Re: [Rats] [Cbor] 🔔 WGLC on draft-ietf-cbor-tags-oid-02

[Carsten Bormann <[e-mail]>]

> 
> CWT seems a little ambiguous as to whether thusly encoded OIDs are allowed as a label/key. Section 3 says:
> 
>    To keep CWTs as small as possible, the Claim Keys are represented
>    using integers or text strings.
> 
> 
> This seems more of a statement of design intent rather than a prohibition against other types of Claim Keys, but one could interpret it as a prohibition.

It is neither: It is a statement of fact.
The dual of this statement is:
If the label (map key) is not an integer or a text string, this label is not a Claim Key.
(And, since Claims use Claim Keys, the key/value pair is not a Claim, and ... the whole thing is not a CWT.)

[We had an interesting discussion over at JSONpath about normative statements.
The upshot was that this is a normal way to phrase a normative statement, even without RFC 2119 language.  RFC 2119 language would not be wrong here, but it is not needed.  Just like an ABNF rule “id = 1*ALPHA” would not need an RFC 2119 statement with it that dollar signs MUST NOT occur in ids, unless that is for some reason a misunderstanding that suggests itself.]

Grüße, Carsten