To support, sustain and protect this vector of values, you decide to create a legal entity.
- If the concrete thing is about making money together, you create a company;
- If the concrete thing is about just your group making money separately, you create a trade association;
- If the concrete thing is about enabling anyone to benefit, you create a charity.
That last one is what open source community members tend to label a “foundation”. And obviously I’m simplifying here, a lot.
So how many of those do we need, seriously?
No Quick Fix
You’ll note that each of these treating-groups-of-people-as-if-they-were-collectively-a-person - “incorporations” - encapsulates existing motivation, trust and treats the result as if it were independent of the individuals who originally came together. It’s important to realise that it does not bestow the vector of values.
If there’s no working community of trust, motivation and resource, creating a foundation will not magically cause it to come into existence. There is no point trying to create or join a foundation to solve absent community values. If you have problems, solve them before you incorporate, as incorporating will just make your problems permanent instead of curing them.
No Force Fit
Similarly, it’s also possible that attempting to join an existing foundation as a short-cut won’t work either. To succeed, the existing foundation’s well-established way of working will need to be compatible with the already-functioning vector of values of the group joining.
There’s thus no One Model To Rule Them All. The world is too diverse. No matter how effective a given structure may be for existing groups, there are in my experience always factors that differ. If those unique factors can’t be eliminated, the only answer will be a new incorporation. Given the bureaucracy involved in starting and sustaining a charity it's worth avoiding it you can, but it's better than force-fitting your community into the wrong structure.
So the answer is, we need as many foundations as there are sufficiently unique communities for them to encapsulate. Maybe we need some patterns for people to follow as they incorporate, maybe there will be plenty who fit an existing incorporation like Apache or Eclipse or Outercurve, but ultimately it’s about the project, not about the incorporation that encapsulates it.