On Tue, Apr 22, 2014 at 06:24:10PM +0200, Thierry Carrez wrote:
The origin of this requirement is the definition of 'ATC' (active technical contributor). Pre-foundation it was simply equivalent to code contributor. You contribute, you are an active technical contributor, and therefore you're allowed to vote in PTL and PPB/TC elections.
Unfortunately, the Foundation bylaws state (in Appendix 4) that ATCs must be individual members of the Foundation. There are two ways to read that -- all contributors must be individual members, or "ATCs" are the subset of contributors that happen to also be individual members.
I read it the second way, FWIW. I also believe that requiring all contributors (even a one-time contributor of a 'drive-by patch') to be Individual Members would have been seen as a significant aspect of Foundation membership policy at the time the Foundation was formed, yet I can recall no discussion on the issue. I am not saying that it is something that ought to be stated in the OpenStack Foundation bylaws necessarily, but I am saying that when the bylaws were initially drafted, if it was really contemplated that all contributors would be required to become Individual Members as a *prerequisite* to making an initial contribution (however trivial), it would probably have been made explicit in the bylaws much like the CLA requirement is stated in the IP policy. In other words I do not believe a policy of "you must join the Foundation if you want to submit a patch" was contemplated when the Foundation was formed. If anyone else here thinks I'm wrong about that, or has a different recollection about this issue, I'd be happy to hear it. Reinforcing that point, if it is correct to read the bylaws as saying that all contributors must join the Foundation, why wouldn't the CLAs be unified with the membership agreements? I have to emphasize how unusual I believe this policy is. I have been trying to find some example of an open source project-related membership foundation (there aren't too many of these) with a similar policy, with no success. I think Apache requires project leads to become members by its notion of membership; that's the closest analogue I've been able to find. It just strikes me intuitively as *wrong* -- isn't it in effect coercing potential new contributors into joining an organization they might not necessarily wish to join, or might not wish to join until later on? If the effect of the policy is that it creates a larger body of Individual Members than otherwise might exist, this makes it more difficult to amend certain provisions of the bylaws, which may be good or bad or of unclear goodness/badness, but is a significant side-effect. Also, it is not beyond the realm of possibility that the Foundation might someday charge a fee for Individual memberships. If that ever happened, and the policy remained in place, I believe it would look really bad.
Since it was quite difficult to map contributors to individual members and ensure that only the subset of contributors that are individual members are considered ATCs, it was simpler to just consider the original sense of "ATC" (active contributor) and consider that the bylaws state the all contributors must be individual members of the Foundation.
Not saying it wouldn't make sense to fix that, just explaining where it comes from.
Understood. It seems that it is well intended as an effort to comply with the TC policy, and maybe I'm in the minority in thinking it is a bad policy, but ... I think it is a bad policy. - Richard