November 13th, 2012
The last year or so I’ve been leading a small team of developers. We’ve been working on a project that involves genomics and molecular biology, bioinformatics, oncology and computional biology. Saying that it’s hugely complex is an understatement. One of the interesting dynamics of this project was that I personally designed and implemented a large portion of the project. A big part of the project was also to solve the actual problem – our client did not have a solution when we came to them, so really we ended up getting access to resources and experts, but no predefined solution.
The question that I’ve been pondering the last week or so is this – if the project had been even more complex; so complex that I wouldn’t have been able to fit it all in my head – could we still have solved it? If I had 50% of the information in my head and someone else had the other 50%, would it still be possible to come up with a working solution?
My intuition is that the answer to this is no. I don’t think a problem of this complexity level could have been solved in a good way by sharing the responsibility for coming up with the solution.
On the plus side, I have never encountered anything even close to this magnitude of complexity before. The projects I’ve been on have been a variety of different enterprise style projects and most of them doesn’t really have much in terms of domain complexity. So maybe this is not a problem in practice for most cases.
But on the other hand, we still have lots of unsolved problems in highly complex and scientific domains. In order to solve them, we need people that can understand both the science and the software aspects at the same time. And based on my experience the last year, I suspect that there are real limits in what kinds of problems we can actually take on in this way. There has to be a better solution. I don’t think we have a solution to this problem yet. Incremental development methodologies really doesn’t help for this.
Another interesting aspect of this project is that we did not have any BAs (business analysts). Most of our projects have BAs and it’s highly unusual to not have them. In retrospect it was the right choice for us and I can now verbalize why – when you have BAs working with the domain, you still have to take into account the communication with the developers and tech leads. If the domain is complex enough and the developers need to have that understanding, having BAs would actually get in the way and the communication surface area would be to large to effectively work. Me and one of my colleagues ended up together doing all the BA work in conjunction with our design and implementation work.
Working on a project like this has been a very different experience. It’s definitely clear to me that our standard ways of working with businesses doesn’t really apply.