It Won’t Work, It Won’t Work, It Won’t Work. It Worked
At the Testing and Finance Conference in Germany we presented a case study that showed how we transformed a financial company from a level 1, ad hoc, organisation into a level 2, managed organisation. We did this by combining management techniques with testing techniques.
Testing is a cornerstone of software development, and therefore it’s a cornerstone of financial software development. Developing financial software is both easier and harder than building other systems. It’s easier in the sense that a single financial engineer might understand a model and be able to translate that into Excel with some VBA code. The software may then be used on a laptop to show a customer what’s going to happen to their money, or it may be used to build a curve for use in a later presentation. It’s harder because once you start to make these tools robust and available to more than one user, you have to deal with data from multiple sources, double point precisions and random numbers, both of which make testing a little bit more interesting.
This was what our talk at Testing and Finance was about. Our blurb said it all:
There is a cultural disconnect between software and financial engineers. They disagree on done, on levels of rigor and where to apply it. Their reward mechanisms, as captured in the brain, the manifestation of years of different styles of learning, are different. They don’t always like each other.
Ugly Duckling’s focus has always been, will always be, the spaces in between, never people vs. process but people via process. Never financial vs. software engineering, but each one via the other. This conference was important for us because it was a continuation of our delineation of the current state of financial and software engineering.