John C Klensin wrote:
Or we could drop it. My instinct at the moment (i.e., I reserve the
right to change my mind) is that dropping it would be better.
Having a specification contain language about hypothetical future capabilities
is almost never useful and often turns out to be just plain wrong. The only
benefit of reference to the future is when the specification contains pedagogy
about extensibility by way of showing how to add capabilities.
A spec should specify. And it should explain the specification. Predicting the
future does neither.