Open source and 17 U.S.C. 203

Section 203 of the US copyright statute* is an interesting provision which allows an author to terminate the right to use a work 35 years after the grant was made. There are good policy reasons behind this rule; an author may not be in an equal position to bargain with a publisher, and the value of a work may not be apparent when the grant is first made. The rule allows an author to re-negotiate for more compensation if it turns out that the right they sold (or gave away) turned out to be more valuable than initially thought.

Does this introduce any uncertainty in the rights you receive when you use free software? Do programmers and companies now need to worry that they can use free software for at most 35 years before the author has the right to tell them to stop? What about other software which builds on top of the first work - can the rug be pulled out from under them at that point, if the author decides they want to revoke the rights granted to you and instead sell the right to use software?

I believe there is little to no risk imposed by Section 203. First, 203(b)(1) specifically says that derivative works created before the revocation may continue to use the underlying work/code. This would generally protect anyone who has created any material modification of the work, or who has included it in a compilation (for example, a CDROM or other collection). A Freedom Toaster might be an example; any Linux distribution such as Debian GNU/Linux would be another.

Second, section 203(a)(4) (and 37 C.F.R. 201-10) impose strict notice requirements on an author, including, for example, a requirement that the revocation be in writing and that delivery be made by US mail or in person. These might make it expensive for a programmer to contact each grantee to revoke their rights.

The third consideration is interesting; section 203(a) states the requirement that the grant be "executed by the author." Both the GNU Public License (GPL) (section 2) and the Creative Commons licenses state that when you distribute/convey a work to someone else, the recipient is granted a license directly from the author(s); the GPL specifically states that no sub-licensing is involved. This seems to imply that an author could revoke the license to use their piece of free software as long as they comply with the rest of the provisions of section 203 and with the C.F.R. Of course, that would require an author to know exactly when you received the software; the 35 years runs from the time of the grant, which they may not know if you got the code from source other than them. Consider too: what would happen if you receive a revocation from an author and a subsequent copy of the work under the GPL? Would the terms of the GPL mean that the author has agreed to re-license the work to you, as is allowed under sections 203(b)(3)-(4)? That interpretation seems consistent with the language in the statute.

Finally, the "life cycle" of software is fairly short; it's pretty unlikely that a piece of code will remain valuable for more than 35 years. Nevertheless to remove risk entirely, it might be prudent for a recipient to document when they first receive a work under a free software license and to plan to replace the code in 35 years if it's still in use.

For all these reasons, I believe the risks to recipients of free software from provisions in section 203 are minimal, can be easily avoided, and can likely be ignored pretty safely.

*In case you were wondering: section 304 paragraphs (c) and (d) only allow authors to terminate grants prior to 1/1/1978. The first version of the GPL was in 1989, well after that date, so those provisions don't apply here. It's my assumption that 1989 is the first time a free software license was used for works in the United States.