WEB: Remove NumFOCUS and Coiled as sponsors

[datapythonista]

pandas is it's not really well funded these days, and we've been updating the sponsors list recently to be more accurate, and I'd like to still remove NumFOCUS and Coiled from the list. Of course I do appreciate the support that both NumFOCUS and Coiled provided and still provide to pandas, but I think having them as sponsors is at this point misleading and create the false impression that pandas is better funded than it is.

For NumFOCUS, we pay them 15% of the pandas income for financial support mostly. We get few other things like legal support or small development grants. But I wouldn't list them as "pandas has the support of NumFOCUS" in periods of very low funding like now, as we probably supported NumFOCUS more than NumFOCUS supported us. For reference OpenCollective US would charge as 10%, and OpenCollective EU 8%.

For Coiled, I saw Patrick made two commits in the last 7 months, and if I'm not wrong he's not very active in PR review or other maintenance tasks recently. Truly thankful to them as Patrick could make lots of quality work for pandas sponsored by them in the past, but as of today, having them as a sponsor it's again misleading visitors of our website into thinking pandas is in a healthy financial state as it was a couple of years ago, when in practice it's not.

I hope having a more accurate list can help find new sponsors, help when we ask new grants, and maybe even help bring awareness of our current sponsors how key is their support at this point.

@pandas-dev/pandas-core @phofl please let me know if any objection, otherwise I'll move forward with this in the next few days.

[WillAyd]

I don't think NumFOCUS should be removed - even if not financially they still help us with legal and administrative tasks as an organization.

I have no opinion on Coiled

[datapythonista]

I don't think NumFOCUS should be removed - even if not financially they still help us with legal and administrative tasks as an organization.

They help us because we paid them more than $100k+ for it. Many people I spoke with believe pandas receives funds from NumFOCUS, when the money goes in the opposite direction. I just don't think creating the false impression that NumFOCUS is funding our development at a time when pandas is mostly a single maintainer project is very helpful. That's my point, but if people like to have their logo in the list of sponsors, no big deal.

[attack68]

I don't think NumFOCUS should be removed - even if not financially they still help us with legal and administrative tasks as an organization.

They help us because we paid them more than $100k+ for it. Many people I spoke with believe pandas receives funds from NumFOCUS, when the money goes in the opposite direction. I just don't think creating the false impression that NumFOCUS is funding our development at a time when pandas is mostly a single maintainer project is very helpful. That's my point, but if people like to have their logo in the list of sponsors, no big deal.

I had exactly this impression until I read your post. I agree with you

[rhshadrach]

Many people I spoke with believe pandas receives funds from NumFOCUS, when the money goes in the opposite direction.

Sponsors support a project. I think the error here is thinking that support only consists of funding (including indirectly by allowing employees to maintain the project on paid time). I don't have a good sense of how much support NumFOCUS provides, but it seems to me removing them as a sponsor just because people think that means they are funding pandas is wrong.

They help us because we paid them more than $100k+ for it.

Do we have a sense that they have not provided support that makes this worth it?

@datapythonista - I also removed the good first issue tag. Agreed this will be a good first issue once the direction forward is decided, but since new contributors use this to find issues to work on now we should wait until then.

[phofl]

Sound good to me re coiled (I actually left coiled a few weeks ago anyway but didn’t get around to do that yet)

[datapythonista]

Thanks all for the feedback.

Do we have a sense that they have not provided support that makes this worth it?

I think this opens a separate discussion, and I don't have enough information to be strongly opinionated about it. Based on my experience, I don't think the service we're getting is as good as it should be. As an example, last contract one of us had to sign it took more than 2 months of waiting for NumFOCUS before it was ready. I also was quite pissed off when few of us (pandas maintainers) considered teaching pandas in a somehow official way to help the project raise funds (and be paid for it), and this was blocked by NumFOCUS. I also had a quite bad experience when I volunteered to help with NumFOCUS with their infrastructure and most of my work was boycotted and thousands of dollars were wasted for no reason I could understand.

So, for now, since our income is almost nothing, we aren't paying much to NumFOCUS, so I guess it's worth. If at some point we get an annual income of $200k, I would personally save $10k on fees and move to OpenCollective US. But I don't think there'll be consensus, so I won't lead the effort.

Back to the discussion here on whether NumFOCUS supports us. It's to me as if we listed AWS as a pandas supported because we spend $30k a year in AWS credits, and they support us with some virtual machines because of it. I think the only difference in both cases is that we are the ones who are happy to support NumFOCUS. I was personally happy with it when we had 10 sponsors, a $200k budget, and more money for maintenance that we could spend. But right now we have Matt hours from NVIDIA, $500 a month from Tidelift, and the ongoing Bodo grant (for features, we finished the maintenance hours they granted us).

All non-volunteering work for pandas I'm aware of is paid by NVIDIA, Tidelift and Bodo. I'd like the community, organizations considering giving us a grant, companies considering supporting us... to see that in our website, so they can make an informed decision.

[WillAyd]

Sorry to hear about all the negative experience with NumFOCUS. I know that they too have undergone a lot of change within recent memory, so there's definitely a lot to be critical of.

On the flip side I think NumFOCUS has been recently helpful with trademark enforcement (pandasAI) and with the legal agreement to get a royalty donation from Packt from the sales of my book. I don't want to discredit entirely where they help, and since we haven't gotten many funds within the past year or so, they in turn have not received much financial support through us.

If we collectively wanted to reassess our relationship with NumFOCUS, I suggest we try and restart that conversation with someone there like Nicole. If that doesn't yield whatever it is we are looking for, then I think a PDEP to explore alternatives would be warranted (similar to what Jupyter did when they left NumFOCUS for the Linux Foundation https://jupyter.org/governance/linux-proposal.html)

For now though, unless someone is willing to lead that charge, I think we should still consider them a sponsor and partner

[datapythonista]

I won't consider them a pandas sponsor since they are not. ;) But agree with all you said.

I'll close here, since Coiled has been removed, and it seems there is no agreement to remove NumFOCUS as a sponsor. Thanks all for the feedback.