Hello @ChrisJefferson , thanks for responding.

We stand strongly with the idea that if you use a piece of software in your research and the software/the authors request citation then you should cite it.

Did our wording suggestion something different to you?

#4 also attracted questions on Twitter https://twitter.com/ctitusbrown/status/712585852178837505

We're looking to see how to clarify our intent there

My equivalent here would be, some person in my field wants me to cite their paper.

Either I should be citing their paper because it contains relevant work, or it isn't relevant so I shouldn't cite it. Their wish of if I should, or shouldn't cite it should be irrelvant to the process.

Similarly, if software A and software B were equally relevant to my project, I would consider it scientific bad behaviour to cite A because they want it, but not cite B because they don't really care. Either they both need citing, or neither.

@ChrisJefferson We are in agreement with you. The decision tree is to help authors get to a firm "yes" or "no" answer as quickly as possible.

I completely agree with Chris and disagree completely with the workflow.

The authors' wish to be cited or not or benefit from being cited or not should be irrelevant EXCEPT in a very small number of cases. Roughly speaking those cases would be "Is it a coin toss situation whether to cite or not." In no situation should you cite somebody just because the author would like you to, and in no situation should you cite something just because it would help the authors out financially or otherwise.

Note that nothing in the preceding paragraph is to do with software. It's to do with any scientific contribution. Software should not be elevated to a special case.

@mr-c : If you are in agreement with me, then I don't see how you could defend points 4 and 1, in particular point 1 being first.

Perhaps point 1 could be changed to "Does the software require you to cite it"? (if that is the case it is meant to cover, including for example gnu-parallel).

In that case one must either cite it, or stop using it (in the case of gnu parallel, I plan to stop using it, and recommend others stop using it).

I agree with the point of this issue.