8000 DOC Inconsistent between checklists in CONTRIBUTING.md and contributing.rst · Issue #9958 · scikit-learn/scikit-learn · GitHub
[go: up one dir, main page]

Skip to content

DOC Inconsistent between checklists in CONTRIBUTING.md and contributing.rst #9958

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
qinhanmin2014 opened this issue Oct 19, 2017 · 8 comments

Comments

@qinhanmin2014
Copy link
Member
qinhanmin2014 commented Oct 19, 2017

In both CONTRIBUTING.md and doc/developers/contributing.rst, we have a checklist which list the things to check before submitting a pull request. (See the list in CONTRIBUTING.md and contributing.rst). The two lists are almost the same (many entries are exactly the same). But both lists are missing something from each other. For example, in .md file, we have an entry to mention expected time and space complexity but that's not in .rst file. In .rst file, we mention PEP8 check but that's not in .md file. I think it might be better to copy-paste between the two files to make the two checklist consistent.
cc @jnothman @lesteve WDYT? Thanks :)

@lesteve
Copy link
Member
lesteve commented Oct 19, 2017

Just to state the obvious, straight copy and pasting will not work because one is markdown and the other one is reST. In my mind CONTRIBUTING.md is a subset of contributing.rst, that aims to be more concise and to the point than contributing.rst. Other core devs may disagree on that.

Maybe CONTRIBUTING.md could be rewritten in a nicer way with less text and more links to the dev doc (that is scikit-learn.org/dev because it will be more up-to-date than the scikit-learn.org/stable one).

For example, I am not convinced that the full PR checklist should be in CONTRIBUTING.md, same thing for things like flake8, autopep8, coverage, etc ...

@qinhanmin2014
Copy link
Member Author

@lesteve Thanks a lot for the instant reply :)
In the long run, I personally agree with your suggestions to improve CONTRIBUTING.md. But currently, I might still think it necessary to make the two list consistent (at least add what's missing from .rst file from .md file because we want .rst file to contain all the information).
If you are sure that we don't need any discussion, please close it directly.

@qinhanmin2014
Copy link
Member Author

@jnothman Could you please share your opinion on this? Do we need to make the two checklists consistent? Thanks.

@jnothman
Copy link
Member

I'm not a big fan of contributing.md. I suspect it is read much less often than the one rendered on the web site. I would much rather work improving the rst one to make the other redundant, or simplifying CONTRIBUTING.md, renaming it to CONTRIBUTING.rst and ideally including it in the docs directly.

@qinhanmin2014
Copy link
Member Author

@jnothman Thanks for the reply :)

I suspect it is read much less often than the one rendered on the web site.

For myself, I read the contributing.rst website. But recently I (surprisingly) find out that ISSUE_TEMPLATE.md, PULL_REQUEST_TEMPLATE.md and part of README.rst all choose to link to CONTRIBUTING.md file instead of contributing.rst website.

Anyway, this PR only focus on the two PR checklists. They are almost the same, but each side is missing something from the other side. It means that before submitting PR, no matter you read the checklist in CONTRIBUTING.md or contributing.rst, you will miss something. Seems that there are 3 alternatives:
(1) make the two checklists consistent.
(2) add what's missing in contributing.rst from CONTRIBUTING.md since we suppose .rst website to contain everything.
(3) simply close it if we don't care.
WDYT? Thanks.

@lesteve
Copy link
Member
lesteve commented Oct 22, 2017

I'm not a big fan of contributing.md. I suspect it is read much less often than the one rendered on the web site. I would much rather work improving the rst one to make the other redundant, or simplifying CONTRIBUTING.md, renaming it to CONTRIBUTING.rst and ideally including it in the docs directly.

A link to CONTRIBUTING.md is shown by github each time you open an issue or PR (probably only in a repo you have not contributed to, I haven't tested this extensively). Here is a snapshot to give an idea:
snapshot

As such I would bet that new contributors are way more likely to read CONTRIBUTING.md than the developers doc. Also a lot of the bullet points of CONTRIBUTING.md are specifically for these new contributors.

Also an advantage of CONTRIBUTING.md is guaranteed to be up-to-date in contrary to the developers docs. Linking to the dev doc is an option I agree.

The original github announcement mentions CONTRIBUTING and CONTRIBUTING.md so I am not sure you can have a CONTRIBUTING.rst. The announcement is 5 year old so it may be worth double-checking whether CONTRIBUTING.rst is working.

(3) simply close it if we don't care.

I am going to give you an honest answer to this one since you asked multiple times. It's not that we don't care but more "do we care enough about this to do something about it?". Core developers have limited availability and we have to make the most of their time to move the project forward. It is super easy to spend a lot of time on bikeshedding issues like this one (no clear cut obvious solution, lots of varied opinions) with no obvious sizeable benefit. Spending time on bikeshedding issues distract us from more important issues. How do we avoid this? I don't know but I think sometimes not starting a discussion is a good option.

To finish on a constructive note, here is what I would be in favour of:

  • a very short CONTRIBUTING.md with links to the dev doc for more details.
  • have links to the dev doc in the issue and PR templates.

On a separate note, contributing dev docs have way too many bullet points if you ask me and they could use some work to be clearer and more concise.

@qinhanmin2014
Copy link
Member Author

@lesteve Thanks a lot for the detailed reply :) Closing this one since the issue seems trivial.
I agree with your notes so if needed, you might instruct me to submit issue/PR for that.
Thanks again and sorry for wasting your precious time.

@jnothman
Copy link
Member
jnothman commented Oct 22, 2017 via email

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

4 participants
0