@rmazor @TrevorABranch @hadleywickham This is a perennial question. The best guidance is https://t.co/w0WXWZhRZL by @arfon @danielskatz @kyleniemeyer & @force11rescomm: there's no single rule, but in most cases we should cite in our paper the pkgs that our research claims uniquely depend on, just as for papers. https://t.co/9AGvYlMikw
@PonteIneptique @up_jors Also, of course, opening your software to direct citation by depositing software citation metadata (https://t.co/HN7wHdHw1N) is preferable to/should accompany writing a paper about it (https://t.co/beDjJwNpdb), but potential public scrutiny ≠ peer review, obviously.
@walkingrandomly @danielskatz Sorry to chime in. If no DOI exists, try version ID + GitHub repo URL. Version ID is optimally a version number (0.12.7), but could also be a commit hash. Cf. https://t.co/beDjJwNpdb, section "Unique identification". Author names remain hard though, perhaps "The X project team"?
@StephanJanosch @haesleinhuepf @Rocha_Lab @joachimgoedhart @AndrewGYork @force11rescomm Re #deRSE19: Possibly. In the meantime, the most important source to refer to remains the Software Citation Principles paper (https://t.co/beDjJwNpdb). Primers can also be found at https://t.co/HN7wHdHw1N.
Update from the Software Citation Implementation Working Group - by @npch @mfenner @danielskatz. Came out of the Software Citation Principles, with aim to get them adopted. See the @PeerJCompSci paper https://t.co/7prPIf7veM #FORCE2018