Dry

A great sin in many data analysts (especially statisticians) is repeating themselves. We need to learn from programmers over the last few decades how to code efficiently and effectively.

Perhaps the easiest and quickest starting point is documentation. Why do we keep code description in our header comments and maintain separate documentation in a word document? One easy solution would be to write a program which extracts the comments from all your header files and generates the required documentation automatically through a markup language. Of course, this will mean one important factor, that all coders will need to be pragmatic in maintaining these comments, with some standards in place to ensure uniform presentation.

There are standards which are put in place, examples being markdown, reStructedText, javadoc. Here is my simple implementation (Python and Ruby).

Why is this important? Imagine the hours required to maintain documentation and ensure that all your code repository is correct. But through automation and being pragmatic, the required time would be close to zero.

We need to nip it in the bud.