Andy Bierman wrote:
So, existence of required running code does not mean much.
I disagree.
It means the specification is implementable.
If a protocol is so complex that its implementability is not
obvious, you have lost from the beginning.
Since the goal of our work is to produce specifications
that will allow multiple independent implementations to
inter-operate successfully,
How can you define successful interoperation of implementations?
I think adequate procedures exist for gathering implementation
experience for the IESG to evaluate protocol interoperability.
Such formalism has killed IETF.
To formally confirm that multiple implementations of a protocol
interoperate, which is required these days, you really need to
have a formal specification of a protocol, which, if any, is very
complex even if an informal specification of the protocol is simple.
If all you want is informal and vague feeling of interoperability,
it is not a very useful review.
Masataka Ohta
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf