I certainly recall instances where features were dropped from the Draft
Standard version of a specification precisely because interoperability had not
As can I on quite a few occasions, but what I cannot provide is any evidence -
or even an anecdote - that in the present climate of high scrutiny before
getting to proposed, such changes demonstrably improved interoperability and
deployability in the field.
Now, it is entirely possible and even likely that this practice carries
benefits that far outweigh the costs. But we should not assume that's the case
jut because unused features offend our design sense.
Ietf mailing list