Re: Interoperability for MUST, SHOULD, MAY

[Thomas Narten <narten@us.ibm.com>]

> I think we had consensus that we need to have two genetically 
> implementations on every feature - and that this is orthogonal to the use 
> of 2119 language (as in "these are not the same thing").

This is the key point.

> I think we also admitted that we've let some of the features only
> used by MAYs escape from being noted in the interoperability
> reports' feature lists.

I don't know that we've done that. I think people asked if MAY is not
required. Some of us have not uses the 2119 language as the main/only
basis for coming up with the features list.

> I don't know if we had consensus to be very much more strict in the future 
> when writing feature lists.

My recollection was that there was no desire to raise the bar in
general over what we have been doing in the past. 

Thomas