--On Wednesday, 31 January, 2007 17:02 +0000 "Steven M.
Bellovin" <smb(_at_)cs(_dot_)columbia(_dot_)edu> wrote:
On Wed, 31 Jan 2007 11:54:26 -0500
John C Klensin <john-ietf(_at_)jck(_dot_)com> wrote:
Except for the fact that the material being cited contains the
specifics of license and IPR releases, and promises to abide
by certain rules, by the authors. Authors can't reasonably be
asked to agree to something that might be published under the
BCP number in the indefinite future, so you are either stuck
with a document (RFC) number or a BCP as of a specific date,
which amounts to the same thing and is harder to track down.
I'll let Jorge correct me if I'm wrong, but referencing by
<number,date> is the norm in the legal world, since statutes
do get amended without necessarily being renumbered.
I believe that is correct. The problem here is that, as a
consequence of that referencing system, they have taken measures
to be sure that the version corresponding to the reference is
readily available. We don't do that: finding out what was
actually BCP NNN on some particular date requires both skill and
out of band knowledge.
I do agree we want to make it easy for non-lawyers. I've
suggested a date-stamped archive of each version of each such
document, for precisely that reason.
That, of course, would do that job and eliminate all of my
objections. But it does mean that, for many purposes, we can't
use a reference to "BCP nnnn", it must be a reference to "BCP
nnnn as of yyyymmdd" or its equivalent.
Ietf mailing list