Everyone can agree that sharing of code and methods are useful for fast progress in research.
However, is it a problem that things easily available may be used without sufficient critical thinking about assumptions and possible flaws in the implementation?
I use to say that:
A good user should know the software and assumptions sufficiently to be able to develop the software himself (with sufficient time available), unless he is prone to others improper assumptions and simplifications. (I am not saying this should be done but unless you understand the software, you are an user not a researcher)
To really know the software (and the methods implemented) one has to read the source code (and be able to modify the code or implement new functions). If the source code is not available, the software is a black box: garbage in, garbage out. Every scientific software should allow the scientist to interact with the software's programs via command-line user interface.
Commercial software, which rarely ships with the source code, tends to be, for obvious (monetary) reasons, more user-friendly (more friendly, more users; more users, more sales). But user-friendly usually means scientist-unfriendly. In contrast, open-source scientific software usually focuses on functionality over user-friendliness, which means fewer users, but higher likelihood of meaningful outputs.
- Badges
- Science method
- Similar topics
- Philosophy