are computers a complex system?

photo by photonquantique under cc license

i'm taking a course called Introduction to Complexity taught by Melanie Mitchell through the Sante Fe Institute.

we are not deep into the course yet, but i've already been enlightened with many of the concepts presented in the lectures so far. one that really spoke to me was an interview with Liz Bradley about work she's done exploring computer performance as a complex system.

i thought that was pretty ingenious. it's seem pretty obvious at first thought that computers are not complex systems. rather, they just do what their told. you have registers and memory and instructions sets. you load these up and the outcome is very deterministic and predictable.

but then, when you think about the performance of computers, there could very well be some behaviors there that the tools of complexity could help inform. these complexities manifest themselves in everyday computer use that we just accept (think locks, reboots, or just random glitches).

one thing that is telling is that she says that she's been unable to get any computer science researchers to be interested in the work and chalked it up as "one of the problems" with cross-disciplinary work. i think she's probably right about that.