At 7:53 PM -0500 3/6/12, Barry Leiba wrote:
> it would be a very broken implementation, but not something
that harmed the Internet or even anyone else whose applications worked
properly. At most it would harm its own user, who I assume would quickly
dump it.
I don't think harm to the Internet is the bar for MUST. If your
implementation would generally be considered broken if you did X, then
I think that rates a MUST NOT X. It's often a bit of judgment whether
to use MUST NOT or SHOULD NOT, and we have some latitude in making
that judgment. I agree with PSA and MNot that this case should be
MUST NOT.
Yeah, harm to the Internet is overblown, sorry. But still, if a
violation doesn't harm anyone else, then I don't think a MUST NOT is
justified. My reading of 2119 is the basis for this:
Imperatives of the type defined in this memo must be used with care
and sparingly. In particular, they MUST only be used where it is
actually required for interoperation or to limit behavior which has
potential for causing harm (e.g., limiting retransmisssions) For
example, they must not be used to try to impose a particular method
on implementors where the method is not required for
interoperability.
--
Randall Gellens
Opinions are personal; facts are suspect; I speak for myself only
-------------- Randomly selected tag: ---------------
He knows nothing, and he thinks he knows everything. That points
clearly to a political career.
-- George Bernard Shaw
_______________________________________________
Ietf mailing list
Ietf@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf