Where to draw the line for authorship for a software-announcement paper?

Question

I am currently working on a scientific software, for which I want to publish a software-announcement paper in a pertaining journal. While I am doing and probably will continue to do the vast majority of the work, it may be that others will contribute in a manner that brings up the question on their authorship on the paper. To avoid disagreements and subsequent issues, I would discuss authorship as early as possible if such a contribution seems plausible.

In a regular paper, the central criterion for authorship is intellectual contributions, while technical contributions such as a straightforward software implementation of some existing algorithm do not qualify. Moreover (at least in my field), it is rather untypical that somebody contributes only a small piece to a paper such the grey zone of authorship is rarely an issue.

In a software-announcement paper, however, I can think of several aspects that could be considered essential and contributors to which could be considered eligible for authorship:

• Developing new algorithms and approaches. While this is clearly an intellectual contribution and could even justify a paper on its own, not all new scientific software features such a thing. Therefore it cannot be the only aspect qualifying for authorship.

• Choosing the algorithms and methods to implement.

• Devising the interface, usually targetting some particular scientific application.

• Actually writing the software.

• Testing the software. Given the application, finding proper test cases with a known behaviour can be a challenge on its own.

• Writing the actual paper.

Moreover, it is much more likely that somebody makes a small, grey-zone contribution such as a bug fix, providing an example for the documentation or testing, suggesting a feature, or similar. This applies in particular if development versions of the software are published.

My target journal does not provide any guidance on this matter and neither could I find any general rules on the Internet. Hence I am asking whether there are any good (preferably established) guidelines on deciding authorship of such a paper, in particular as to which type of contributions can qualify for authorship at all and where to draw the line.

Answer 1

As the discussion underneath this post shows, there is no set of established guidelines within a single subfield, let alone across all academia, as to where to draw the line.

A different, not unrelated question: what are you trying to achieve with this paper?

If it's about trying to properly apportion credit for the work done in producing the software, then this is a hard problem: there's the work of the FORCE11 Software Citation Group and their paper. But as the software grows and changes, you're going to have to keep making these decisions.

If it's about growing a community around the software, to improve it, get more people to use it and contribute, then there's a thoughtful post here. That's about much more than just a paper, but notes many problems with giving the proper credit when a software project is updated and the members of the community change.

My opinion, for what it's worth: the more broadly you extend authorship invitations, and the more broadly you're seen to do so, the more likely you are to get community buy-in and future support. By making clear expectations about concrete contributions to the paper-writing process itself, you may reduce conflicts around the free-rider problem.