend an ever greater amount of time either trying to ensure that different code from different programmers was compatible, or correcting bugs in different code from different sources.
Sometimes it may have felt as though development has not been as swift as it might be, but for myself I am confident that the approach we have taken has meant we have had to spend much less time correcting code than would have been the case if we had decided to throw programmers at the project from the outset.