Re: copyrights assignments

[jg pa dec com (Jim Gettys)]

I am VERY leery of requiring copyright be assigned...

The reason why is my experience with X:

A bit of history: MIT (and the X consortium) did not require copyright
assignment.  They did require a contributor to sign a letter stating that
the contributor had legal ownership of the copyright, and that redistribution
was allowed by the X Consortium under those terms (copyrights no more
restrictive than the original MIT copyright).   This is needed to cover
things legally (and there should probably be patent language in such letters
today: at the time, software patents weren't much of a problem).

When the X consortium was disbanded, the MIT/X consortium copyright
went with it to (eventually), TOG (the Open Group).  With 20-20 hindsight,
it would have been better to have assigned the copyrights back to the
members of the X consortium, rather than assigning them to the successor
organization.  MIT and the X Consortium were certainly trustworthy
organizations: but that does not mean that the sucessor organization was.

As you all probably remember, TOG, in one of its spasms trying to avoid 
going broke, changed all the MIT/X consortium copyrights to a restrictive 
copyright. Legally, the officers of an organization have fiduciary 
responsibility: one of the arguments made was that their budget forced 
them to do it. It took a year or two, great anguish on all sides, and 
a threatened fork in the code base by XFree86 to get TOG in line (and 
X.org got founded to try to get things on a better footing).  Thank
God we had NOT required contributors to X to assign copyright.

This community actually owes the big UNIX vendors a large amount of thanks 
for fixing this (all behind the scenes: one of my major complaints with 
X.org is that they are completely opaque to outsiders, myself included). 
One of the saving graces that allowed this mess to get fixed was in fact 
that most of the code in the X distribution was copyright others than 
T.O.G. (e.g. Digital, HP, Sun, IBM, etc.). This allowed the UNIX vendors
to cry foul with a very straight face.

There is a moral to this history: while the people who run things now may
be very trustworthy, they may move on to other things: in setting up an
organization like this you need to take a long view.

I believe that (widely) distributed ownership of copyrights is MUCH safer 
against change and problems than ANY central organization, and that 
provisions should be made up front that for any copyrights assigned end 
up getting assigned to a bunch of people/organizations simultaneously, 
to avoid the kind of mischief that happened with T.O.G.  That we trust 
Gnome folks today, or FSF (I don't, and many others don't), doesn't mean 
they or their assigns will be trustworthy in the future.

I certainly will not assign any personal copyrights to a Gnome foundation, 
and I suspect many/most large companies will not either (as it would then 
prevent them from using the same code in other circumstances).  I think 
the best solution is the one adopted by MIT and the X Consortium, with 
more carefully drawn donation letters, and provisions for where copyright 
ends up for those the foundation might hold itself.

Remember, (s)he who owns the copyright can cause IMMENSE mischief (if 
they get a cricital mass of the copyrights, that is). Once burned, twice 
shy. 			- Jim

--
Jim Gettys
Technology and Corporate Development
Compaq Computer Corporation
jg@pa.dec.com