A distributed version control system doesn't actually mean much for the 
direct management of our project, any more than, say, incorporating 
ReviewBoard into the GeoServer management policy would for GeoServer.  
Basically, what a DVCS provides is a way for contributors (blessed by 
the project with committership, or otherwise) to supply patches in a 
format that breaks up a large changeset into discrete, annotated steps, 
and (at least in git's case) a merge tool that can take advantage of 
that extra info to do merges more automatically.

We still have a central, managed repository: it's the one at 
http://github.com/GeoNode/geonode . This is the main repository, the one 
that you have to get changes into if you want them to be in a release of 
GeoNode.  We can enforce this pretty easily by just requiring that 
releases are tagged and built from that repository.  We can call people 
with write access to this repository 'reviewers' instead of 'committers' 
to emphasize that anyone can fork the project and commit to his own git 
repo, and that those with write access have the duty to review incoming 
patches before pulling in changes from those without it.

That said, I think a little clarity about whose vote counts for what 
would be very good for the project. Perhaps Seb and I should sit down 
and hash out some voting rules before just telling everyone to vote on 
stuff. I definitely agree that we should encourage feedback from anyone 
who wants to offer it though.

--
David Winslow
OpenGeo - http://opengeo.org/

On 07/15/2010 01:49 PM, Chris Holmes wrote:
> Sounds great, but who is 'allowed' to vote?  PSC?  Committers?  Anyone
> who cares to?
>
> I'd say it's good to encourage everyone to vote and express opinions,
> but you probably want to only 'count' some subset, so someone couldn't
> just get a bunch of friends to vote on their issue.  Not that this is
> likely, just an example.
>
> With GeoServer it's PSC votes on GSIPs (big proposals), commit rights is
> 3 other committers.  But I think with distributed versioning
> 'committers' are a bit fuzzier, no?
>
> On 7/15/10 10:51 AM, Sebastian Benthall wrote:
>    
>> I had a conversation with David the other day about a topic that I'm
>> concerned about: how GeoNode is going to transition from a production
>> team driven project to a community driven project.  One thing he noted
>> was that there hasn't been a lot of voting going on.  I don't know if
>> that's because of general unspoken consensus or because it hasn't been
>> how we've done things in the past several months.
>>
>> So I just wanted to say:
>>
>> +1 for more voting.
>>
>> I want to encourage all developers to vote,  even if they don't have
>> much else to say.  I think that it will speed things up and make our
>> community decisions more history and legitimacy.
>>
>> As far as process goes, we all know it informally.  But here's a cool
>> document from the Apache foundation, who apparently started the voting
>> style we all use, on the subject:
>> http://www.apache.org/foundation/voting.html
>> --
>> Sebastian Benthall
>> OpenGeo - http://opengeo.org
>>
>>      

Reply via email to