[GRASS5] Proposal: RFC 1: Project Steering Committee Guidelines

Frank Warmerdam warmerdam at pobox.com
Fri Apr 28 13:41:13 EDT 2006

Hamish wrote:
> Yes, a new raster format will need a well thought out and published
> plan. I am not all that familiar with the formulation of the new vector
> library- can someone provide some history? I expect this is something
> that is much harder to do well by a committee of many slightly
> interested parties than it is to be done by a small team of full
> knowledge coders. 


Note that the proposal / lazy consensus model is aimed at letting a
knowledgable person (willing to do the work) propose the document
rather than editing-by-committee which I abhor.  And then folks with
knowledge and an interest can put in their oar.

> A system where 50 people have voting rights and it only takes 2* +1
> votes to pass something clearly has problems. Two people could pass a
> strange idea before anyone else notices and then the result is ignored..

Note they can only pass something if no one objects.  The vote as proposed
must stay open for 2 days.

> Ok, then it could be buried but then what was the point of the 2* +1
> rule? Leave votes open for 5 days? 

Well, for small things the longer the gap the more slowdown in the process.

 > Do Glynn's votes on raster matters
> and Radim's votes on vector matters count more than mine? (honestly they
> probably should..) Do I get veto rights on stuff which I am the primary
> or sole author of?

According to the document anyone on the PSC gets a veto if they can
express a reason clearly.  Ultimately though the veto can be overridden.

Radim and Glynn possess emmense moral authority.  Are you going to
veto a proposal of theirs in their area of expertise?  I doubt it unless
you have a very compelling reason.   Obviously the danger of casting the
"PSC member" net too wide is you might include some folks with no sense
of cooperation and respect for areas of expertise.

> I have no idea how to kick someone off of a committee / cvs access /
> etc. in the free software context. Hopefully never, but in the event
> would you have a public or private vote? (private means only exec
> committee sees who voted what) 

Certainly it is necessary to be able to do this very occationally.
And when it has to happen it helps to have a clear and open process
to make the decision.  I think making such a decision in private
would be a worst case scenario.   I have seen folks stripped of
CVS commit privs on projects for disagreeing with project admins on
things that ought to happen.  That is very ugly.  But if someone
is clearly misusing their CVS commit privs., not following projects
procedures and so forth, then there needs to be a way of removing
their commit access and having them submit changes through the
bug system.

> Personally I can see the need for a PSC and am willing to help, but to
> be honest would rather spend my finite time and energy fixing bugs than
> playing politics. We have meetings at work. They are boring and don't
> solve much. Some people love scoring points and winning arguments, but
> this doesn't help solve any problems. What the meetings do provide is
> forced communication, and this is very very important. However the
> nature of the GRASS project, being coordinated over a central mailing
> list(s), means that communication is usually very good here. The main
> communication failures can be avoided if all CVS committers understand
> that they must post to grass5 detailing any far-reaching changes days
> before doing so (and some understanding by others is needed that it is
> often hard to know how big your small change will be when you make it).
> Public IRC meetings need to be logged and archived.

Note, that as far as I know MapServer never has formal meetings in IRC
or elsewhere.   This isn't manditory.  But some projects, such as GeoServer
make a habit of weekly meetings.  I don't see a reason for GRASS to start
having meetings without a particular need.

> I guess I am having problems with what the PSC could do for us as I find
> it hard to think of a decision which we have had to make in the past
> that would have been better solved using a PSC vs grass5. Can anyone
> give examples of decisions that have ended up being weak or never had a
> good answer?

How about, joining OSGeo?   Whether to have a PSC?

Ok, those are rather self referential. :-)

> e.g. I don't know if a PSC could help solve the problem of 3 different
> teams working on 3 different GUIs, all with too light a focus to get
> done. We can't stop people but I think it is a shame whenever a lot of
> work goes into something that never gets widely used due to bad
> integration with the central effort.
> some obvious examples that a PSC could help with:
> Would a committee decide which widget set to use? (java/gtk+/kde/..)
>   (2/3 majority, but majority doesn't work well with >2 answer
> questions. auto-transferable voting is great but who'll set it up?)

I think a PSC could help decide which approach to GUI SDKs and so
forth would be considered officially supported by the GRASS project.
Of course it doesn't stop others from pursuing their own approach as
an addon or layer.

> Would a committee decide if a submitted module should become part of
> the main CVS? (2* "+1")

Sure, why not?

> OSGeo stuff (2/3 majority?)
> I must admit I'd feel a bit silly voting using the term "I +1 agree".
> The exec committee +5 each? Markus have the option to use atomic +20
> if he ever feels the need to use it?  (assuming 50 voters) +inf+1?
> Very hard to ever get the balance right with people coming and going;
> 1st class, 2nd class, 3rd class voters is a bit unfriendly; we should
> have the balls to make decisions ourselves instead of hiding behind some
> complex algorithm that forces an answer.

Note +1 is just intended to be a sign of clear support for an idea, and
not a mathematical construct.  It comes out of use on a variety of projects
over many years (IETF, Apache, etc).

The idea of an executive committee is quite reasonable though it isn't
exactly clear to me what decision would need to be made by an executive
committee rather than a larger group.   To me, the more "executive" the
decision, the more important it is to have broad support in a community
oriented project.  But perhaps I am unclear on the sort of thing an
executive committee might decide.

Best regards,
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | President OSGF, http://osgeo.org

More information about the grass-dev mailing list