NASA requests help finding gov't use of standard OSS licenses.

Brian Behlendorf brian at behlendorf.com
Mon May 2 07:51:15 UTC 2011


On Sun, 1 May 2011, Robin 'Roblimo' Miller wrote:
> And then there's VistA - http://en.wikipedia.org/wiki/VistA -  which is 
> public domain and was released as the result of a FOIA request. I love VistA 
> and wrote about it earlier today at htttp://roblimo.com -

This is also on the spreadsheet Scott is maintaining that Karl mentioned. 
However, it's listed incorrectly - the U.S. government did not release it 
as GPL or under any other open source license, but as public domain, as a 
response to a FOIA request (and subsequent updates under subsequent FOIAs, 
a situation they are trying to fix).  This is because VistA was written by 
VA employees and/or contractors whose IP output was assigned to the VA.

This is significant, as some will argue (as I might) that taxpayer-funded 
bespoke software should be released as CC0/public domain, which should 
also be seen as covering any patent claims by the developers.  When the 
developers are government employees, this is already the case (or should 
be).

When the software is written by a contractor, then the terms of that 
contract & any supplements apply.  That's generally whatever the 
contractor can negotiate, and what the procurement office will ask for. 
Procurement offices don't seem to be particularly careful about IP issues 
beyond ensuring the government's own rights to use the software.  So, 
all too often, the contract will pay 100% of the development costs of 
software that is then 100% owned by the contractor, who can then license 
it to others under the terms of their choosing, including non-open-source 
terms.  This is a wasted opportunity.

With NHIN CONNECT (http://connectopensource.org/), no software was 
authored by government employees, but by government contractors, who were 
asked to release the source under the BSD license (which they did).  So 
that should be on the list.  BSD isn't perfect, Apache would be better as 
it provides guarantees against patent claims by contributors over their 
contributions, though some might argue that patent grants are implicit in 
the BSD license.

IMHO, if the NASA software is written by government employees, it should 
be released CC0.  If it's a derivate work of a pre-existing codebase, it 
should be offered back upstream, offering to sign a contributor agreement 
when appropriate, so that the upstream source can then license it as per 
the rest of the upstream codebase.

If the software is written under contract, NASA procurement rules should 
stipulate the same terms from contractors.  After all, contractors argue 
that they provide better value per hour than government employees do - to 
properly compare apples to apples, they should provide the same work 
product.  If they hinder the value of what they create - either by putting 
non-open-source terms on their code, or even using a copyleft license - 
they should do this at a significantly discounted hourly rate to 
compensate for the more limited rights conferred. After all, the fewer 
restrictions there are on code (and BSD has the fewest aside from CC0), 
the more re-usable it is, the more miscable it is, the more fungible it 
is, and the more valuable it is.

Finally, after spending a lot of time on this, I've come to the conclusion 
that government should not be in the business of aggregating copyrights of 
others into collective works - that's much more efficiently and clearly 
handled by non-profit organizations, or secondarily by individuals or 
corporations.  There is a great set of reasons why governments can be 
contributors to and even instigators of open source projects, and we 
should encourage that.  But NASA and other federal agencies, and other 
governments I've seen, have a very different (and well-intentioned) set of 
policies they have to follow around procurement, public communication, and 
community governance that can significantly complicate their dealings with 
others.  They can not reach what I consider the ideal state, which is a 
neutral core a la the ASF or Linux Foundation or Drupal Association, 
around which revolves an ecosystem of individuals and institutions that 
contribute to and benefit from the code.  When HHS fumbled the CONNECT 
contract renewal, they set back the development of the nascent CONNECT 
community by a year or more.  If it were just a function of timing it 
would merely have been defensible as bad luck, but I believe it's inherent 
in the model.

By that reasoning, the best thing for NASA to do with its open source 
projects is not determine which open source licenses to release them
under, but to either contribute them to existing organizations interested 
in being a home for them, or seek a contractor to set up a governance 
structure and release it themselves (as the VA is doing with VistA).

Brian




More information about the License-discuss mailing list