[OSGeodata] Project request

[Jo Walsh]

dear Frank, Arnulf, all,

On Mon, Apr 10, 2006 at 09:20:55AM -0400, Frank Warmerdam wrote:
> Arnulf Christl wrote:
> > http://www.mapbender.org/index.php/ALK_mit_Freier_Software
> > http://sourceforge.net/projects/edbs2wkt
> >Now - where to put this?
> >
> >There is definitely an Open Source software project (GPL) involved but 
> >it is of type benevolent dictator. Because this project addresses only a 
> >tiny subset of the worlds population (Germans with an affinity to survey 
> >data) this is fine with me but it does not justify to create a separate 
> >OSGeo subdomain. On the other hand it is an exemplary example of 
> >combining Public Geodata with Open Source Software to give people at 
> >large better access to spatial data.
> >
> >It could be branched in Mapbender SVN but that does not make sense, same 
> >is true for MapServer. Maybe this would best fit into the Public Geodata 
> >SVN? Ideas, comments?

Arnulf, this is an interesting hybrid case; I'm glad you raised it. 

Just to summarise; this is a 'sub-project' that cross-sects the
interests of several projects (software and non-software) within
OSGeo, yet as Arnulf points out, it doesn't "belong" to any of them 
to the extent that it belongs in their core code repository.

In this case, a Foundation member has been explicitly asked if this is
the sort of thing we *can* support. It fits with some of the
Foundation goals (provision of resources to projects; encouragement of
standards-based interop; promotion of public access to geodata) but
how well does it sit with other goals ('To ensure a high degree of quality
in foundation projects in order to build and preserve the foundation
"brand".')? 

There have been some general concerns that OSGeo not tend to become a
"sourceforge for geospatial"; that whatever is offered through our
hosted facilities (software, data) conform to a "gold standard". This
process for full projects is still being worked out in Incubation.
And the potential users of EDBS2WKT aren't going to be very widespread
inside the Foundation. 

There's a balance to be found here; on the one hand, I'm very
reluctant to put people off contributing, or to appear unhelpful. On
the other hand, it's worth considering what example would be set by
taking this on; if OSGeo agreed to offer managed facilities for this
project in *whatever* committee's remit, then that would be opening
the door to other efforts just like it. Some of the graduating
criteria for Incubation definitely apply (code legal status,
demonstrable commitment); others for a project so narrow in scope
('democraticness' of maintenance process) don't really apply.

I really don't want to dismiss this idea out of hand. I'd like to
consider where Foundation 'sets the bar' for software contributions by
members, that aren't part of the core projects in Incubation. But I 
reflect that, this might be too much too soon, until the Foundation 
itself has a more stable foundation, until at least a couple of core
projects have successfully incubated. (Mapbender included...) 

I'd be interested in keeping this on a think-about stack, and if more
project suggestions of this kind come along, think harder about what
OSGeo would risk by taking responsibility for them, versus what we
might sacrifice by turning them away. 

Arnulf, what would you do in the meantime; keep hosting at
sourceforge? Look for another organisation that might maintain this in
Germany where it would have more direct relevance? Would you be
disappointed? 

> My suggestion would be that the project could live in one of the existing
> project repositories (perhaps MapBender would be the most logical) as long
> as the project PSC is prepared to be responsible for it.  It would be up
> to that PSC to approve committers, monitor code licensing requirements and
> so forth. 
>
> I don't especially think it belongs in the geodata SVN, partly because that
> would seem to be abusing the geodata project's "free pass" past incubation.

This is the first time that the idea of even using SVN facilities has
come up in the Geodata Committee - it has not seemed relevant so far
to what we are trying to achieve. If you are looking for a
long-term-persistent, stable home for a software project, *especially*
given the uncertainty about where OSGeo's SVN facilities will live after
next year, then Geodata is probably not the go-to-group; The PSCs 
definitely are working a lot more out, in terms of governance processes. 

I still think it's worth us (Geodata) discussing this as a test case
at the next IRC meeting, because it does fit with a couple of the
stated goals of the Committee (# Promote the use of open geospatial
formats # Promote public access to state-collected geodata) In the same 
sense that software is useless without data, data is useless without software.

At the same time, when there are clear code and code governance
standards that need to be upheld for the Foundation to maintain
gravitas, there is no way that we (Geodata) should be taking action
that controverts them. I don't think there is a "free pass" past
incubation; there's a potential opening into "something more", and I
see no harm in exploring, without making resource committments
(because we're not really able to) what that can be.

I remember Markus' phrase not exactly, but it was something like
"we're really working out the process for a non-software project
within OSGeo as we go". There's already a *lot* that a group with such
a diversity of interests in it wants to achieve, and for the 2-3 weeks
that we've been in existence, we've been through a bootstrapping
process that looks like a microcosm of the Foundation's macrocosm -
figuring out what it is that everyone within the group wants to
achieve, how that fits with the projects we're already working on, in
order to later narrow focus on what can be achieved with members' free
code time energy and the potential resources (hosting, visibility) 
that are available. 

best wishes, 


jo