[Top] [All Lists]

Re: draft-hall-deferrals-00.txt

2007-02-01 17:15:39

--On Thursday, 01 February, 2007 14:46 -0800
ned+ietf-smtp(_at_)mrochek(_dot_)com wrote:

PS: we might use an extension with a version, e.g.,
220-DRR v1 v2
MAIL From:<...> DRR=v1
where the version numbers correspond to those listed above,
and make one of them (v1?) mandatory to implement. That
might allow for "future enhancements" with "little" effort.

I really don't think this is necessary.

Let me go a little further.  Unless there is a very clear model
of how extension/enhancements would work, e.g., as to whether
Vn+1 was required to be backward-compatible with Vn, whether a
system that encountered "v99" and didn't recognize it
could/should assume v1 (or v15) and whether it would have to
tell the client it had done that, and so on, our experience is
that this sort of thing is not merely unnecessary, but harmful.
And I don't think there has been nearly enough demonstration of
likely payoff to justify the work it would take to try to define
such an extension mechanism.  

Fewer options yields more interoperable implementations.


<Prev in Thread] Current Thread [Next in Thread>