GNU Copyright Assignment

[John Keiser]

Etienne has brought up some concerns about GNU copyright assignment being
required before any contributions are made to the project, and I have some
concerns as well (albeit different ones).

Specifically, I think required copyright assignment is hindering the
project.  People who would otherwise contribute are being turned off, not
because they don't *want* to assign copyright, but because it's *hard* and
IMO it seems like unnecessary lawyer bullshit has been forced into what was
once an uninhibited, exciting process.  People have to receive the papers,
fill them out, send them in the mail, wait for GNU to process them, get
confirmation back.  I think the first time I did it, my papers actually were
returned to sender because of the wrong address ... it was months before I
did it again, and frankly I can't remember whether I *did* do it again (I am
willing to if anyone has a problem).  It slows down the wheels, and hackers
who are contributing a patch for the first time are the *last* people we
want to slow down.

There is at least one example of this--someone submitted a patch like 3
weeks ago that I would like to commit, but we have to wait for him to fill
out papers and send them back, which he may never do--because it's hard.
When something waits for a week or more, many people tend to forget about
it.  I know I can't even remember what the patch was for at the moment
(something to do with ClassLoader I think), and with every passing day an
important change stands more and more chance of getting lost.  I don't keep
detailed to-do lists.  I

I understand the rationale for it.  I think it's fine to require it.  Just
don't make it hard.

Here are my suggestions for solving the problem (some are mutually
exclusive):
1. Allow contributions without assignment of copyright.
2. Make the assignment of copyright easier by going electronic.  Have them
go to a page explaining the assignment, maybe even sign their name with a
mouse (in Java), and click submit.
3. As a temporary compromise, perhaps up the # of lines required to call a
patch "substantial."
4. Check with the lawyers and find out whether it's possible to contact
people *after* legal proceedings begin.  This is probably a less happy
solution, but nonetheless a good one.
5. If nothing else, make a "pending patch archive" so that people can make
contributions that aren't accepted yet, and so it stays fresh in our minds.

I know Etienne is worried about assigning copyright at *all*.  This is a
whole separate issue, especially for corporations who are trying to give
something back to the community.  Can we really accept huge patches from
Intel, for example?  I suspect that that's not as big a concern me as an
individual, though, since my work becomes public domain at that point anyway
and if it is misused by the new owner, I can fork.  As long as my name is
associated with my work, I am happy, and I suspect most other hackers are
too.

--John