Add community docs

[effigies]

This PR adds the CONTRIBUTING and CODE_OF_CONDUCT documents that GitHub will privilege to give hints to new contributors as to where to find them. I am mostly linking/reproducing the Nipy community code and the developer guide in the documentation, since we do already have these resources.

The main difference is that I add an "Incident Reporting" section to the community code, so people know to email me. I think ideally it would be good to have a second person who would be willing to be listed in that document, so that if somebody's having a problem with my behavior, there's an obvious person to talk to.

I had initially tagged #733 with a 2.5.0 milestone, which was a good prompt to get me to actually work on this, but I'm going to remove it so that we don't rush this. I'm not interested in trying to rush this through and leaving people turned off about the tone we're setting for the project.

Links with current rendering:

• CONTRIBUTING.md
• CODE_OF_CONDUCT.md

Closes #733.

[matthew-brett]

Oh no - please not the contributor code of conduct. It's horribly written, uses impersonal legal-sounding language, rates poorly on objective readability metrics, and, ironically enough, rates on the extreme male end of the language spectrum on readability.io. Any chance of adapting the Scipy code of conduct instead, for which we did a reasonable amount of work trying to fix those problems? I can do it if you like.

[effigies]

Yeah, I have no problem with scipy's. My impression from earlier discussions was that nipy was generally agreeable and that scipy will take more work to adapt to a small project, but if you'd like to take a shot, by all means.

[matthew-brett]

Yes, the Nipy bit is fine - the ugly bit is the added enforcement section, from the Contributor Covenant, which, as you can tell, is a document I loathe with a deep passion.

The general difficulty is that our (rather sweet) Nipy code of conduct, does not try to define what a violation of the code of conduct is, with the consequence that it's not at all clear people would email you about, or how you'd distinguish between moderately irritating and frank harassment. For reasons I can go into, I think that's a bad position to be in, and I think we need to either:

• do some kind of a job saying what we / you consider a breach, what's not a breach, and how appeals work OR
• (my preference) keep it loosely worded, and say something like:

We put great value on respectful, friendly and helpful communication. If you feel that any of our Nibabel communications lack respect, or are unfriendly, or unhelpful, please try the following steps: If you feel able, please let the person who has sent the email or comment that you found it disrespectful / unhelpful / unfriendly, and why; If you don't feel able to do that, or that didn't work, please contact <ChrisM> directly by email, and he will do his best to resolve it. Alternatively, if <ChrisM> is himself being disrespectful / unfriendly and unhelpful, please contact <Alternative person>. 

[effigies]

Seems reasonable to me. Any interest in being person no. 2? Or have a suggestion?

[codecov-io]

Codecov Report

Merging #778 into master will decrease coverage by 0.04%.
The diff coverage is n/a.

@@ Coverage Diff @@ ## master #778 +/- ## ========================================== - Coverage 90.03% 89.99% -0.05%  ========================================== Files 94 94 Lines 12018 12009 -9 Branches 2136 2133 -3 ========================================== - Hits 10821 10807 -14  - Misses 852 859 +7  + Partials 345 343 -2

Impacted Files	Coverage Δ
nibabel/ecat.py	86.27% <0%> (-1.72%)	⬇️
nibabel/streamlines/trk.py	94.31% <0%> (-0.12%)	⬇️
nibabel/affines.py	100% <0%> (ø)	⬆️
nibabel/quaternions.py	99.02% <0%> (ø)	⬆️
nibabel/nifti1.py	91.48% <0%> (+0.01%)	⬆️
nibabel/casting.py	87.28% <0%> (+0.05%)	⬆️
nibabel/nicom/dwiparams.py	71.69% <0%> (+0.54%)	⬆️
nibabel/__init__.py	93.75% <0%> (+1.02%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update cee6bf3...028e987. Read the comment docs.

[matthew-brett]

I'm happy to do it, but if anyone else wants to do it, I'm happy with that too.

[effigies]

@matthew-brett I added you as a backup.

[effigies]

Thanks for the feedback. Let's see how this goes.