JOSS submission?

[gallais]

TIL https://joss.theoj.org/ gives us the opportunity to "publish" the stdlib as a small journal paper.
This would give people a standard way of citing the lib & crediting the contributors

[JacquesCarette]

That's a really good idea. Count me in on those willing to do the admin work to get this done.

[MatthewDaggitt]

Sounds good to me! How does authorship work? Do we add anyone who has ever contributed?

[JacquesCarette]

My personal view is that the paper should "thank" people with a very small number of contributions (and those who have been active on issues without necessarily doing PRs), but be authored by those who have been steady participants (at any point in time of the development of the library). We should ask each potential author if they want to be named as author - some may decline! Probably want to include all who've made 10+ commits.

[MatthewDaggitt]

Sounds reasonable to me!

[gallais]

From the submission guidelines:

As a rule of thumb, JOSS’ minimum allowable contribution should represent not less than three months of work for an individual. Some factors that may be considered by editors and reviewers when judging effort include:
(...)
The authors themselves assume responsibility for deciding who should be credited with co-authorship, and co-authors must always agree to be listed. In addition, co-authors agree to be accountable for all aspects of the work, and to notify JOSS if any retraction or correction of mistakes are needed after publication.

[JacquesCarette]

Since stdlib amounts to multiple person-years' worth of effort, the first bit is passed no problem (and all the others from the same guidelines too). [Too bad my student's contributions don't count as mine, I'd be higher up on the list... darn agda-categories!]

[jamesmckinna]

It seems that I am now just (but still some way) behind @JacquesCarette in the contributors list, so I'd like to add myself to any potential author list for such as submission, and I suppose (rewarding self-publicity! ;-)) raise a PR to add my name to the LICENCE? (DONE: #1893 ;-)

[jamesmckinna]

Ob which, @gallais, you are conspicuous by your absence from the list of contributors there?

[JacquesCarette]

It's down to working habits: I make more, smaller PRs, comprising multiple commits. You do larger chunks. If you sort by lines added, you far, far outstrip me as a contributor @jamesmckinna . Those same differences in working styles show up in @MatthewDaggitt and Nils' "rankings".

I'll spare you my rants on metrics, gamification and metrics+gamefication.

[jamesmckinna]

@JacquesCarette for sure, all the possible ways of measuring things, and all the possible ways of gaming that (to what unimaginable ego-driven ends, in a volunteer/unfunded development project ... ;-)) need setting to one side. But I guess there is some sort of (internal) threshhold, as well as a community one, before one (I) really feels to have made a (meaningful) contribution... partly as well (separate rant, also likewise suppressed) because by contrast with other sciences (naming no disciplinary names), we don't have a culture in CS of multi-tens-of-author papers... The current JOSS suggestion might be a gesture in the direction of rebalancing that arithmetic.

As for styles of PR/commit policy, I guess a number of my contributions this year have been more 'global' changes, on the basis of trying to grok exactly how wide/deep might be the consequences of changes in one place (and thereby gaining some more global perspective on the design of the whole library), as well as a sprinkling of more local contributions of 'new', rather than 'refactored', material.

[JacquesCarette]

Bumping this back up onto people's radar. 2.0 would be a good milestone to send to JOSS.