[Board] Eclipse meeting report

[Michael P. Gerlek]

Dear Board:

I met with the still-forming Location Working Group (LWG) of the Eclipse Foundation today for a couple hours, the largest agenda item being possible interactions between OSGeo and the LWG. Attendees included OSGeo notables Geoff Zeiss, Chris Holmes, Andrew Ross, and Dave McIlhagga.

My slides can be found here (for now): http://dl.dropbox.com/u/17690616/Eclipse-120522.pdf

The Eclipse LWG page can be found here: http://wiki.eclipse.org/Location

* * *

The Eclipse Foundation is setting up group to develop, support, and promote open source projects in the geospatial arena, with a strong emphasis on being commercial-friendly. Their intended efforts directly intersect with the largest part of our charter here at OSGeo, and so it is of interest to both communities to make this be a positive situation for all.

This new Eclipse LWG would like to be able to use some OSGeo projects, as determined by the needs of their members. They could choose to use an OSGeo project "at arm's length", just like any other consumer of open source code (and as they currently do for some Apache projects) -- but some projects they might wish to bring closer under their foundation, for reasons of stronger/different licensing models, collaboration between projects, etc. Such allied projects would, in turn, be likely to get significantly more funding and strategic support than OSGeo provides for projects do today.

While Eclipse has made it very clear that they have no intention of "poaching" away our projects, this could potentially become a divisive situation in the open source geospatial world.  An OSGeo project steering committee might well wish to work within the Eclipse Foundation if they felt that would be better for their project in the long run -- such as money for codesprints,  more developers, broader industry adoption, and so on. Some projects already feel that being part of the OSGeo community has not brought them sufficient value. We need to address this very real issue.

There are other aspects of mutual interest, such as conferences that we could collaborate on -- but the focus of discussions today was on the projects themselves.

* * *

Eclipse has asked us to work with them, and so it is now OSGeo's job to decide how to react to this situation. There are three possible reactions I see.

First, we could declare Eclipse to be a "competitor" and discourage projects and people from working with them. This would harm and possibly fracture our community, benefitting no one and going against OSGeo's stated mission to "support the collaborative development of open source geospatial software and promote its widespread use".

Second, we could ignore the issue. This too would fracture the community, albeit without as much rancor. In two or three years, however, we might find that the two organizations are competing for scarce resources -- developers, sponsors, conference attendees. That would be unfortunate. And, presumably, one of the two organizations would suffer in the long term.

Third, we could proactively choose to work with Eclipse now. Projects that are of interest to Eclipse and whose members want to be part of the Eclipse stack could be encouraged to do so, with the Board's full support and blessing, because there is no reason a project can't be part of both OSGeo and Eclipse.

I favor this third approach.


* * *

Historically, OSGeo hasn't been able to raise sufficient funding and sponsorship to promote and further projects in the way that many of us would have liked. The reasons for this are many, but it is my considered opinion that as an entirely volunteer-based organization made up of a very diverse set of interests OSGeo is not likely to be able to develop and sustain a successful fund-raising campaign. Indeed, in six years we have made little progress in this area. Others may disagree; my position in this letter is based on this premise, however.

A number of projects have goals they would like to achieve, ranging from attracting more developers to holding codesprints to writing better documentation to producing binary distributions.

Our charter says we must support and promote these open source geospatial projects: but how are we to support and promote those goals of our projects without significant funding or resources?

We now have one possible answer: by encouraging our member projects to work with the Eclipse Foundation. 

As I posted here a few weeks ago, I think OSGeo has successfully achieved many of the goals we set out for ourselves six years ago. But just as the geospatial world has evolved over those six years, so to must OSGeo. Our market and ecosystem has grown, new players are entering, and there's no longer any reason we need to be the only voice for open source geospatial software. Where there are other such voices and where they can accomplish things we cannot, we should work with them on those things, and concentrate instead on the things we can do.

At the risk of sounding trite, we should, in fact, embrace the differences in the two foundations. Although writing code is the most significant effort we do, our charter is broader than that: we encourage data sharing, we promote academic curricula, we educate and inform potential new users. These are things that Eclipse does not directly represent, things that OSGeo needs to continue to stand up for.

* * *

In summary, then, I would like the board to discuss a motion saying that 

  (1) We welcome the Eclipse Foundation as a new, valuable member of the open source geospatial world;

  (2) We recognize the Eclipse Foundation is aimed at commercial-friendly open source software and may be able to support the goals of some OSGeo projects towards that end, in ways better than OSGeo can; and

  (3) We encourage OSGeo projects that are of interest to the Eclipse Foundation to work with Eclipse, to the benefit of the project and its goals, the OSGeo Foundation and its community, and the Eclipse Foundation and its community.

* * *

As I mentioned, there are other areas where Eclipse and OSGeo could work together -- such as supporting the annual FOSS4G conference -- but there's no need to move on all fronts at once. I know some in our community have already expressed fear of a takeover by Eclipse, it's commercial orientation, and a dilution of the OSGeo brand. I believe that moving towards the motion above represents a good, solid first step to build mutual understanding, trust, and confidence between the two Foundations.


_mpg