[dnsext] Re: [Editorial Errata Reported] RFC4035 (7972)

["Rose, Scott W. (Fed)" <[e-mail]>]

On 5 Jun 2024, at 23:02, Rob Austein wrote:

>> Type: Editorial
>> Reported by: Nate Choe <[e-mail]>
>>
>> Section: 4.2
>>
>> Original Text
>> -------------
>>    Security-aware resolvers MAY query for missing security RRs in an
>>    attempt to perform validation; implementations that choose to do so
>>    must be aware that the answers received may not be sufficient to
>>    validate the original response.  For example, a zone update may have
>>    changed (or deleted) the desired information between the original and
>>    follow-up queries.
>>
>> Corrected Text
>> --------------
>>    Security-aware resolvers MAY query for missing security RRs in an
>>    attempt to perform validation; implementations that choose to do so
>>    MUST be aware that the answers received may not be sufficient to
>>    validate the original response.  For example, a zone update may have
>>    changed (or deleted) the desired information between the original and
>>    follow-up queries.
>>
>> Notes
>> -----
>> "MUST" is a key word according to RFC 2119/BCP 14 and should be capitalized.
>
> Well, it's been nearly twenty years, and I don't feel terribly
> strongly about this, but since Nate raised the point:
>
> I think I disagree.  "MUST" is keyword, "must" is not.  It probably
> would have been advisable to avoid the non-keyword "must" due to the
> risk of this confusion, but "MUST be aware" does not seem actionable
> to me, nor does it seem likely to be what the author of the original
> text (who may or may not have been me) intended.
>
> So I would reject this erratum, but as noted above, I don't feel
> strongly about this, and will not stand in the way if the (long
> disbanded) WG or my co-editors (most of whose contact information has
> probably changed by now) agree with Nate.

Don’t know what it says about me, but my contact information hasn’t changed…

I agree - it’s hard to code “aware” into a resolver as that was advice to implementors not a hard requirement. We should not have used “must” there, just to avoid using RFC2119 keywords even in lowercase supporting text.  Probably should have written it as “...implementations that choose to do so would need to account for answers that are not sufficient to validate the original response” or something.

Scott

=====================================
Scott Rose
NIST/CTL/WND
[e-mail]
ph: 301-975-8439
GoogleVoice: 571-249-3671
=====================================