[apps-discuss] CBOR and BULK (was Re: Gen-ART review of draft-bormann-cbor-04)

["Pierre Thierry" <[e-mail]>]

On Thu, Aug 08, 2013 at 09:58:17PM +0200, Carsten Bormann wrote:
> [CBOR] was written to some specific objectives, clearly stated in
> the document.

Hi all,

I just subscribed to this list and just learned the existence of
CBOR. A few days ago, I just submitted a few independant I-D
specifications also for a binary object representation.

It's very interesting because many of our objectives are the same. The
main differences, as I see them, are that I didn't base my data model
on JSON's and that I didn't specifically cared about constrained
environments. I was more aiming at implementation simplicity in
general and it paid off as a side-effect: I think my current design
can be parsed on a class 0 node.

Also, the balance between the requirement on future extensibility and
size of wire data is different, my design being a bit more verbose
than CBOR.

I also envisioned a jump table parser like CBOR's, but in my case, I
always use a whole byte as marker, where CBOR use 3 bits.

> As Paul and I have repeatedly said, CBOR is not trying to be the
> final, definitive, binary object representation for all purposes.

I would expect it to be an established fact that no single binary
format can accomplish both simplicity of parsing, generality,
extensibility and optimal size of wire data and processing speed. I
don't see how a general and extensible format could do as well as
formats like Ethernet, TCP/IP, Ogg or DivX.

But for the rest, from documents, images, or executables to RPC
messages, I suspect the jury should still be out. One format to rule
them all might be possible. Maybe I suffer from a bit too much
enthusiasm, but I hope BULK might be able to do just that (at least
I'm sure it could be the basis of something that does).

I'm implementing a simple BULK reader/writer and I plan to implement a
few utilities before March (a file compressor/encrypter, BULK/XML and
BULK/JSON translators, an archive creator), as I was hoping to attend
the IETF meeting in London.

Here are the I-Ds:

https://datatracker.ietf.org/doc/search/?name=thierry-bulk&activedrafts=on&sort=

I would be glad to hear feedback on BULK from CBOR's authors.

Regards,
Pierre Thierry
-- 
[e-mail]
OpenPGP OxD9D50D8A