Hi,

I'm just wondering what your code of conduct is for this project?

Thanks, Alison

Comment From: TomAugspurger

Nothing formal at the moment, AFAIK it hasn't come up.

This would be a good discussion for the mailing list, as not everyone in the community tracks the issues. @alison985 could you propose something there?

Comment From: jreback

I think we should steal this from other projects. prob ipython, numpy, scipy and others have this. we can modify.

Comment From: jorisvandenbossche

NumFOCUS has apparantly a Code of Conduct: http://www.numfocus.org/code-of-conduct.html, and expect from all their sponsored and affiliated projects to adopt a code of conduct. So I suppose what is there is regarded as a good starting point for projects.

Comment From: TomAugspurger

That sounds good for me (that's based off version 1.2, there is a version 1.3 available too).

I can send out a notice to the mailing list if you all think we should formally adopt it.

Comment From: jorisvandenbossche

Yes, sending to the mailing list is good Here is the discussion on the Jupyter mailing list about it: https://groups.google.com/forum/?utm_medium=email&utm_source=footer#!msg/jupyter/XlunILbJUJ8/t7nIumg0AAAJ

Apart from the questions whether to adopt it and where to put it, the main question will be about how to report problems I think. In the convenant text, it is stated to contact a project maintainer / certain email address. In the jupyter discussion, Fernando suggest to

An alternative would be to officially list a group in charge of handling these reports, staffed by say three volunteers. By default we could suggest that reports be sent to all three to increase reporting response speeds, but that gives the option to the reporter of omitting someone if they feel uncomfortable with one of them for any reason (such as that person being the source of conflict, for example).

And on the NumFOCUS page, there is a NumFOCUS contact. But it is not clear to me if it is the intention that subprojects should use that.

Comment From: hayd

Firstly: Big +1.

such as that person being the source of conflict

(Clearly we would hope that this was an edge case!) IMO Sending emails to a private list has the benefits of a bigger paper trail so having people send complaints to a single person should NOT be the default - which I fear it would be if individual's email addresses were published there.

From the covenant:

will result in a response that is deemed necessary and appropriate to the circumstances

What is the suggestion that the outcomes could be? Is this decided by a single person on the team? IMO this process must be transparent (whilst individual cases remain private).

I was going to suggest having a private list where these things can be discussed (if need be). This came up in julia quite recently, where someone was eventually banned from posting on GH, it helped for collaborators to privately discuss and agree on this (understandably hard) decision.

Comment From: TomAugspurger

For the record: https://groups.google.com/forum/#!topic/pydata/ulfquSriMR0

Comment From: shoyer

I'm also supportive of adopting a COC -- we've had at least one case I can think of in the past where this would be helpful. The NumFOCUS one looks fine to me.

Comment From: jreback

@TomAugspurger you want to put the NumFOCUS one up as a PR?

Comment From: TomAugspurger

Going to submit a PR with a draft today. I think the appropriate place is pydata/pandas-governance, and then we link to it from our README, the website, the docs maybe, Stackoverflow tag, mailing list.

In this mail @ellisonbg mentioned he would add one to their governance repo, but I don't see one (Brian, was it added elsewhere? Or consider this a nudge if it fell through the cracks :wink:).

Comment From: jorisvandenbossche

Yep, putting it https://github.com/pydata/pandas-governance seems good!

Comment From: TomAugspurger

Submitted at https://github.com/pydata/pandas-governance/pull/2

Comment From: jorisvandenbossche

This is moved to the pandas governance repo (https://github.com/pydata/pandas-governance/pull/2), so closing here.