As a grad student in physics, I not only wrote code, but also designed my own (computer controlled) electronics, mechanics, optics, vacuum systems, etc. I was my own machinist and millwright. Today I work in a small R&D team within a larger business, and still do a lot of those things myself when needed.

There are many problems with using a dedicated programmer, or any other technical specialist in a small R&D team. The first is keeping them occupied. There was programming to be done, but not full time. And it had to be done in an extremely agile fashion, with requirements changing constantly, often at the location where the problem is occurring, not where their workstation happens to be set up. Many developers hate this kind of work.

Second is just managing software development. Entire books have been written about the topic, and it's not a solved problem how to keep software development from eating you alive and taking ownership of your organization. Nobody knows how to estimate the time and effort. You never know if you're going to be able to recover your source code and make sense of it, if your programmer up and quits.

With apologies to Clemenceau, programming is too important to be left to the programmers. ;-)