[Cbor] Octets (was: Re: Paywalled IEEE standards referenced in draft-bormann-cbor-time-tag)

[Carsten Bormann <[e-mail]>]

Hi Emile,

I’ll get to your very useful input later, as it requires some study, but I can react to this editorial one now:

> On 3. Apr 2021, at 01:49, Emile Cormier <[e-mail]> wrote:
> 
> Also, "octet" would be a better word than "byte", since there are architectures (such as DSPs) where a byte is 16 bits.

Actually, no.  
The term “octet” stopped to be useful approximately 1983, when the last architectures that had non-8-bit bytes ran out of steam.

Its continued usage over the decades is a symptom of a stilted style that values the superficial appearance of precision over the usefulness of the resulting documents.
It may be a small instance of the many symptoms of this kind in standards documents, but it is worth stamping out with a vengeance.

(Note that this is about the English language — octet of course is the right word in French and other languages.)

If a technical document today uses the term “byte” for something else than 8 bits, somebody is trying to appear very smart by regressing to the time when that actually still was an option.  Technical documentation should never be made misleading just to appear that kind of smart.  I actually haven’t run into misuse of the term “byte” for a long, long time — can you point to an example from this century?

Oh, and to get rid of the residual argument that “byte” might be perceived to be less accurate: since RFC 7049, many documents’ terminology sections include the sentence:

   The term "byte" is used in its now-customary sense as a synonym for
   "octet".

Just checked the time-tag document: 
It does so explicitly as well (many other documents just include the terminology of RFC 7049/8949 and so do that only implicitly).

But now back to your excellent technical input...

Grüße, Carsten