On Wed, 16 Dec 1998, Wietse Venema wrote:
> If there is a conflict about a code module, say, the one that
> emulates strerror() on pre-ANSI systems, IBM would be happy to
> provide a non-infringing replacement.
> If there is a conflict about, say, the use of buffered I/O over
> UNIX-domain sockets for IPC between chrooted daemons (comparable
> to someone patenting the exclusive OR cursor), then that pretty
> much nukes the entire Postfix mail system.
> I can understand there is pain when you are only using, say, the
> vstream module because you needed something that stdio could not
> do for you. In that case it would suck if you could not use the
> vstream code just because there was a conflict about the mailbox
> locking routine that you're not even using.
> I will have to ask if partial license revocation is possible.
IBM's commitment to open source in the current time frame is unambiguous.
No one really believes that IBM is going to cut the rug out from under us,
given its current stance and clear interest. But policies do change, and
the more the license commits IBM to look for alternatives to complete
revocation, the more comfortable the user community is going to be about
long-term availability of product.