[OSGeo-Board] Contributor agreements

[Rich Steele]

Dave McIlhagga wrote:
 
> I have to say that I'm concerned by the idea that only some OSGEO
> projects would be covered by CLAs and sets a precedent I'm not too
> comfortable with for the future.
> 
> 1. How significant is the 'anti-CLA' sentiment within OSGEO? Is this a
> vocal few, or does this represent the will of project contributors
from
> many of the open source projects currently entering the Foundation?
> Would not enforcing a CLA for all contributors prevent any projects
from
> entering?

Dave, I don't know the answer to any of those questions.  I do know that
the original objections to the CLA were not related to the points raised
by ESR.  Rather, they were related to the CLA's granting of a license to
the foundation that was perceived as being too broad (such that the
foundation could relicense the code in the future if the project so
desired, without contacting every single contributor).  I believe these
objections were limited to certain GRASS contributors, which is
perfectly understandable given that GRASS is a GPL project.  Thus, at
the IncCom meeting, a reasonable suggestion was made that the
relicensing option be removed from the CLA.  This suggestion didn't
pass, so then a motion was made to simply eliminate the CLA requirement
altogether (throwing out both the baby and the bathwater, in my
opinion).  I think some people were swayed by ESR's arguments, and I
think others thought that if relicensing was not an option, then the
utility of the CLA was marginal and not worth the trouble.  But as I
pointed out in my last email, I strongly believe the CLA requirement is
important for other reasons.

> 2. Image and Promotion - My hope has been that OSGEO could present a
> strong united message regarding IP. A message that organizations can
> adopt any OSGEO project knowing that it has been through an identified
> level of rigour for auditing purposes and commitments by contributors.
> Commitments that ensure the PSC and the Foundation have the neccessary
> tools to act in the best interest of the Project. Two-tiers of
projects
> will inevitably reduce this overall credibility, including the
> credibility of projects that enforce a CLA. A 'lowest common
> demonitator' effect is difficult to avoid.

At some level, I think it will be difficult for OSGeo to present a
united message regarding IP at anything other than a fairly high level.
As we discussed in Chicago, OSGeo is different from other foundation in
that we are not born of a single project nor do we have a single
licensing philosophy.  But license philosophy notwithstanding, I do
agree that OSGeo projects can (and should) share a commitment to a code
diligence process, and I share your concerns about two tiers of
projects.

By the way, there is an interesting debate going on at Apache right now
about the Apache "brand" and what it means in terms of expectation when
you download Apache code.  This is discussed in the context of
licensing, but I think applies equally well to Apache's IP review
policies.

http://mail-archives.apache.org/mod_mbox/www-legal-discuss/200603.mbox/%
3c2f8a5bd60603161651y5760ee56g970790e981ce38eb at mail.gmail.com%3e

> 5. Barriers to Contribution - Contrary to the fear that CLAs may
inhibit
> contributions - some organizations will not contribute resources and
> efforts to a project UNLESS CLAs are in effect for all contributors.
> This barrier in my mind could be dramatically more significant then
the
> loss of a few contributors unwilling to sign a CLA. 

Contributing to a project does not raise the same concerns as using code
from a project.  However, a corporate contributor probably isn't too
interested in contributing to projects it doesn't also use, so I think
your point is still valid.  
 
> One further thought:
> Is the primary concern about adoption of the CLA revolve around
> granting the right to re-license to the Foundation -- in particular
> for GPL projects?
> 
> If this is the main issue - has there been any consideration of
> allowing each project to select from one of two CLAs, one that grants
> the ability to re-license and one that does not?

That seems like a reasonable compromise.  Some projects may choose to
retain the relicensing option, and others could choose not to.

Frank Warmerdam wrote:

> While there is little doubt in my mind that we need to offer a CLA
> without
> the relicensing option if we are going to have a CLA, for me the most
> compelling argument against the CLA is the amount of overhead it
> introduces
> for a (to my mind) questionable legal benefit.

Some question the legal benefit and see only overhead, while some
question the overhead and see only legal benefit :)

But seriously, there are plenty of open source organizations out there
that are dealing with this issue every day and still cranking out great
software used by millions of people around the world.  I think before we
throw up our hands and say this is too much hassle, we should talk to
people in these other organizations and find out what their experience
has been.  I spoke briefly with Cliff Schmidt, Apache's VP Legal
Affairs, at a conference today.  I can ask him (or Brian B) what
Apache's experience has been on this, and also try to reach out to some
other organizations.

-Rich