Open-source software for data processing and analysis
To be fully reproducible, software used for processing and analysing data needs to be made available open-source.
One of the most frequent arguments against publishing source code is its (lack of) quality. Understandably, most scientists are not (and will never be) professional programmers. Hence, the quality of the software they are forced to write is not necessarily according to their own standards. Therefore: regardless of the quailty of the software, make the code available open-source and free of charge.
Finally, and quite important: make sure to provide a license for your code, as most legislations do not allow code use or reuse by any other person than its original author without a clear license. Use available licenses for this, and never try to come up with your own license – except you're a lawyer with far too much time.