NumPy/SciPy foundation

[Travis Oliphant]

Hello all,

I really appreciate the input from everyone. I think there is some interest, but also a lot of concern about the overhead of administration, which really translates to funding for the thing. I understand this concern very well.

Setting up and running a foundation does take effort / money, but it also serves a very useful purpose for something like the Scientific Computing community using Python. I believe it is a large enough effort that a bit more control will be needed than provided by the Software Conservancy which needs to support a large number of projects that are unrelated to Science. But, I think their model is a good one and could be used for sending support to a number of different projects all in the Scientific Software space.

I don't think there is a lack of people who could help administer the foundation given some money. It will take 3-6 months to get organized, but I will be pursuing this further. We will likely re-use the mscomp.org (Mathematical and Scientific Computing) domain name that Jarrod Millman currently has. My availability for helping with this is a little unclear for the next 3 months, but I am very hopeful to be able to contribute some money to the problem over the next 6-9 months.

If you are interested in contacting me about getting involved, feel free to send me email directly or post to this list.

Thanks,

-Travis

[Robert Bradshaw]

On Wed, Oct 19, 2011 at 9:36 PM, Travis Oliphant <teoli...@gmail.com> wrote:
> Hello all,
>
> I really appreciate the input from everyone.   I think there is some interest, but also a lot of concern about the overhead of administration, which really translates to funding for the thing.    I understand this concern very well.
>
> Setting up and running a foundation does take effort / money, but it also serves a very useful purpose for something like the Scientific Computing community using Python.   I believe it is a large enough effort that a bit more control will be needed than provided by the Software Conservancy which needs to support a large number of projects that are unrelated to Science.   But, I think their model is a good one and could be used for sending support to a number of different projects all in the Scientific Software space.
>
> I don't think there is a lack of people who could help administer the foundation given some money.

I'm not sure funding is an adequate solution to the administration
issues, especially if the load falls somewhere between someone
pitching in once a month and a part-time job. You can't hire a
secretary for 4 hours/week, but that's a lot of load for a someone to
just pick up on top of their other responsibilities (including,
presumably, the non-administrative portions of keeping the foundation
going). There are also issues like having access to legal council that
are not easily resolved by volunteers or (modest amounts of) funding.

The Software Conservancy might not be the right target, nor give us
the appropriate amount of control, but an umbrella (like a University,
the way the Sage foundation is set up) can be really handy.

[Anthony Scopatz]

On Thu, Oct 20, 2011 at 12:38 AM, Robert Bradshaw <robe...@math.washington.edu> wrote:

On Wed, Oct 19, 2011 at 9:36 PM, Travis Oliphant <teoli...@gmail.com> wrote:
> Hello all,
>
> I really appreciate the input from everyone.   I think there is some interest, but also a lot of concern about the overhead of administration, which really translates to funding for the thing.    I understand this concern very well.
>
> Setting up and running a foundation does take effort / money, but it also serves a very useful purpose for something like the Scientific Computing community using Python.   I believe it is a large enough effort that a bit more control will be needed than provided by the Software Conservancy which needs to support a large number of projects that are unrelated to Science.   But, I think their model is a good one and could be used for sending support to a number of different projects all in the Scientific Software space.
>
> I don't think there is a lack of people who could help administer the foundation given some money.

I'm not sure funding is an adequate solution to the administration
issues, especially if the load falls somewhere between someone
pitching in once a month and a part-time job. You can't hire a
secretary for 4 hours/week, but that's a lot of load for a someone to
just pick up on top of their other responsibilities (including,
presumably, the non-administrative portions of keeping the foundation
going). There are also issues like having access to legal council that
are not easily resolved by volunteers or (modest amounts of) funding.

I feel the that the funding issue is commensurate with want we want the 

foundation to *do*.  The purpose of a foundation is to build a community 

or formalize an existing community.  This can be done very cheap.

It is other activities that have higher capital requirements.  For instance, 

sponsoring the SciPy conferences, running a GSoC-esque program, putting

out a journal, etc.

I think that it is important to not 'shoot for the moon' initially.  This won't

be the PSF or the FSF overnight, nor does it need to be.  Rather, I advocate

for taking some time to set up the foundation, figure out a governance 

plan, come to consensus on worthy long-term goals, drum up institutional

backers (which Robert noted would help scale up), and proceed from there.

These steps could realistically take a year.

Such is my two cents.  Personally, I would really like to know what the foundation

would do, but that can only be known after the framework has been set up.

Be Well

Anthony

[Travis Oliphant]

Agreed that this is going to take some time to setup.   6 months to a year seems like a reasonable time frame.     I think there is plenty for the foundation to do.  These ideas were laid out in an earlier email.  I think the efforts of the foundation could employ at least 2-3 people full time.   My expectation is the foundation should plan to: 

* raise money

* sponsor sprints

* sponsor and run conferences

* issue development grants (call for proposals style)

* sponsor awards

* manage releases of NumPy / SciPy and others

This is only possible with money, of course.   But, I'm hopeful to be able help with that problem over the next year or two.   I can't promise anything at this point, but am working to find a way to get the foundation jump-started.    My main purpose in writing the email was to find people who are interested in seeing the foundation succeed. 

Thanks, 

-Travis

[Robert Bradshaw]

A foundation large enough to absorb 2-3 full-time salaries in the
overhead was admittedly larger than I had envisioned, but if magnitude
of funding can be directed to these causes that would be great.

> My expectation is the foundation should plan to:
> * raise money
> * sponsor sprints
> * sponsor and run conferences
> * issue development grants (call for proposals style)
> * sponsor awards
> * manage releases of NumPy / SciPy and others
> This is only possible with money, of course.   But, I'm hopeful to be able
> help with that problem over the next year or two.   I can't promise anything
> at this point, but am working to find a way to get the foundation
> jump-started.    My main purpose in writing the email was to find people who
> are interested in seeing the foundation succeed.

In case it wasn't clear, I am (though admittedly I don't know that I'd
be able to contribute much in the short term).

- Robert