On May 28, 2004, at 1:18 PM, David Brodbeck wrote:
What we need to watch out for are attempts to "embrance and extend" the
protocol so that Microsoft mail servers can interoperate but other
servers
can't. This is exactly what Microsoft did to Kerberos, for example.
If
we leave Microsoft an opening to do that, they *will*.
This is the big worry for me. Even if everyone at MS working on this
now has the best of intentions, if five years down the road MS can
exploit their intellectual property rights in way that hampers the use
of son-of-SPF among their competitors, they will. Indeed, it would be
irresponsible (to their shareholders) not to.
It doesn't matter whether it is MS (with their particular track-record
wrt to standards) or, say, Eric S. Raymond. The standard (and
extensions) must be fully implementable by any party without
restriction, and we must know that that will be true for the entire
life of any patents.
It would be irresponsible of us to rely on anyone's good intentions.
We need a structure that protects us. I have not studied the IETF
patent policy, but it may, indeed, provide that protection.
This question, as others have said, is a separate question than whether
something like XML is needed. As recently pointed out, various
plausible extensions do require context-free parsing of the data. Once
we are convinced that there may legitimately be a need for such an
extension, then something like XML is a reasonable (but not only) way
to go.
-j
--
Jeffrey Goldberg
Jeffrey(_at_)goldmark(_dot_)org http://www.goldmark.org/jeff/