I have questions regarding the interpretation of the governance related to contractual work.
You should think about it in terms of individual council decisions: if you are voting on an issue that may be influenced by contracting work you are doing (or the vote will influence your work), then you need to declare the conflict. If the conflict is serious, you should recuse yourself from the vote.
Since NF is our financial sponsor, its hard for me to think of a potential conflict with receiving money from them. Perhaps if the NF contract was related to a different project?
Thank you but I don’t know if I should have used the terms “conflict of interest” which are perhaps another issue, because what I meant was to ask specifically about this
the limit for council members reporting to the same person/company is 2 according to our governance
and what precisely is considered “reporting to a person/company”
Working for the same company.
I have moved this to its own topic so we can have a proper discussion on this. The governance currently has this:
No more than 2 Council Members can report to one person or company (including Institutional Partners) through employment or contracting work (including the reportee, i.e. the reportee + 1 is the max).
This was inspired by a condition in the ArviZ-NumFOCUS sponsorship agreement (which I assume is not too different than the PyMC one but I have only seen the ArviZ one):
A majority of the ArviZ Random Variables Council should not be employed by the same entity or share a common affiliation beyond that of the Project.
And both aim to be a safeguard to keep the project a community driven one. Given the current situation with PyMC Labs, we might want to ask NumFOCUS about guidance on how to manage this, but in general I think aiming to have enough representation on the council with no affiliation whatsoever to Labs is a good idea.
Regarding NumFOCUS contracting, we might want to be explicit on this too, but I don’t think it counts as when contracting for numfocus, we don’t really report to numfocus. What is to be done and how much is paid is decided by ourselves (as in pymc team/consensus among core contributors)
Numfocus/ GSOC are grants, which Labs also provides. Those should probably not count as being affiliated with the company providing the grants.
Contractor work, if it’s the main source of income should definitely count. Most labs data scientists fall under this category. If it’s only a secondary source of income it may need a more precise guideline.
Ideally we can get unaffiliated people involved, but there’s also the risk of scrapping the bottom of the barrel if all the people deeply involved with the project are collaborating with labs.
Previous work should not count as affiliation, at least after some agreed upon time passed.
Thank you, this is the kind of answer that I was looking for. I indeed wish that the governance was clearer on the case of people doing contracting work as a secondary activity.
I disagree with most of this.
I find the statement “NumFOCUS is grants” between confusing and misleading.
PyMC basically pays NumFOCUS to act as our legal umbrella, manage the project money and ensuring we are complying with any legal requirements related to that. When funders like CZI award a grant to PyMC it isn’t a wire transfer to our accounts, there is much more overhead and bureaucracy involved, which NumFOCUS takes care of (and consequently keeps a small percentage of any grant PyMC gets to pay this overhead). Focusing strictly money-wise, that is the main thing we get from NumFOCUS, without that we wouldn’t have been able to receive the CZI grant at all.
It is confusing though because NumFOCUS doesn’t limit to doing only these administrative tasks, it also seeks its own grants/sponsors, and organizes events like PyData. And then distributes part of the benefits it can get from that back to the project via small development grants. And PyMC has also gotten a couple of those. They represent less money than CZI and also than funds from PyMCon sponsors but they would be “numfocus grants” strictly speaking; however, I think this isn’t what was meant above I think and it was meant for any money paid via NumFOCUS.
All the money that gets paid via NumFOCUS is allocated by the project[1], and NumFOCUS only has a say on these movements in making sure we are not violating conditions of the grant or legal requirements as a non-profit. The grant proposal is written as PyMC project and it rebudgeting is needed, it is also discussed and agreed on as project wide decision.
GSoC (and similar initiatives like Outreachy for that matter) is irrelevant here as they are specific to newcomers in OSS which would make anyone receiving such stipend not eligible as council member. Still, it starts with a project wide generated list of ideas, candidates then submit proposals and who gets selected is again decided by the project.
Whatever the structure and similarity between the grants like CZI that have been awarded to PyMC and Labs there is a key difference which I think is the main thing that should be taken into account when considering to whom does someone report: Labs decides who gets the money, how they get it and why. Small note to make super clear that there is nothing wrong with that, in fact it is encouraged even and one of the main things that makes labs an institutional partner.
I might try to frame that more as a “minor activity” than secondary. I know this is quite subjective so should not be verbatim what goes into the governance, but if having or not the income from the extra contracting gig is relevant to the steering council member then it is close to impossible for them to remain unbiased. The goals of Labs, Intuitive Bayes (of which 3 candidates are also part of) or Google (cutting it close at 2 candidates but had 3 nominees too) might not be the same as the goals of the project at all times. But as @Armavica used as extreme example in the first post, receiving a 1€ payment should probably not count
I completely agree with that, adding a caveat though which is we need to combine this with isolating “top of the barrel” people who are not affiliated with Labs if PyMC project decisions end up happening at Labs. We might be closer to this second situation despite efforts to the contrary (with both of you being among the most dedicated to ensuring PyMC discussions and decisions happen in PyMC spaces).
Sometimes it is more a “should be” than an “is” because the governance is not always completely followed, but I think this is a different issue ↩︎
What do you disagree with? That two people receiving money through Numfocus/GSOC shouldn’t count as being affiliated with the same company?