[Ltru] Re: Adding more 2119 keywords

[Frank Ellermann <nobody at xyzzy.claranet.de>]

Randy Presuhn wrote:

>| 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

ACK, can't say this often enough.  If a standard says "xyz are
letters", then it's unnecessary to replace it by "xyz NUST be
letters" for some imaginary normative effect.  The standard is
normative without any 2119-keywords, unless it explicitly says
"informative" (examples etc.).

> I suggest considering whether the value of making the change
> is worth the risk of the changes cumulatively necessitating
> a second last call.

So far I thought that an "internal last call" is something like
a dress rehearsal, it's not really required, and it's also not
necessary to repeat it.

Mark wrote in his reply:
| ask John to do his next review before last call ;-)

Better we find any "bugs" now than somebody else later in the
real "last call".  But IMO using _no_ 2119-keywords where they
are _not_ needed is no "bug", quite the contrary.

BTW, in <http://mid.gmane.org/42D272D3.3060802 at zurich.ibm.com>
Brian mentioned the case of a draft standard (3596) obsoleting
BCP 49 (3152), apparently that's possible:

<http://purl.net/net/rfc/3596> + <http://purl.net/net/rfc/3152>

                         Bye, Frank




_______________________________________________
Ltru mailing list
Ltru at lists.ietf.org
https://www1.ietf.org/mailman/listinfo/ltru