Re: [Cbor] Updated Drafts for dCBOR I-D and Gordian Envelope Structured Data Format I-D & IANA Tag Registration

[Christian Amsüss <[e-mail]>]

On Wed, May 31, 2023 at 12:35:03PM +0200, Carsten Bormann wrote:
> What I’m trying to point out here is that mixing this issue with
> deterministic encoding might be blurring it more than helping.

To put a point from earlier in the interim into mail-referencable form:

Maybe the numeric part can be better expressed as a variant of the CBOR
information model. Applications would describe that they're not built on
the (strongly typed) CBOR information model, but described in terms of
the more losely typed numeric-types-are-interchangable information
model. That model might have different APIs to interact with; for
example, a weakly typed language may have a `.append_int()` and an
`.append_float()` function in the encoder of CBOR, whereas the
numeric-types-are-intercangable (dCBOR) encoder has a
`.append_numeric()` function -- most of its API will be similar, though.

That model may also describe that numeric arrays are always expanded,
or encoded as numeric arrays under some particular conditions (but I
guess it'll be always-expanded), or for even weaker typed languages also
define a high-level "maybe UTF8" type that gets encoded as tstr or bstr
depending on its content.

The model would then describe how to do the number reduction, and CBOR
deterministric encoding (as described as sufficient in section 3.1 of
mcnally-deterministic-cbor-01).

BR
Christian