Code of Conduct for Stan

I like it! I picked the one from my draft just because it’s the one I saw and said, “Yes this seems reasonable”. The NumFocus one seems good as well! I can remove mine from that draft and once we have our official one sorted out I’ll add a link to it in Stan Math.

1 Like

I have pinned this to the banner. I can un-pin if you like but this seems very important.


Great! It looks similar? Is it the NumFocus one? Or do they all just use similar wording?

I think this is ultimately a SGB decision, but they seem super busy so maybe we can write a proposal to make it easier for them to decide?

What do you think of the idea of a central code of conduct (perhaps even the numfocus one), and “interpretations” of it for specific subgroups?

@breckbaldwin thanks for the pin! I think you can only pin one thing at a time though. What if we had a thread that was pinned that was “important topics for Stan users atm” and had links to the threads. I think @martinmodrak’s intro is pretty important too and don’t want it to get lost.

1 Like

Ok I tried to create a weekly page. @martinmodrak should probably run it. Important Stan Topics This Week


1 Like

it’s a slightly modified version of the one from I added some things from tensorflow and gtests code of conduct that I like (the sections on conflict resolution and ‘generally don’t be a jerk’)

Yeah def! How do you want to go about it? We could make a fork of the project proposal repo

A few random thoughts on @stevebronder’s proposal :

  • It should be clear who “project maintainers” are. That’s not a term that’s publicly defined afaik.
  • It needs to be clearer that there is a difference between “harassment” (which should not be expected to be resolved between the two parties) and “conflict” (which could conceivably be). Reporting harassment can be very hard (depending on who in the community you are reporting and what they have done) and I worry that making people have to separate between the two puts an additional barrier to getting help.
  • Question about Scope: What happens if someone asks me (a person associated with Stan) for Stan help in a non-Stan-controlled forum (eg twitter) and I do something that clearly violates the CoC (eg use extreme sexual language). As I read the scope section, this is not covered (the event is not an sanctioned Stan event so I am not officially representing Stan). Should it be?
  • Enforcement: There needs to be more than one point of contact. IIRC board@mc-stan goes to Daniel Lee and Breck. What if one of them violated the CoC?
  • What is a Project Steward?
  • Why is notifying the harasser not decided in consultation with the person making the complaint?
  • The range of potential consequences are too vague. People should know what to expect.
1 Like

where is this proposal?

Link is here:

so this only applies to the Stan math library?

I think the idea is for this to apply to the whole project. (It’s got languae in it about covering all of stan)

I also worry about codifying “I was having a bad day” an an excuse for bad behaviour. I think that is a very very very poor idea. The onus of a CoC should be on the effects of the action. People are very good at justifying their poor behaviour and if we literally lead with that as an option it has the potential to encourage us to treat Person A as someone who was having a bad day, but Person B as someone who violated the code.

1 Like

I’ll stop spamming the thread now, but I think the NumFOCUS Code of Conduct Lauren linked at the top of the thread is a much richer and more detailed CoC that seems like something that could be adopted almost entirely.

  • Project Steward == Project Maintainers I agree that wording is to vague. Should we define Project Maintainer as "Anyone who is authorized to accept and merge pull requests on the stan-dev repo. I’d like wording to add non-programming contributors but I’m not sure on the verbage. We could make it anyone in the Stan electorate.

Agree! Tensorflow lists two people to email. Is there someone at NumFocus we could have as a neurtral 3rd party to email? And/Or alternatively several possible people to email

A more clear definition here would help clear up this line

If the behavior is threatening or harassing, or for other reasons requires immediate escalation, please see the section on enforcement for escalating the issue.

It’s hard to define harassment vs. conflict. In my brain something relative to magnitude and volume makes something go from conflict to harassment. imo a good solution may be have multiple easy to access channels to escalate. Would lead to more false negatives but a higher capture rate.

Or we could remove the section on trying to sort it out between the two parties

Probably right? We could expand the wording from “Project Maintainers” to ‘Recognized members of Stan’? Or something larger in scope

The wording comes from the original Contributor Covenant CoC. I think it’s meant to make clear that if someone goes through the official channels about harassment we won’t be sending it to everyone willy nilly.

This line here?

Project maintainers who do not follow or enforce the Code of Conduct in good faith may face temporary or permanent repercussions as determined by other members of the project’s leadership.

We can could add some bullet-points. tbf I’m not super good at knowing about these those sort of conflicts and the judiciary process around them. But the doc is up as a PR so if you want to add anything there do feel free!

I saw we didn’t have a CoC and didn’t know if anyone was working on one so while I was making the for stan math I worked on the CoC as well. Mostly wanted to get the ball rolling. That and tbh the math library is like 98.23% commits from dudes and I’d like us to be more diverse. I figured a CoC and good contributor guide are a helpful step for that

eod I believe the idea is to have one CoC that covers the whole project

Yeah reading that twice I agree we can get rid of that. I put that in since I saw it in the Tensorflow and MxNet CoCs

It’s appreciated! The NumFOCUS one is good to me.

This is NumFOCUS’s version of the same thing

Be careful in the words that we choose. We are careful and respectful in our communication and we take responsibility for our own speech. Be kind to others. Do not insult or put down other members of the community.

I think it’s incredibly important to make sure that taking responsibility for the effects of our speech and actions is at the core of a CoC.


is it not the case that in order to be a part of NumFOCUS, we have to adopt the NumFOCUS Code of Conduct? so that is the de facto CoC already?

1 Like

I also think that, while everyone should be held to a high standard, people who are being paid to work on Stan by Stan should be held to very stringent standards. And that any deviations from the absolute best should have a decent chance in ending their paid position.

With volunteers we get what we get. But we don’t pay everyone and we should not pay people who are not good member of and ambassadors for the community