relationship with projects

[Matt Terry]

Hi,

What is the relationship between NumFOCUS (NF) and its associated projects?  Reasonable people might want to know if NF "owns" numpy, or its IP, or something else.  I know that to be associated with NF, we need a couple representatives from each projects community.  There may be some legal stuff we offer if we are sponsoring development.

In general, it would be good to have an "About" page that explains (in plain English) why we are here, how we help the community, and why you should support us.

I started a bug [1] on this topic if folks want to chime in there, too.

[1] https://github.com/numfocus/numfocus.org/issues/30

-matt

[Jacob Barhak]

Hi Matt,

What kind of legal support do you mean?

I am asking just out of curiosity. 

You were mentioning IP and ownership. These are sometimes hard to determine in open source community - especially with older code. There may be a need to reach current consensus in some cases and advertise to see if this consensus is contended by anyone - you may not have a clear cut resolution in all cases.

Never the less, you can formulate new rules to help navigate new participants. This may help in the future. 

       Jacob

Sent from my iPhone

--
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/groups/opt_out.

[Matt Terry]

If I were a company considering sponsorship, I'd want to have a general idea of how my donation is being used.  What benefits it provides to the community and making sure it wasn't secreting supporting my competitors or some such.  basic due diligence information.

Likewise, if I were a part of a project and was considering a NumFOCUS (NF) affiliation, I'd want to know how much control I'd have to assign to NF and what benefits I'd get in return.

I'd like the answers to be along the lines of:

NF does not own or direct any affiliated projects.  The projects and their IP belong to the community.  NF may provide funds to assist development, but the community decides the direction for the project and whether sponsored development is even accepted.

To be affiliated with NF, NF needs two community members willing to represent the community.  Affiliated project are eligible to receive grants from NF to support sprints, testing, and otherwise assist development.  If assigned the trademark to a given project, NF can defend that trademark against attack.  At some level of affiliation, NF is able to receive donations on behalf of the community and disburse those funds as directed by the community.

Note: there are my thoughts and opinions, not NF's

-matt

[Travis Oliphant]

This is a very nice start for the language that very much describes what NF is about and what it does .

I am +1  about putting this language on the website.

[Jarrod Millman]

+1 from me as well.  We are still working on what (if any) formal/legal relation we will have with various sponsored projects.  For now, this seems like a good description.

Thanks,

Jarrod

[Anthony Scopatz]

For some historical perspective, I think that what this definition has been has changed over time and is currently in a state of transition. Original associated projects where simply ones than we thought we would like to fund eventually.  Then it became projects where the leads agreed to put their name and logo on our website. Now we hope to have FSAs signed with certain projects so 'associated' potentially means a lot more.

Be Well

Anthony

[Frédéric Bastien]

I understand that NF moving. But I'm not sure sure we need absolutely some signature in all cases. For example, what if a community is divised? Probably it is better that the sponsored work is one agreed part of the community, but it would be sad to end up in a case where nothing can be done due to divergence. It is not ideal, but I think it should be allowed for NF to support only one side of a divided community. I trust NF to handle those not easy case in the best way they think. But making rules that don't allow NF to handle some hard cases isn't good I think.

Also, about the version of text by Matt, why put the need that 2 community members agree to represent the community? If a project is just starting and have only 1 member, NF can't support it? What if only one member have time to represent the community? I trust NF to select wisely the project and I'm not sure it is good to have this limitation. An alternative is to have multi level of association with projects. No much restriction for the entry level, but this allow just low level of support from NF, to help start the project for examples. Then a second level for project with more money?

Another questions, how do you define community? On a per software/package level? Or at the python scientific community level? Do the representent of the community member need to contribute to the code? To the idea/direction of the project?

This is just questions, I don't have the answers.

Fred

[Anthony Scopatz]

On Wed, Sep 18, 2013 at 10:40 AM, Frédéric Bastien <no...@nouiz.org> wrote:

I understand that NF moving. But I'm not sure sure we need absolutely some signature in all cases. For example, what if a community is divised? Probably it is better that the sponsored work is one agreed part of the community, but it would be sad to end up in a case where nothing can be done due to divergence. It is not ideal, but I think it should be allowed for NF to support only one side of a divided community. I trust NF to handle those not easy case in the best way they think. But making rules that don't allow NF to handle some hard cases isn't good I think.

Hi Frédéric, 

Nothing prevents anyone from applying to NumFOCUS for funds, whether they have signed an FSA or not.  The FSA is really about establishing a long term relationship with projects and possibly holding assets such as grants, trademarks, and TLDs on behalf of the projects. For the case of a divided community, we would very carefully which each issue as a separate task.  And in the case of a Grantor/Grantee model, each issue is a separate task anyway so this is not that different.  Hopefully this won't happen.

 

Also, about the version of text by Matt, why put the need that 2 community members agree to represent the community? If a project is just starting and have only 1 member, NF can't support it? What if only one member have time to represent the community? I trust NF to select wisely the project and I'm not sure it is good to have this limitation. An alternative is to have multi level of association with projects. No much restriction for the entry level, but this allow just low level of support from NF, to help start the project for examples. Then a second level for project with more money?

It is a consensus of the board that you see reflected here.  We generally feel that a 'community' is more than one person and more than one institution.  If a project cannot find 2 - 3 individuals who work for different institutions who are willing to stand up and take ownership of a project, the project is probably too young or underutilized to warrant the kind of fiscal sponsorship discussed here.   In essence, the project hasn't 'made it' yet.  

Still, this doesn't preclude individuals from applying for grants from NumFOCUS for their projects.  It is just that a separate legal agreement is redundant (individuals are legal entities) and not worth the overhead.

 

Another questions, how do you define community? On a per software/package level? Or at the python scientific community level?

I think that we are in the process of defining that here =)

 

Do the representent of the community member need to contribute to the code? To the idea/direction of the project?

This is largely up to the discretion of the project itself.

Be Well

Anthony

[Frédéric Bastien]

thanks for the clarification that there is already 2 levels. It make more sense to have such type of check for the higher level of engagement.

thanks

Fred

[Radim Řehůřek]

Hello Anthony,

this topic and your reply are extremely interesting for me. I'd like to promote gensim, the "topic modelling for humans" software [1], with sprints and tutorials at conferences too [2].

What are the concrete steps I should take to get NumFOCUS support?

Cheers,

Radim

[1] https://github.com/piskvorky/gensim/

[2] http://radimrehurek.com/2013/10/five-years-of-gensim/

[Anthony Scopatz]

Hello Radim, 

You should make a concrete proposal to the board.  We don't have a standard template for this yet, but if you want to develop one, that'd be great too!

Be Well

Anthony

[Radim Řehůřek]

On Wednesday, November 6, 2013 10:18:06 AM UTC+1, Anthony Scopatz wrote:

Hello Radim, 

You should make a concrete proposal to the board.  We don't have a standard template for this yet, but if you want to develop one, that'd be great too!

I clicked the "Board" tab on the NumFOCUS website, but I saw no contact/proposal info there.

What should that proposal look like and where should it be made?

I can release it as a template once I know what it is, no problem.

Cheers,

Radim

[Anthony Scopatz]

On Wed, Nov 6, 2013 at 1:33 AM, Radim Řehůřek <m...@radimrehurek.com> wrote:

On Wednesday, November 6, 2013 10:18:06 AM UTC+1, Anthony Scopatz wrote:

Hello Radim, 

You should make a concrete proposal to the board.  We don't have a standard template for this yet, but if you want to develop one, that'd be great too!

I clicked the "Board" tab on the NumFOCUS website, but I saw no contact/proposal info there.

What should that proposal look like and where should it be made?

I can release it as a template once I know what it is, no problem.

Hello Radim,

So the problem is exactly that there is no concrete proposal form.  I think you would be doing us all a huge favor if you helped create this form in the first place.  I would do this as a pull request to the website repository: https://github.com/numfocus/numfocus.org

Here are some things that probably need to go into the form:  

* Project Name

* Submitter Name

* Request Type (sprint, travel, infrastructure, development time)

* Request amount / service

* Location

* Justification (this should be a paragraph or two paragraphs to few pages.)

I think that is basically all we need.

Be Well

Anthony

[Radim Řehůřek]

So the problem is exactly that there is no concrete proposal form.  I think you would be doing us all a huge favor if you helped create this form in the first place.  I would do this as a pull request to the website repository: https://github.com/numfocus/numfocus.org

Here are some things that probably need to go into the form:  

* Project Name

* Submitter Name

* Request Type (sprint, travel, infrastructure, development time)

* Request amount / service

* Location

* Justification (this should be a paragraph or two paragraphs to few pages.)

I think that is basically all we need.

Will do. Thanks again Anthony.

Radim

[Radim Řehůřek]

Hello again Anthony,

I just submitted a pull request for the proposal form to github. And immediately used it to submit the gensim proposal, to test it :)

I went with a simple email template, available from the Projects page. You probably had something fancier in mind, but it seems the site is completely static, so I'm not sure what kind of backend form processing is possible/needed.

Best,

Radim

[Anthony Scopatz]

Thanks Radim, 

The only thing fancier I possibly had in mind was a google survey / form.  Not sure that it strictly needed though.  I'll let Matt review the PR.

Be Well

Anthony

[Matt Terry]

I haven't forgotten about it.  I've just been busy on several different fronts.

[Anthony Scopatz]

Not to worry Matt!

[Radim Řehůřek]

Thanks! I was actually starting to worry the email got lost somehow :)

Radim

[Radim Řehůřek]

A quarter-of-a-year bump.

Any progress?

-rr

[Radim Řehůřek]

Hi all,

a four month bump.

Keep well,

Radim

[Anthony Scopatz]

Hi Radim, 

Sorry that this has gotten so lost.  Hopefully the board can discuss this this weekend.

Be Well

Anthony

[Andy Ray Terrel]

We did discuss it and I'm sitting down with a web developer today to
see about updating the site appropriately.

-- Andy

[Radim Řehůřek]

Five month bump.

How did the discussion go Andy? Has the board reached a decision yet?

I thought you might humor me on April Fools at least :)

Keep well,

Radim

[Radim Řehůřek]

Half a year "anniversary" bump.

Time to thank NumFOCUS for their exemplary support and promotion, I hope I can continue bumping this thread until I'm 50! :)

Keep well,

Radim

[Andy Ray Terrel]

What exactly are you bumping about? I can't tell anymore.

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

[Anthony Scopatz]

Also, Radim, 

If you'd like to me more engaged and proactive about numfocus promotion, we'd love any help you could provide.

Be Well

Anthony

[Radim Řehůřek]

On Sunday, May 4, 2014 4:22:21 PM UTC+2, Andy Terrel wrote:

What exactly are you bumping about?  I can't tell anymore.

Now, that's a reply in style :-)

Andy, you can scroll up the thread.

Or check the email I sent numfocus half a year, using a proposal form I was asked to contribute to you then (because you didn't have a "project proposal" form).

Keep well,

Radim

[Radim Řehůřek]

Anniversary 1 year bump! :)

Any progress on this? I'm still interested in getting help from NumFocus.

And I'm still curious what that "helping open source Python projects" means in practice, and whether any other project managed to surmount the NumFocus wall of silence (and how they did it).

Best regards,

Radim

[Andy Ray Terrel]

Hello Radim,

Our website goes through the various projects we support (
http://numfocus.org/programs/index.html ) . Including:

- Fellowship for open source contributors
- Help for travel to conferences
- Running the PyData conference
- Hackathon support
- Fiscal sponsorship for organizations (
http://numfocus.org/projects/join-as-project.html )

As you can see on our projects page, we have quite a few projects we
are currently working with. ( http://numfocus.org/projects/index.html
)

-- Andy

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

[Radim Řehůřek]

Thanks Andy.

As a reminder, some years ago I asked for support, and helped NumFOCUS create your "proposal form" for new projects in the process.

I never heard back, and after a few polite bumps, I admit I started making fun of the "support" in not-so-polite way.

If you're serious about supporting the Python data world, I'm still open to collaboration of course. I'm co-organizing a Prague meetup, I sponsor Python open source projects financially (apart from maintaining my own), and would be happy to see the Python world thrive.

Replies like "What did you want again?" or "How about you promote us instead?" are not helping, I imagine the collaboration more two-sided.

Regards,

Radim

[Andy Ray Terrel]

On Tue, Dec 2, 2014 at 3:39 AM, Radim Řehůřek <m...@radimrehurek.com> wrote:
> Thanks Andy.
>
> As a reminder, some years ago I asked for support, and helped NumFOCUS
> create your "proposal form" for new projects in the process.
>

AFAIK, that proposal form has been place at
http://numfocus.org/projects/join-as-project.html

I do not find your gensim proposal in my inbox. Please, send it to us again.

> I never heard back, and after a few polite bumps, I admit I started making
> fun of the "support" in not-so-polite way.

Sorry our support was not up to your satisfaction, unfortunately, we
do not have a paid staff and rely completely on volunteer efforts for
supporting our community. We are a very small organization with a lot
of balls in the air. We could certainly use help supporting folks if
you are drawn to doing so. Right now our focus has be on a handful of
projects and the programs I have mentioned below.

>
> If you're serious about supporting the Python data world, I'm still open to
> collaboration of course. I'm co-organizing a Prague meetup, I sponsor Python
> open source projects financially (apart from maintaining my own), and would
> be happy to see the Python world thrive.
>

I'm not sure why you would say we aren't serious.

> Replies like "What did you want again?" or "How about you promote us
> instead?" are not helping, I imagine the collaboration more two-sided.

Once again we are a small volunteer organization, perfection will
never be achieved. Your queries to the NumFOCUS community have come
through many channels and are not easy to follow. Since we are a bit
overwhelmed with our current activities, tracing down and finding what
the heart of your queries really is turns to be a bit too much. Just
as you deserve an answer to your queries, we deserve a well
articulated question that helps us know how to answer. But rather
than help us answer your query, creating a tradition of poking fun of
our small organization does not empower our volunteers to stop what
they are doing and interrupt that fun.

-- Andy

[Katy Huff]

So, Radim and Andy,

I've been watching this thread and I just had a Eureka moment. I now understand both of your reactions for the first time. It looks like a big misunderstanding. 

Until andy's response just now, I was also wondering, like you, Radim, why NumFocus was dragging their feet reviewing your proposal after having promised in March that they would talk as a board about it.

But, they never were! They were reviewing your pull request with the board and didn't ever receive your proposal. This is what I've now finally understood.

That is In november 2013, Radim said:

"

>>>>> I just submitted a pull request for the proposal form to github. And

>>>>> immediately used it to submit the gensim proposal, to test it :)

"

And Anthony said "Thanks". But, no one followed up on the gensim proposal. When Radim bumped, he got a reassurance from anthony:

"

Hi Radim,

Sorry that this has gotten so lost.  Hopefully the board can discuss this
this weekend.

Be Well

Anthony

"

Radim (and I, actually) probably thought this meant that they were going to discuss your gensim proposal, submitted through the form. However, from Andy's response just now, it's clear that they merely discussed *your pull request* at the board meeting.

Probably this is clear to both of you as well, so this email is a little silly. But, in the interest of everyone seeing where everyone else is coming from (and not being combative anymore), here's a timeline with the key moment in red (I think this moment was missed completely by numfocus):

Timeline:

• 9/2013 
• Matt Terry opens an issue about there not being a very detailed project page.  https://github.com/numfocus/numfocus.org/issues/30
• 9-11/2013
• There is some discussion
• 11/2013
• Radim chimes in on the list to ask "What are the concrete steps I should take to get NumFOCUS support?" for his gensim project. https://github.com/piskvorky/gensim/ http://radimrehurek.com/2013/10/five-years-of-gensim/
• Anthony responds : There is no concrete application, but there is some basic information we need. Please just email it. Suggests that if he wants to help out, Radim can help create the project pages Matt needs or make a google form or something. 
• Radim makes a pull request creating the project proposal form page: https://github.com/numfocus/numfocus.org/pull/62
• Radim submits proposal through the form and begins his long wait for approval. "I just submitted a pull request for the proposal form to github. And immediately used it to submit the gensim proposal, to test it :)"
• 2/2014
• Radim bumps the thread. Gets no response.
• 3/2014
• Radim bumps the thread. Anthony says: "Sorry that this has gotten so lost.  Hopefully the board can discuss this this weekend."
• Andy says the changes have been pushed live to the project page (the one you just directed Radim to, actually, as if he'd never seen it) and closes the pull request promising to open another one with a google drive solution . Andy thinks the discussion is over, but Radim is still waiting for the board to look at his gensim application.
• More bumps hereafter.

https://github.com/numfocus/numfocus.org/pull/62

http://katyhuff.github.com

[Naveen Michaud-Agrawal]

To clarify Katy's excellent timeline, Radim is bumping a request for NUMFOCUS  funding support for his Gensim project (http://radimrehurek.com/gensim/) which to quote from the page, is a python machine learning library for unsupervised learning (while looking through the history of this thread, I noticed that Gensim was only explicitly mentioned a few times). Unfortunately it seems general discussion about Numfocus supported projects was mixed in with Radim's explicit request regarding Gensim (and his subsequent bumps lacked that clarification, so confusion regarding the purpose of his bumps is understandable if one just looked at the subject heading).

Naveen

...

[Andy Ray Terrel]

Hi Katy and Naveen!

Thanks for the summary. I'll make sure the topic of Gensim as a "other
supported project" will be discussed at our next board meeting. To
have a more enduring legal status, Fiscal Sponsorship, we should find
a time to have a call about it. The website explains what it is and
there is a lot of documentation on the web about FSAs, but its usually
best to go over what it means in a longer phone call.

-- Andy

[Radim Řehůřek]

Katy is a treasure!

Andy, I can look up the old emails & resend.

In fact, a lot has changed since last year, so I'll modify the info accordingly. (The "request" wasn't only about gensim, I was looking for more personal involvement with the community too, though that part is less relevant now.)

Best,

Radim

[Andy Ray Terrel]

Hello Radim,

It would be great to have you send those updates. I can forward them on to the board.

-- Andy

[Skipper Seabold]

Was just reading this article [1] and thought about this old thread.

"The most commonly used implementation of word2vec is in the
open-source gensim project (with some growing interest in the Spark
implementation as well). Most of these projects are just scraping by
without a stable source of funding, which seems ridiculous. It seems
that few of these open data and software communities are compensated
by the (large) corporations that use these tools. This is perfectly
legal of course given the licensing but one wonders if there isn't a
better funding model for all of us to consider in the future (like at
bountysource)."

Skipper

[1] http://www.wise.io/tech/five-takeaways-on-the-state-of-natural-language-processing

[Stefan van der Walt]

On 2015-04-30 10:03:09, Skipper Seabold <jsse...@gmail.com>
wrote:

> "The most commonly used implementation of word2vec is in the
> open-source gensim project (with some growing interest in the
> Spark implementation as well).

I'm curious whether the conversation between Andy and Radim ever
happened, and whether gensim signed the Fiscal Sponsorship
agreement?

Stéfan

[Andy Ray Terrel]

No I never got a proposal. I assumed interest was lost.

[Stéfan van der Walt]

On Thu, Apr 30, 2015 at 7:20 PM, Andy Ray Terrel <andy....@gmail.com> wrote:
> No I never got a proposal. I assumed interest was lost.

Your last message to Radim states:

'''

Thanks for the summary. I'll make sure the topic of Gensim as a "other
supported project" will be discussed at our next board meeting. To
have a more enduring legal status, Fiscal Sponsorship, we should find
a time to have a call about it. The website explains what it is and
there is a lot of documentation on the web about FSAs, but its usually
best to go over what it means in a longer phone call.

'''

Action items from that paragraph:

- Discuss support of Gensim as "other supported project"
- Find a time to have a Skype conversation

So, it sounds like the ball is in the NumFocus court, at least to
report back on the meeting (or to discuss it, if that discussion was
never held).

Stefan

[Gael Varoquaux]

On Thu, Apr 30, 2015 at 11:26:45AM -0700, Stefan van der Walt wrote:
> I'm curious whether the conversation between Andy and Radim ever happened,
> and whether gensim signed the Fiscal Sponsorship agreement?

The fiscal sponsorship agreement of numfocus isn't exactly something so
trivial. When I looked at it, it was too binding, and it wasn't clear
that it was in the interest of an independent project to sign it.
Specifically, in the wording of 6 month ago, the following section were a
problem:

3c. NumFOCUS Activities on Behalf of the Project. Subject to §3(b) of this
Agreement, all community programs, fundraising events, processing and
acknowledgment of cash and noncash revenue items, accounts payable and
receivable, negotiation of leases and contracts, disbursement of Project
funds (including grants), and other similar activities planned by the
Project shall be the ultimate responsibility of NumFOCUS and shall be
conducted by NumFOCUS or by the FIXME­LEADERSHIP­BODY­NAME with oversight
from the Board. These and other services provided by NumFOCUS are
described in §4.

3e. The FIXME­LEADERSHIP­BODY­NAME Acting on Behalf of NumFOCUS. When
raising money for the project the FIXME­LEADERSHIP­BODY­NAME is acting on
behalf of NumFOCUS and subject to Board of Director oversight.

In other words, with regards to fund raising, the project is handing out
it's governance to numfocus.


The reason that this is written this way is to protect the non-profit
status of numfocus with regards to the IRS: wrong actions taken by
projects in the name of numfocus would put numfocus at risk. However,
these provisions are very binding for the project and the project core
members.

Things aren't always easy :).

Gaël