How systems programming is like biochemistry

Recently I've been taking a lot of classes in computer architecture and systems programming, someone asked me why, with a biology background, I would be interested in such things. Here's what I said.


In biology right now, we're having to deal with exponentially growing amounts of data, so parallel algorithms and cluster computing are highly relevant. Also, more and more labs are starting to use automation to do experiments and collect data, so systems programming is relevant for that also.

My background is in biochemistry, and I kind of think of computer architecture and operating systems to be at the same level of abstraction, and involve similar kinds of thinking. In biology, the lowest level of abstraction is the physics and theoretical chemistry, things that are modeled with differential equations and other kinds of math that I don't understand very well, the higher levels of abstraction are things like psychology and ecology, where they mostly use statistical models. In between is the biochemistry and molecular biology, where logic and algebra are generally the basis of our models.

For CS, you've got the physics and electrical engineering at the bottom with diff eq models. At the highest levels there is applications programming (and maybe things like data science and machine learning? which use lots of statistical models), and in between there's systems programming, which is all about logic, just like biochemistry!


In fact, one of the more prominent figures in the field of synthetic biology, Tom Knight, spent most of his career working in computer architecture and only in his 50s switched over to biology.

Another parallel I see is between where synthetic biology is right now and where computer architecture and systems programming were in 70s-early 90s. There is an explosion of new start-up companies in the life sciences right now, just like in the past there was an explosion of small hardware and OS companies.