Story: Anger over EC medical data-sharing scheme
Naive politicians constantly waste public funds on grand IT schemes
Politicians who don't understand the complexities of software development naively think that just because they set up a contract with a supplier who claims to be able to deliver, the supplier can deliver. I thought it was only British politicians, but I am alarmed to discover that EC politicians/administrators are falling for this.
As a result of reading about the NHS disaster I produced an analysis here http://www.cs.bham.ac.uk/research/projects/cogaff/misc/isoft
showing why the problems of forecasting resources required grows exponentially with the size of the project, except for projects that are minor variants of what has been done before, which the medical scheme is not.
Moreover, there are many reasons why it is *impossible* to make such systems secure. If it were possible the big banks, who have been in this field for years, would not be losing vast sums of money through fraud, etc. Again politicians are naive in thinking that just pumping money into software companies will ensure that security is achieved.
Moreover, the academics who think that if only proper mathematical/formal methods are used it will be possible to ensure correctness of designs forget that big systems are embedded in a physical, psychological, social and economic environment that they have not a hope in hell of accurately representing in their models, and even if they could, the combinatorics would defeat them.
I think the answer is only to grow small systems, and to run many small experiments in parallel, learning from experience. If public funds are spent, make sure that all results are guaranteed to be in the public domain so that other developers can take the sources and find and fix bugs and improve designs. Don't employ companies that will not agree to this. Others will.
Aaron Sloman
www.cs.bham.ac.uk/~axs
Full Talkback thread
