Scientific Programming

By Daniel Chen

March 27, 2017

Software-Carpentry played an integral part of who I am today. I am always trying to learn and follow best practices in the context of scientific programming, which I think is a neglected area in research.

The fundamental problem is the incentive structure in academia, where productivity is measured by the number of papers published. The downside of tis pressure is that quality of analytics and code will suffer to get the results for the paper.

Those of us who are doing research and realized how much coding is needed for analytics see programming as a means to get results. Teaching yourself a technical skill is extremely difficult without mentorship. It’s very easy to learn and pick up bad habits.

I’ve seen and heard multiple times that the code is written in a particular way solely because it ‘works’, without considering that it can potentially have bugs in it and might need to be patched. It’s extremely frustrating when I try to provide guidance and try to teach a skill, but am confronted with “No. I don’t have time for that”.

It only makes me wonder how much this happens in scholarly research, and can also explain why science has a reproducibility problem.

Posted on:
March 27, 2017
1 minute read, 202 words
teaching higher ed pfps17
See Also:
Using OBS for Online Teaching
Changes in Higher Education
Preparing for the Summer