Hi, I'm pursuing a Phd (CS, Software Engineering) at the German Aerospace Center (DLR) & Humboldt-Universität zu Berlin. I'm also a #ResearchSoftwareEngineer (RSE) in linguistics.
💻 My research is on #SoftwareEngineering, #ResearchSoftwareSustainability and #SoftwareCitation.
I also teach computational basics to researchers (#SoftwareCarpentry), promote RSE (#deRSE) and am a Special Collaborator of the #SoftwareSustainabilityInstitute (SSI).
I'm based in Berlin, Germany.
@sdruskat Curious about #SoftwareCitation can you elaborate on that a bit? Thanks in advance.
@jpvanacken Sure can, I think (and so do larger parts of the research software community) that software used in/for research should be cited on par with papers, etc. This is currently not always part of practice.
Some issues around how this would work have been solved (https://peerj.com/articles/cs-86/), currently solutions are being worked on (https://www.force11.org/group/software-citation-implementation-working-group).
I'm working specifically on making software citation work for dependencies (i.e. software citing software, like papers cite papers).
@sdruskat Concrete example: I listed the software used for my master thesis in the appendix by naming the software and the version number.
Based on Smith, Katz & Niemeyer (https://peerj.com/articles/cs-86/) I presume that I should also have included
1. a unique identifier
2. the authors of the software
3. the release date of the version used
4. the location / repository that I obtained the software from
to make my list comply to #SoftwareCitation ideals?
Wondering: are there specific styles?
@jpvanacken Sorry for the delay. I think you can fit most of the important info in most of the styles, and styles are usually defined by the media you publish in.
I personally like to use IEEE w/ URL to make sure I can fit in a DOI + URL in the reference.
In general I think a PID is actually more important for citing versions than a link to the repo, but - as usual in research matters - "it depends" :).
@sdruskat Oddly the PID was the one thing that I never would have considered if I had not read some of the material you mentioned before. I was simply unaware that this sort of ID was a thing for software, as it is for journal articles etc.
Is there a default resource for looking up a PID for any / most pieces of software that you can recommend?
The two main reasons behind that being:
1. Software work is important for research and should be creditable and credited (e.g., towards better career paths, getting recognition, etc.)
2. Reproducibility: (Proper) software citation identifies the different parts of a computational process that has led to research results. In order to reproduce results, these parts must be known, e.g., down to software versions being used, etc.