Initial pandas governance documents #1

wesm · 2016-01-07T02:24:46Z

For review and comment by @pydata/pandas core team based on pandas-dev mailing list discussions.

I drafted this from https://github.com/jupyter/governance which is a more complex OSS project so may have made some typos or left some irrelevant points.

I've also revised the project's license to define the "pandas Development Team" in line with the governance document and I recommend that we begin to apply copyright notices to all pandas code files. Maybe there is a script floating around to make it easier to fix up copyright headers.

…ce documents

shoyer · 2016-01-07T02:27:11Z

governance.md

+defer that authority to the consensus of the community discussion channels and
+the Core Team. It is expected, and in the past has been the case, that the BDFL
+will only rarely assert his/her final authority. Because rarely used, we refer
+to BDFL’s final authority as a “special” or “overriding” vote. When it does


missing a vowel in "Because rarely used"

I meant "verb" not "vowel" :).

should be: Because *it is* rarely used

jreback · 2016-01-07T02:43:11Z

people.md

+## Project Core Developers
+
+- Tom Augspurger
+- Phillip Cloud


is this the steering committee? are we using that nomenclature?

I used the "Core Team" terminology in the governance doc. Let me fix that here

TomAugspurger · 2016-01-07T05:00:59Z

Not sure if we're going for the 80% quorum on this, but +1 from me.

hayd · 2016-01-07T05:28:45Z

+1. This is great, thanks for putting this together @wesm.

jorisvandenbossche · 2016-01-07T10:17:04Z

governance.md

+
+To become eligible for being a Core Team Member an individual must be a Project
+Contributor who has produced contributions that are substantial in quality and
+quantity, and sustained over at least one year. Potential Core Team are


'Potential Core Team' -> 'Potential Core Team Members' ?

sinhrks · 2016-01-07T14:24:36Z

+1, It's great to have a clear and fair policy.

…ance

wesm · 2016-01-07T18:56:08Z

Thanks for the comments. Will leave this open for comments for several more days and we can vote to merge sometime next week

jankatins · 2016-01-10T09:37:31Z

projectlicense.md

+==============
+
+Copyright (c) 2012-2016, pandas Development Team
+All rights reserved.


This ends up as one line, reading ...pandas Development Team All rights reserved. Add a period after "Team"?

Same for the next two copyrights...

wesm · 2016-01-11T16:58:31Z

@cpcloud @shoyer @lodagro @jreback @jseabold @changhiskhan @jtratner @jorisvandenbossche can you review and indicate your +1 so we can merge these documents? We can add the code of conduct in #2 shortly afterward. thanks!

jorisvandenbossche · 2016-01-11T17:10:13Z

governance.md

+- Ability to influence the project through the participation of their Core Team
+    Member.
+
+**Tier 2** = an institution with at least one Institutional Contributor


This distinction in two levels seems a bit strange if there is nothing specific about the second level ? (no benefits listed)

The Jupyter governance gives the same benefits to to Tier 2 as to Tier 1 (from the ones we have listed), while numpy just not makes the difference (http://docs.scipy.org/doc/numpy-dev/dev/governance/governance.html#institutional-partners-and-funding) which actually is equivalent to giving Tier 2 the same benefits as Tier 1

In both cases, the Institution will be recognized as an Institutional Partner and will be able to claim as such. The Tier 1 partners will receive additional acknowledgement (e.g. websites and so forth). I don't have strong views, but the Jupyter-style IP's seems fine to me.

anyone know whom are institutional partners for Jupyter or other projects? (if any)

For Jupyter, they are listed here: https://github.com/jupyter/governance/blob/master/people.md#institutional-partners, and they are also listed on the website with their logo: http://jupyter.org/about.html (scroll down), they have quite a lot.
Numpy has also one institutional partner at the moment: http://docs.scipy.org/doc/numpy-dev/dev/governance/people.html#institutional-partners

Contributing to pandas is not part of my official duties at Cloudera, but if that changes in the future I will advise the Core Team.

(For example, I'm working on pandas today but I am on PTO)

It's a pandas vacation! 🎉

@jreback I suggest that we propose adding Institutional Partners 1-by-1 to the document through separate patches after we merge this initial draft if that works for you

sounds good

jseabold · 2016-01-11T17:18:44Z

lgtm modulo the one style comment

jorisvandenbossche · 2016-01-11T17:32:14Z

Two other small nitpicks:

The text states:

+Traditionally, Project leadership was provided by a BDFL (Wes McKinney) and
+subset of Contributors, called the Core Team, whose active and consistent
+contributions have been recognized by their receiving “commit rights” to the
+Project GitHub repositories.

In practice, 'commit rights' and the core team were coupled in the past (as there was no official definition of the core team). Do we want to keep this coupling? Or can we give more quickly give commit rights without necessarily becoming a Core Team member?
The financial sponsorship agreement document speaks about a "Governance Council" instead of a "Core Team" (I don't know how important this is to be the same)

But in any case, +1 !

@jreback, I am not sure, but is it correct that you can work partly on pandas during your working hours? If so, should Continuum be listed as Institutional partner?

lodagro · 2016-01-12T20:10:10Z

+1 for me

I wondered if there would be a section on inactive Core Team Members,
clearly there is.
Given that my last commit is from June 2013, i can be removed from the core
team, maybe also revoke my commit rights.

2016-01-11 17:58 GMT+01:00 Wes McKinney notifications@github.com:

@cpcloud https://github.com/cpcloud @shoyer https://github.com/shoyer
@lodagro https://github.com/lodagro @jreback
https://github.com/jreback @jseabold https://github.com/jseabold
@changhiskhan https://github.com/changhiskhan @jtratner
https://github.com/jtratner @jorisvandenbossche
https://github.com/jorisvandenbossche can you review and indicate your
+1 so we can merge these documents? We can add the code of conduct in #2
#2 shortl y afterw ard.
thanks!

—
Reply to this email directly or view it on GitHub
#1 (comment)
.

lodagro · 2016-01-12T20:10:12Z

governance.md

+umbrella and cannot use the Project trademarks in a way that suggests a formal
+relationship.
+
+To acknowledge institutional contributions, there are two level of Institutional


Institutional Contributions ?
... two levels ...

Fixed, thanks.

wesm · 2016-01-13T09:19:52Z

@lodagro we can visit the core team membership in a follow up pull request

wesm · 2016-01-13T09:25:17Z

@jorisvandenbossche to your points

I think core team and committership is a reasonable coupling. Patch acceptance and merging is an important feature of OSS project governance -- the ASF model has both committers (equivalent of core team) and a PMC (more of a project board) so this is somewhat consistent.
I hope it will be reasonably clear that the Core Team == Governance Council as it relates to the NumFOCUS fiscal sponsorship agreement. If somehow it becomes an issue we can amend these documents

wesm · 2016-01-13T09:29:24Z

@cpcloud @shoyer @jreback @changhiskhan @jtratner if you have further comments, it would be good to wrap this and #2 up by weekend

jreback · 2016-01-13T13:28:53Z

governance.md

+===========
+
+The pandas Project (The Project) is an open source software project affiliated
+with the 501c3 NumFocus Foundation. The goal of The Project is to develop open


wesm · 2016-01-13T21:33:13Z

I see @sinhrks on the core team list -- do we have a proper name for this person or we can leave as is?

jreback · 2016-01-13T21:37:12Z

Masaaki Horikoshi

sinhrks · 2016-01-13T22:01:48Z

Ah thanks. Pls include my proper name.

wesm · 2016-01-14T23:59:39Z

Barring concerns from @cpcloud, @changhiskhan, or @jtratner I'll hit the merge button tomorrow. Thanks all for helping out with this

wesm · 2016-01-16T16:32:08Z

hi folks -- it's been pointed out to me (quite correctly) that the pandas-dev mailing list has a much smaller circulation than the broader pydata google group, and the community would benefit from an opportunity to scrutinize these documents and raise concerns even if they aren't coming from someone named on the core team.

I'll send an email to pydata@googlegroups and we'll leave the discussion open for a couple more weeks. I would also like everyone named in the People document to +1 before we merge. Thanks

aterrel · 2016-01-18T15:45:42Z

Hi all, Thanks for sending our to the pydata list. I have a couple of comments.

You might add that the Pandas Project has the right to receive tax-free donations in the US through NumFOCUS (Europe is a complicated can of worms but many countries allow donations tax free). We find many corporate donors can work this way better and for the project it protects you from having any obligation to the money donated.
I wouldn't call the Project NumFOCUS Board a board. Unfortunately, the word board is used in US legal contexts to hold liability to the activities of a project. While in the US you will have a greatly limited liability and even more so with the Pandas License, it is usually better to avoid the word. Calling it a Project NumFOCUS Subcommittee would be better.
This document doesn't give a path of mediation upon breach of the document. Many people assume that a NumFOCUS affiliated project means that breaches should be the arbiter but we are really the fiscal entity not the governing entity. So as it stands the only mediation to this document if a person sees her rights as a core contributor or institutional partner being abused is to appeal to the courts. I would suggest adding something to the effect that non-compliance should be reported to the core membership. This is a place where private communication might be more appropriate.

aterrel · 2016-01-18T15:58:30Z

I guess to further clarify point 2, the Jupyter document gives the new BDFL decision to the main NumFOCUS Board not its NumFOCUS subcommittee. Reading this document as it stands it seems unclear if it is giving this right to the Pandas NumFOCUS subcommittee or the NumFOCUS Board, i.e., I find the wording Project NumFOCUS Board confusing.

wesm · 2016-01-18T16:44:44Z

Thanks @aterrel, I'll propose some edits to incorporate this feedback.

…lection of a new BDFL

wesm · 2016-01-18T20:37:32Z

Have a look. I adopted the same process for changing the governance documents for BDFL election in the event that I am unable. I also removed the "board" terminology re: our NumFOCUS subcommittee and noted that we are able to receive tax-deductible donations.

aterrel · 2016-01-19T14:46:58Z

@wesm thanks for the edits! I think they look good.

jreback · 2016-02-13T03:45:11Z

we should prob merged these at some point....

wesm · 2016-02-13T04:33:44Z

wesm · 2016-02-13T04:35:58Z

@cpcloud @jorisvandenbossche @changhiskhan @jtratner can you please review and give your +1 for merging this.

To the others -- if I misrepresented in carrying over your +1 from the discussion please let me know, not intentional!

jorisvandenbossche · 2016-02-15T00:19:37Z

A +1 from me!

cpcloud · 2016-02-15T00:21:41Z

+1 from me too!

changhiskhan · 2016-02-15T17:22:58Z

governance.md

+Question: if the BDFL is not on a subcommittee, do they still have override
+authority?
+
+Suggestion: they do, but they should appoint a delegate who plays that role


Should this be formalized rather than formulated as a "suggestion"?

Possibly. Until we create more subcommittees (unclear when it would extend beyond the NumFOCUS fund-management side of things) it's fairly hypothetical, and we could always establish a more formalized process around appointing a subcommittee "BDFL lieutenant" if deemed helpful.

changhiskhan · 2016-02-15T17:40:10Z

+1 from me other than the 2 review comments above. 🚢

… the same direct manager

femtotrader · 2016-02-16T15:35:52Z

+1 I'm member of Pandas-DataReader http://pandas-datareader.readthedocs.org/ project not a "core" dev... but anyway that's a good idea to write a governance document / code of conduct

wesm · 2016-02-16T17:24:05Z

Thanks @femtotrader -- I think we should definitely encourage more projects like pandas-datareader. If there's anything we can be doing to help that please let us know. Definitely happy to add core team members who are working on affiliated projects but not necessarily core pandas

femtotrader · 2016-02-16T19:33:41Z

@davidastephens was also very involved in pandas-datareader https://github.com/pydata/pandas-datareader/graphs/contributors

jtratner · 2016-02-20T02:54:03Z

This document looks excellent! Thanks for putting it together and I think it's a great step in favor of pandas future development. I've been on the sidelines but I'm incredibly excited about the direction of pandas as it goes to 1.0 and this is a great companion to that.

jtratner · 2016-02-20T02:54:30Z

(Also I apologize for the delay in my reply, initial email got caught in my filters and I've been a little snowed under)

wesm · 2016-02-23T20:01:00Z

We have a quorum, so I am going to merge these documents. Thank you everyone for the feedback.

Initial pandas governance documents

wesm added 2 commits January 6, 2016 18:20

First draft of governance documents based on Project Jupyter governan…

946dd5c

…ce documents

pandas uses a simplified BSD license

b9ab8a5

shoyer reviewed Jan 7, 2016
View reviewed changes

Grammar fix

5a2a062

jreback reviewed Jan 7, 2016
View reviewed changes

Consistency with Core Team

0be435a

jorisvandenbossche reviewed Jan 7, 2016
View reviewed changes

Incorporate comments and add Subcommittee details from Jupyter govern…

a4cb7b6

…ance

jankatins reviewed Jan 10, 2016
View reviewed changes

jorisvandenbossche reviewed Jan 11, 2016
View reviewed changes

lodagro reviewed Jan 12, 2016
View reviewed changes

Fix typos and style issues

ab9d226

wesm changed the title ~~Draft of pandas governance documents~~ Initial pandas governance documents Jan 13, 2016

jreback reviewed Jan 13, 2016
View reviewed changes

Fix typos and add proper name for sinhrks

f68e5ff

Remove NumFOCUS board nomenclature. Provide a new process for team se…

97fa8b7

…lection of a new BDFL

changhiskhan reviewed Feb 15, 2016
View reviewed changes

Clarify that two individuals can share reporting lines without having…

6c92ecf

… the same direct manager

wesm added a commit that referenced this pull request Feb 23, 2016

Merge pull request #1 from pydata/governance-draft

c737e3b

Initial pandas governance documents

wesm merged commit c737e3b into master Feb 23, 2016

wesm deleted the governance-draft branch February 23, 2016 20:01

Uh oh!

Initial pandas governance documents #1

Initial pandas governance documents #1

Uh oh!

Conversation

wesm commented Jan 7, 2016

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

TomAugspurger commented Jan 7, 2016

Uh oh!

hayd commented Jan 7, 2016

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

sinhrks commented Jan 7, 2016

Uh oh!

wesm commented Jan 7, 2016

Uh oh!

Choose a reason for hiding this comment

Uh oh!

wesm commented Jan 11, 2016

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

jseabold commented Jan 11, 2016

Uh oh!

jorisvandenbossche commented Jan 11, 2016

Uh oh!

lodagro commented Jan 12, 2016

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

wesm commented Jan 13, 2016

Uh oh!

wesm commented Jan 13, 2016

Uh oh!

wesm commented Jan 13, 2016

Uh oh!

Choose a reason for hiding this comment

Uh oh!

wesm commented Jan 13, 2016

Uh oh!

jreback commented Jan 13, 2016

Uh oh!

sinhrks commented Jan 13, 2016

Uh oh!

wesm commented Jan 14, 2016

Uh oh!

wesm commented Jan 16, 2016