Re: license change for POSIX manpages
On Tue, Jun 08, 2004 at 09:07:52PM -0400, Nathanael Nerode wrote:
> Andre Lehovich wrote:
> > The upstream source for the manpages has received permission
> > from IEEE to include text from the POSIX documentation in
> > Linux manual pages. Debian has not distributed the POSIX
> > man pages because until recently the license prohibited
> > modification.
> > The latest version (1.67, 20 May 2004) now allows
> > modification, "so long as any conflicts with the standard
> > are clearly marked as such in the text". Joey Schulze,
> > Debian's manpages maintainer, thinks the need for clear
> > marking may be a problem.
> Yeah, it's not actually free. :-(
> If it said something like "So long as no conflicts with the standard are
> represented, explicitly or implicitly, as conforming to the standard", it
> could be free.
> As it is, it seems to quite effectively prohibit (for instance) adapting a
> printf manpage for use as a manpage for weirdf, a non-POSIX command. I see
> no sane way to make sure that "conflicts with the standard are clearly
> marked as such in the text" for such reuse. Correct me if I'm wrong.
Since there's no weirdf command in POSIX, there's no conflict with the
standard. A conflict would be "POSIX says foo does A, but you say foo does
B", not "POSIX says foo does A, but you say bar does A+B".
The only way I can think of to create conflict in that instance would be to
say "POSIX says the command to do A is foo, but you say the command to do A
is bar", but I think that might be taking things a little too far. And in
any case, conflicts aren't forbidden, you merely have to flag them as such
-- and a note saying "The POSIX-approved method of doing A is foo" should
suffice in that instance.
I don't feel that forcing someone to add what amounts to a symlink to your
original version in their derived work is necessarily free, though. I'm up
in the air about the matter -- it feels freeish, but at the same time
something tells me there's an alligator somewhere. I think asuffield's
clarified version, if we could get the IEEE to accept it, would be a nice
solution to the problem.