Re: [Cbor] Considerations for converting from other serializations formats like Protocol Buffers and FlatBuffers

Very helpful, thank you Cartsen! I just posted to the T2TRG.

On computation - I've also noticed formats that weren't originally designed
cross the open internet lack a framework for signing and encryption, like
JOSE & COSE. I don't immediately see challenges when going between X and
CBOR, especially if X doesn't define S&E. Anything that comes to mind?

On Fri, Apr 10, 2020 at 8:59 AM Carsten Bormann <cabo@tzi.org> wrote:

> Hi Jonathan,
>
> Such a conversion is always “interesting" because of the different generic
> data models of the different data formats.
>
> CDDL essentially is used to define a subset of the generic data model of
> CBOR.  It also can “parse” or “construct", so it could be used as a
> front-end/back-end to such a conversion format, but we would need a
> Format-X back-end/front-end as well.  And some of this will be idiomatic
> (e.g., very idiomatic in JSON), so you don’t want to need to say the
> idiomatic things all the time.  Some conversions will even need
> computations (say, a JSON enum into a CBOR bit string).
>
> We have done something related to a conversion process (uncompressed vs.
> compressed) in RFC 4997.  This notation was also saddled with needing to do
> its own binary decoding/encoding (which still would be useful as another
> front-end/back-end of the above!), but the idea of having statements that
> connect on two sides to two different worlds is immediately applicable.
> SCHC <
> https://tools.ietf.org/html/draft-ietf-lpwan-ipv6-static-context-hc-24>
> is a contemporary attempt to do a bidirectional conversion, but it is
> somewhat specialized to “packet”-type data.
>
> Note also that T2TRG has done a little work on “YOUPI” <
> https://tools.ietf.org/html/draft-petrov-t2trg-youpi-01>, which is
> another way to look at binary data.
>
> As you can see, I think some of this is research, and it borders on the
> WISHI/One-data-model work that we are doing in T2TRG.  I would really like
> it if you could bring in your ideas there.
>
> Grüße, Carsten
>
>
>
>
> > On 2020-04-10, at 17:34, Jonathan Beri <jmberi@gmail.com> wrote:
> >
> > Hello,
> >
> > I'm looking into a use case that would require converting data from
> existing systems that output in well-defined serialization formats
> (specifically Protocol Buffers and FlatBuffers) to CBOR. The goal is in
> part to use other specifications like CoAP with this data as as well an
> existing implementation of CoAP.
> >
> > Has there been any investigation into converting to these formats? Any
> lessons or points to consider? I've been reading up on CDDL - might that
> help in implementing a conversion process?
> >
> > Many thanks.
> >
> > --
> > Jonathan Beri
> > linkedin.com/in/jonathanberi
> > _______________________________________________
> > CBOR mailing list
> > CBOR@ietf.org
> > https://www.ietf.org/mailman/listinfo/cbor
>
>

-- 
Jonathan Beri
linkedin.com/in/jonathanberi <https://www.linkedin.com/in/jonathanberi/>