On 2025-03-04, at 16:27, Anders Rundgren <[email protected]> wrote: > > Since there is no interest in discussing these topics in a more constructive > way, I might as well finish with https://www.json.org/fatfree.html. This is > (IMO) how you obtain ubiquity (=universal).
It is interesting to note that JSON is not deterministically encoded. It is definitely way easier to deterministically encode CBOR than JSON, so there may be a larger percentage of applications that want to enjoy its benefits (or, should I say, are prepared to incur its onus?). > It is OK (indirectly) claiming that Universal CBOR is c**p and its author is > a charlatan. However, Universal CBOR is [currently] the only show in town. I don’t know which town. When you say “universal”, you seem to assume every town. But what we have seen is that the set of applications (or of processing steps in applications) that need deterministic representation is limited, and certainly not “universal”. > Hashing "raw" (non-wrapped [*]) CBOR is not a mystery and enveloped > signatures are featured in PDFs. The fun part in embedded (enveloped) signatures is the transformation for obtaining the signing input, and how it should also be the transformation for obtaining the signed input. That may seem simple until it isn’t. CBOR makes it really easy and inexpensive to do an enveloping signature (such as COSE) by byte-string-wrapping the to-be-signed input, so the incentive to do something else is quite low. That is not to say those applications don’t exist (and I’m happy CBOR supports them very well); they just aren’t “universal”. Grüße, Carsten _______________________________________________ COSE mailing list -- [email protected] To unsubscribe send an email to [email protected]
