Hosting git repositories within the PyData organization

[Matthew Rocklin]

Hi All, 

How do we decide which git repositories to host within the PyData GitHub organization?

Today, I believe that the answer is "find someone with access and they'll do it for you if they think it's a good idea".  Are there guidelines here?  Should there be?

Thank you for your time,

-matt

[Ralf Gommers]

On Thu, Dec 28, 2017 at 12:55 PM, Matthew Rocklin <mroc...@gmail.com> wrote:

Hi All, 

How do we decide which git repositories to host within the PyData GitHub organization?

Today, I believe that the answer is "find someone with access and they'll do it for you if they think it's a good idea". 

That's correct I think.

Are there guidelines here?  Should there be?

No and yes, respectively. It's a bit of an odd mix of PyData conference content and project repos (pandas, xarray, numexpr, patsy), plus some defunct repos that should be cleaned up.

It would be great if one of the current admins could summarize the org history, that'll help with formulating some guidelines.

Cheers,

Ralf

Thank you for your time,

-matt

--
You received this message because you are subscribed to the Google Groups "NumFOCUS" group.
To unsubscribe from this group and stop receiving emails from it, send an email to numfocus+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Andy Ray Terrel]

I help manage it, but for now the github repos are mostly the conference archives. Pandas has moved off to pandas-dev and only a few active projects keep using it.

For the most part I would rather only NumFOCUS Affiliate projects are using it. Such projects agree to a default code of conduct (which could use an update) and governance under the NumFOCUS board. Other than that I don't see much need for many policies, but perhaps I'm being short sighted.

-- Andy

[Nathaniel Smith]

On Wed, Dec 27, 2017 at 3:55 PM, Matthew Rocklin <mroc...@gmail.com> wrote:
> Hi All,
>
> How do we decide which git repositories to host within the PyData GitHub
> organization?
>
> Today, I believe that the answer is "find someone with access and they'll do
> it for you if they think it's a good idea".

That's certainly how it worked in 2012 when I was looking to get patsy
out of my personal space. I wasn't actually aware of any connection to
NumFOCUS at the time, and actually had the impression that the org is
older than NumFOCUS. I could easily be wrong though.

-n

--
Nathaniel J. Smith -- https://vorpus.org

[Matthew Rocklin]

> For the most part I would rather only NumFOCUS Affiliate projects are using it. Such projects agree to a default code of conduct (which could use an update) and governance under the NumFOCUS board. Other than that I don't see much need for many policies, but perhaps I'm being short sighted.

I don't feel strongly that non-NumFOCUS projects should be part of the github org.  However I do think that that there is value to having some community maintained organization for smaller projects.  

> That's certainly how it worked in 2012 when I was looking to get patsy out of my personal space.

Patsy is, I think, a good example here.  It's clearly small-but-useful part of a greater whole.

[Andy Ray Terrel]

That is correct, the original repo was a place for pandas. I believe it was created around 2010. It became home to PyData conference sites when I set up some of the sites in 2012. Since then most of the projects have become NumFOCUS affiliated or sponsored projects and the pydata web team (Martey and I) end up manage the other web details (like dns, cdns, and some hosting services).

 

-n

--
Nathaniel J. Smith -- https://vorpus.org

[Andy Ray Terrel]

On Wed, Dec 27, 2017 at 11:23 PM, Matthew Rocklin <mroc...@gmail.com> wrote:

> For the most part I would rather only NumFOCUS Affiliate projects are using it. Such projects agree to a default code of conduct (which could use an update) and governance under the NumFOCUS board. Other than that I don't see much need for many policies, but perhaps I'm being short sighted.

I don't feel strongly that non-NumFOCUS projects should be part of the github org.  However I do think that that there is value to having some community maintained organization for smaller projects.  

So a project doesn't need to be BIG to be a NumFOCUS affiliated project. From: https://www.numfocus.org/open-source-projects/affiliated-projects/

 """

make meaningful use of NumFOCUS-sponsored tools,

have a significant and consistent community of contributors, 

have supported the open source data science computing community through contributions of code

"""

the NumFOCUS Board could add clarity around what "signficant and consistent community" means but we've added a lot of smaller projects here. 

Perhaps it would be good to have a NumFOCUS Incubator, the only challenge is that no one has signed up to help maintain such a thing. Really the bigger issue there is the cost of maintaining all the pet projects that might crop up and our mighty web team not have enough time to support them. 

[Matthew Rocklin]

Coming back to this.  

To make this general question more concrete I would like to add a sparse ndarray package to the PyData github org and maybe also use the sparse.pydata.org subdomain for a readthedocs website.

It sounds like one approach from Andy's last e-mail is send a message to in...@numfocus.org and ask them for this to be an affiliate project.  I'm happy to do this if it's a simple process.  I would hesitate to do this if there is significant long-term administrative maintenance to NumFOCUS affiliation.

Best,

-matt

[Brian Granger]

+1

>>>> an email to numfocus+u...@googlegroups.com.

>>>> For more options, visit https://groups.google.com/d/optout.
>>>
>>>
>>> --
>>> You received this message because you are subscribed to the Google Groups
>>> "NumFOCUS" group.
>>> To unsubscribe from this group and stop receiving emails from it, send an

>>> email to numfocus+u...@googlegroups.com.

>>> For more options, visit https://groups.google.com/d/optout.
>>
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "NumFOCUS" group.
>> To unsubscribe from this group and stop receiving emails from it, send an

>> email to numfocus+u...@googlegroups.com.

>> For more options, visit https://groups.google.com/d/optout.
>
>
> --
> You received this message because you are subscribed to the Google Groups
> "NumFOCUS" group.
> To unsubscribe from this group and stop receiving emails from it, send an

> email to numfocus+u...@googlegroups.com.

> For more options, visit https://groups.google.com/d/optout.



--

Brian E. Granger
Associate Professor of Physics and Data Science
Cal Poly State University, San Luis Obispo
@ellisonbg on Twitter and GitHub
bgra...@calpoly.edu and elli...@gmail.com

[Robert McLeod]

Might be prudent to post this on the PyData mailing list as well:

pyd...@googlegroups.com

In case there are people subscribed there and not here.

Robert

To unsubscribe from this group and stop receiving emails from it, send an email to numfocus+u...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "NumFOCUS" group.

To unsubscribe from this group and stop receiving emails from it, send an email to numfocus+u...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "NumFOCUS" group.

To unsubscribe from this group and stop receiving emails from it, send an email to numfocus+u...@googlegroups.com.

[Ralf Gommers]

On Thu, Jan 11, 2018 at 5:06 AM, Brian Granger <elli...@gmail.com> wrote:

+1

On Wed, Jan 10, 2018 at 7:30 AM, Matthew Rocklin <mroc...@gmail.com> wrote:
> Coming back to this.
>
> To make this general question more concrete I would like to add a sparse
> ndarray package to the PyData github org and maybe also use the
> sparse.pydata.org subdomain for a readthedocs website.

I'm +1 on this too.

>
> It sounds like one approach from Andy's last e-mail is send a message to
> in...@numfocus.org and ask them for this to be an affiliate project.  I'm
> happy to do this if it's a simple process.  I would hesitate to do this if
> there is significant long-term administrative maintenance to NumFOCUS
> affiliation.

In this particular case, I think affiliated status is neither needed nor desirable. Not needed, because the intention is for these sparse arrays to end up in scipy.sparse, and SciPy is already an affiliated project. Not desirable, because while the bar is not as high for affiliated projects as for becoming a sponsored project, we do want those projects to (a) be willing to be a project with a name, community, etc., and (b) wanting the affiliated status for more than just getting a domain name.

If you haven't gotten access to a pydata repo in a week, maybe ping Andy and me directly.

Cheers,

Ralf

 

>>>> an email to numfocus+unsubscribe@googlegroups.com.

>>>> For more options, visit https://groups.google.com/d/optout.
>>>
>>>
>>> --
>>> You received this message because you are subscribed to the Google Groups
>>> "NumFOCUS" group.
>>> To unsubscribe from this group and stop receiving emails from it, send an

>>> email to numfocus+unsubscribe@googlegroups.com.

>>> For more options, visit https://groups.google.com/d/optout.
>>
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "NumFOCUS" group.
>> To unsubscribe from this group and stop receiving emails from it, send an

>> email to numfocus+unsubscribe@googlegroups.com.

>> For more options, visit https://groups.google.com/d/optout.
>
>
> --
> You received this message because you are subscribed to the Google Groups
> "NumFOCUS" group.
> To unsubscribe from this group and stop receiving emails from it, send an

> email to numfocus+unsubscribe@googlegroups.com.

> For more options, visit https://groups.google.com/d/optout.

--
Brian E. Granger
Associate Professor of Physics and Data Science
Cal Poly State University, San Luis Obispo
@ellisonbg on Twitter and GitHub
bgra...@calpoly.edu and elli...@gmail.com

--
You received this message because you are subscribed to the Google Groups "NumFOCUS" group.

To unsubscribe from this group and stop receiving emails from it, send an email to numfocus+unsubscribe@googlegroups.com.

[Matthew Rocklin]

Update here.  I sent an e-mail to info@numfocus a couple weeks ago and haven't heard back.  

> If you haven't gotten access to a pydata repo in a week, maybe ping Andy and me directly.

I actually already have sufficient rights on the pydata github org to do what I want to do.  I just want to make sure that the community doesn't mind.  We would also be taking the name sparse in the PyData namespace (until something better comes along).  

It would be useful for us if this could be resolved quickly one way or the other.

[Andy Ray Terrel]

yes I think everyone is good with sparse.pydata.org being used for this purpose. Sorry for the delay in response.