[Advisors] Updates and Next Steps
David.Nalley at citrix.com
Tue Dec 18 22:46:31 UTC 2012
On Dec 18, 2012, at 3:38 PM, Jim Jagielski wrote:
> Hi David! ;)
> Just some inline comments... please note, these are my comments and
> not necessarily those Gluster or JMW.
> On Dec 18, 2012, at 1:48 PM, David Nalley <David.Nalley at citrix.com> wrote:
>> My apologies for slow feedback. Even this is incomplete, but I needed to get something out rather than let it dwindle as a draft.
>> So a couple of side comments - Apache and Eclipse are very different open source orgs (as is oVirt). I understand some of the attraction to the Eclipse model, but the mixing and matching causes some concern. Each of those foundations has been successful in its own right, can you tell us why not copy one or the other?
> We are trying to copy oVirt as much as possible, but "fill in the
> holes" (as it were) by leveraging Apache mechanisms... This is
> because after we created oVirt, we realized that there were
> some actions and procedures that were either not well
> defined (or defined at all!) or not really suited to what we
> wanted to implement. That's the reason for the mix you see.
Interesting. I'd love to hear about what you learned from the oVirt experience sometime.
>> So this is somewhat confusing to me. So plenty of assumptions here on my part - it appears that we are setting the stage for the Gluster Software Foundation - there will be multiple projects within the 'GSF'. What isn't clear is how membership is handled across those groups, there is talk about different standards being set within the group once a project is large enough. So I guess a point of clarification - are you a member of the 'Gluster Project' - or are you a member of Gluster sub-projects? or??
> Some of this is murky on purpose: we want to empower the Gluster
> *community* to answer these kinds of questions. The feeling is
> that if we define everything, then there's little for the
> nascent community to do, which will drastically impact uptake.
OK - so defining the minimum.
>> Unwritten things:
>> So there is plenty of talk about the responsibilities, and making the board more active. But there doesn't seem to be much about the authority of the board. Where does authority end and begin?
> Ideally, the board exists to enforce a neutral, well-run project... It
> does not, and should not, make technical decisions, or anything else
> related to the development or "business" aspects of the project
> and codebase. It's there to ensure collaboration and consensus-
> based operations.
Completely agree with this sentiment.
And completely happy for the definition to be that essentially the Board exists to safeguard the assets (trademarks, etc) of the foundation and ensure that it continues delivering on its mission, but doesn't meddle in the technical/business end of the software projects themselves. (actually - that is exactly the way I think it should be)
> Also, at the start, the board is there to be a wellspring of
> experience, knowledge and insight. Think of at least some members
> of the board as "midwives", helping the birth of the community.
>> There are no officers defined - not necessarily a problem, but…..
>> So not to call out the 800lb Gorilla in the room - but there is zero mention in the governance document or this email of the relationship to Red Hat.
>> Is the project completely free from RHT strings? Is it truly independent? Is there to be a legal entity holding various pieces of IP and providing some framework?
>> Assuming the answer to this is 'no'. What is the relationship to Red Hat, what are the limits on project (and board) authority?
> This is key: we (we == JMW, and myself, and Red Hat and everyone else)
> want Gluster to be run as a *real* open source project. Not some
> front for Red Hat, not as some kinda pseudo-FOSS project where
> RH has ultimate control (ala Oracle and Java), but a real FOSS project.
> RH will provide support and resources to the effort, of course,
> but not use or abuse any sort of "control" over it, and the
> board will be empowered to make sure that's the case (and not
> only w/ RH of course).
So not to indicate that its still confusing to me….. but it is.
And for the record, I think corporate-sponsored/controlled projects can be done well. (I'll happily cite Fedora - which I think has generally shown incredibly good stewardship and generally precious little corporate meddling) They can also be done abominably as in some of the examples that you cite, and they can also be done with the best of intentions, but with results that make one go 'meh'. But I do think that it is important to know what the standing is. Fedora, for instance, makes that perfectly clear in the person of the FPL who has absolute veto power over anything in the project as being the 'control' that is able to be exerted by RHT - though it is incredibly rarely used, because I think there was recognition that giving up 'control' results in greater investment by others. That's a fine solution for that particular problem.
Anyway, sorry for taking that segue - let me ask a question that is more to the point. What/who is the legal entity that is assuming the risk behind the operations of the project? This, in my experience, determines where the real boundaries are, because that underlying organization has obligations, typically legal or regulatory, that necessarily impact the project, albeit in hopefully small ways. And if it is the primary sponsor, there also becomes a point at which it becomes no longer in the business interest of that sponsor to continue operations.
More information about the Advisors