[Cbor] Re: Private tag numbers / 1010

[Anders Rundgren <[e-mail]>]

On 2024-12-15 08:20, Carsten Bormann wrote:
> On 15. Dec 2024, at 08:06, Anders Rundgren <[e-mail]> wrote:
>>
>> OK, then I would settle for bstr (like COSE's "kid") because that gives you highest flexibility.
> 
> COSE’s kid [0] records a unilateral decision to call something (a key) something (an opaque byte string).
> It is essentially used as a hint if this information is reflected back to the system that gave it a name.

This what the RFC says but that is just a theory.  Nobody (in their right mind...) is "trying out" signature or encryption keys.
That is, kid (in reality) depends on mutually decided conventions.


> (kids are also instance identifiers, not type identifiers.)

No argument about that.

> 
> Type identifiers are different:
> 
> * first of all because there are intended as types,
> * but also because they usually are meant to transfer processable information from one player to another one,
> * in particular if they are intended to have a unique meaning and are dereferenceable [1].
> 
> bstr won’t do.

The only reason I didn't use bstr in COTX was for printouts in diagnostic notation :)

Anders

> 
> Grüße, Carsten
> 
> [0]: https://www.rfc-editor.org/rfc/rfc9052#section-3.1-3.8
> [1]: https://www.ietf.org/archive/id/draft-bormann-t2trg-deref-id-04.html
>