Re: [Ietf-and-github] Warren Kumari's Discuss on draft-ietf-git-using-github-05: (with DISCUSS and COMMENT)

[Mark Nottingham <[e-mail]>]

> On 12 Mar 2020, at 10:05 am, Martin Thomson <[e-mail]> wrote:
> 
> On Thu, Mar 12, 2020, at 06:58, Warren Kumari wrote:
>> Thank you -- I would even be OK with just changing to Abstract to be:
>> "This document describes best practices for Working Groups that choose
>> to use GitHub for their work."
>> and leave it as BCP.
> 
> I should read all the thread before replying...
> 
> https://github.com/ietf-gitwg/using-github/pull/46 makes this change.

I have some sympathy for Warren's concerns (regardless of the process intricacies).

I didn't review this document closely in LC because my recollection when we started the group, there were numerous reassurances that it was advisory only, not requiring things from groups that didn't want to opt into these processes. The charter seems to support this view (note "that choose to adopt the *practices*", not "Github"):

"""
The GitHub Integration and Tooling (GIT) working group will select a set of such practices and document policies that support those practices. The policies will each detail how work is conducted by working groups that opt to follow the work practice. The goal is to provide both process and tooling support for working groups that choose to adopt the practices.

...

Whether working groups choose to use GitHub or the documented policies to support their work will remain entirely at their discretion.
"""

So this change doesn't go far enough; it needs to say something like:

>  This document describes a set of practices that Working Groups using GitHub can choose to adopt.

Changing the document to Informational would also make me feel _much_ better about it.


With apologies for the late feedback, a few more notes FWIW:

* "Chairs MUST involve Area Directors in any decision to use GitHub for anything more than managing drafts."  This seems onerous; does it mean that I need to consult with the AD to host some tests (e.g., <https://github.com/httpwg/structured-header-tests>), a repository for a wiki (e.g., <https://github.com/httpwg/wiki/wiki>), working group administrivia (<https://github.com/quicwg/wg-materials>), or a translation (<https://github.com/quicwg/zh-translations>)?

* "Maintaining multiple documents in the same repository can add overhead that negatively affects individual documents.  For instance, issues might require additional markings to identify the document that they affect.  Also, because editors all have write access to the repository, managing the set of people with write access to a larger repository is more difficult (Section 3.3)."  This is well-worn territory, but there are also costs associated with having many repositories -- both in terms of WG chair load, and in terms of potential errors (as we're discovering in QUIC).

* "Issues that have reached a resolution that has Working Group consensus MUST NOT be reopened unless new information is presented."  This seems to be pre-emptively establishing both the status of Github issues in the IETF process, and the semantics of their state. 

* "Editors SHOULD create a new Internet-Draft submission two weeks prior to every session..."  How is this requirement specific to Github?

* Generally, there are a lot of RFC2119 terms that seem advisory, not normative. Are they all well-considered? To give one example: "Editors SHOULD make pull requests for all substantial changes rather than committing directly to the "master" branch of the repository." What does it mean to violate that SHOULD?

Cheers,


> 
> 
> _______________________________________________
> Ietf-and-github mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/ietf-and-github

--
Mark Nottingham   https://www.mnot.net/