Let me first state some conflicting assertions of the matter to illustrate what are the issues.
Personally I would like to have my code open at every stage of development, since
- others shall see and take advantage of what I am doing
- I also like to reuse existing code
- third parties can contribute
- the public is funding me, so the public has the right to see
But my boss says
- he needs to approve what becomes public under his name (or the institute's) and he cannot approve every single step
- There are guidelines referring to intellectual right properties of the institute
and my colleagues say
- others will come and steal my unpublished ideas
- my experimental code is of little use for others
To come up with a blueprint for code publication and open source development in my lab, we want to raise the following questions.
- Is there already such a guideline, covering the important issues of open software in academia?
- Which issues have to included into such a guideline?
- What do you think is the right way to implement and use such a guideline?
The issues of crediting, reproducability, code documentation, and where to publish, we want to address in a separate guideline.