[Beowulf] All Your BASH Are Belong To Us

Lux, Jim (337C) james.p.lux at jpl.nasa.gov
Thu Aug 11 11:19:00 PDT 2011


> -----Original Message-----
> From: beowulf-bounces at beowulf.org [mailto:beowulf-bounces at beowulf.org] On Behalf Of Joe Landman
> Sent: Thursday, August 11, 2011 10:54 AM
> To: beowulf at beowulf.org
> Subject: Re: [Beowulf] All Your BASH Are Belong To Us
> On 08/11/2011 01:40 PM, Lux, Jim (337C) wrote:
> > It's really depends on a corporate/organizational commitment to open
> > source to institute processes to keep all this stuff straight.  (and
> > we won't even get into "open source" vs "able to redistribute")
> There are profoundly incorrect views running around out there, as to
> what "open source" means.  I had someone tell me that GPLv2 prevented
> distribution of binaries (it doesn't).  I've watched people slap
> additional legal bits in conflict with GPL onto GPL source.
> I don't want to say "its a mess" but I do want to say that "there is a
> profound need for a very simple statement of what is and isn't allowed
> by each license."  Including what is involved in altering licensing.

Closer to home for me, the NASA Open Source License (which was conjured up a decade or so ago) is apparently incompatible with just about everyone else's licenses.  They had a "How do we encourage Open Source use at NASA" symposium a few months back hosted at Ames with lots of remote participants and licensing issues and complexities is, in my opinion, probably one of the bigger problems.  It's been a royal pain for me trying to release stuff to the general public in a useful form.  It sure would be nice to be able to give someone an .iso and say, here, load this, run make clean; make all, and you'll have your stuff ready to run.  But no, that .iso will be a derived work comprised of a multitude of components with all sorts of different license agreements.  What we have to do is the (to me) accursed approach of: here's a list of eleventy-seven URLs and FTP sites, go get these files, check their MD5 to make sure they're the same one we used, and have at it.

The complication is that in general, work funded by NASA and performed by government employees is a "government work not subject to copyright" although work funded by NASA and performed by an educational institution (e.g. JPL, which is part of Cal Tech) is subject to Bayh-Dole, and is presumed to be owned by the educational institution, with a fully paid, non-exclusive license granted to the government for government purposes.  (there is, of course, litigation about what those "government purposes" might happen to be).

The incompatibility arises because NASA is legally obligated to distribute their products with no downstream restrictions on use, which is not the same as, for instance, GPL, which imposes restrictions on downstream use.   NASA (and the government in general) doesn't care if someone takes their product and uses it to make a subsequent closed source product which is totally proprietary. (and in fact, NASTRAN would be a fine example of this)

More information about the Beowulf mailing list