Unit testing, readability version control, documentation, etc are all engineering practices for the purpose of making ongoing development organized (especially

Unit testing, readability version control, documentation, etc are all engineering practices for the purpose of making ongoing development organized (especially for teams). Why would a researcher need to do this, when in most cases all that they use is the output, and in CS/math it's only a minimal prototype demonstrating operation of their principle? All of the other stuff would certainly be nice, but they don't need to adopt our whole profession to write code