[OpenAFS-devel] Re: "openafs" man page

Russ Allbery rra@stanford.edu
Tue, 23 Feb 2010 13:52:06 -0800


Andrew Deason <adeason@sinenomine.net> writes:

> By not 'touch'ing I meant: does not link against IPL code, is not
> derivative of IPL code, is not copied from IPL code, etc.

I suppose we could have a case like that, but I'd be reluctant, in the
abstract (the specific example may of course lead that way) to accept
something like that because it's a significant limitation on how we
transform that code going forward.  If some of that code ends up fitting
better into a library, we can't move it.  We can't link it with a standard
command argument parsing library.  Etc.

The one exception that comes to mind is a new Linux kernel module or
something else that's entirely kernel space, since that doesn't pose as
many of the same issues and having Linux kernel modules licensed under the
GPL is useful for a bunch of reasons.  But even then, it would be nice to
have it dual-licensed under the IPL for future reuse elsewhere if needed.

-- 
Russ Allbery (rra@stanford.edu)             <http://www.eyrie.org/~eagle/>