Software through the lens of evolutionary biology

My preferred job title is ‘theorist’, but that is often too ambiguous in casual and non-academic conversation, so I often settle for ‘computer scientist’. Unfortunately, it seems that the overwhelming majority of people equate computer scientists to programmers or some general ‘tech person’, forgetting M.R. Fellows rallying cry: “Computer science is not about machines, in the same way that astronomy is not about telescopes.” Although — like most theorists — I know how to program, the programming I do is nothing like what (I hear) is in industry. In particular, all of my code is relatively small and with concentration, or maybe a single sheet of paper, I can usually keep the whole thing in my head. In fact, the only time I’ve worked in a large code base was developing extensions for MediaWiki during my first summer of college to be used by some groups at the Canadian Light Source. Combined with the preceeding semester of drawing UML diagrams and writing up req&spec documents, I was convinced that I would never be a software engineer. However, I did learn a valuable lessons: real world projects are big and unwieldy, logistics have to be taken seriously, comments and documentation are your friends, and for a sufficiently large software project there is no single person that knows the whole code.

With that much unknown, it is not surprising that bugs abound. Even back in 2002 software bugs cost the US $59.5 billion annually or 0.6% of the GDP, and I imagine the cost has only gone up. If you count ultrafast extreme events or flash crashes of automated hight-frequency traders as bugs, then some argue that you have part of our recent financial crises to blame on software errors (Johnson et al., 2013). To get a feel for the numerosity, a big project like Mozilla Firefox can easily get 2000 new bugs in a year (see figure at left), and Yet most of these bugs are not particularly difficult, and don’t require major overhauls to fix. Even the most serious failures can be fixed by a 12 year-old, why not let evolution have a go?