Re: Re:... replacement for 14.4 and C6.2 (interchange encodings)
On 2014-06-25 15:31:34 +0100, John Pryce wrote:
> On 2014 Jun 25, at 02:31, Dmitry Nadezhin wrote:
> > 1) Where will be a place to define mapping to octets ?
> > A) It is level 5.
> > Level 3 is a pair or tripple of Level 2 floating-point datum and an abstract decoration;
> > Level 4 is a bit string;
> > Level 5 is a sequence of octets.
> > B) It is level 4.
> > Level 3 is a pair or tripple of Level 2 floating-point datum and an abstract decoration;
> > Level 4 is a sequence of octets.
>
> I don't know enough to have a strong opinion, but I tend towards
> (A), if A.4 means the conceptual bit string we had before which is
> independent of machine architecture.
I tend towards (A) too, but IMHO, it should be levels 4a and 4b, since
if I understand correctly you don't have a mapping from a bit string
to a sequence of octets, just 3 -> 4a and 3 -> 4b mappings.
> > 2) What is a name for a sequence of octets ?
> > Level 3 is called representation;
> > Level A.4 is called encoding;
> > Whare is the name of A.5 or B.4 ?
> > I think that "mapping" is too general word.
>
> Is "octet-encoding" OK?
Then, bit-encoding and octet-encoding respectively.
> > 3) Is A.5/B.4 required, recommended or annex ?
>
> Good question. I would say required. Why spend all this time
> stressing out at the last minute, if the result of our effort is
> something implementers can ignore?
IMHO, the same requirement level as A.4.
> > 4) Is signature an ascii string or an octet sequence or a
> > structure of C language ?
>
> It needs to be language-independent. IMO, it should be convenient
> for humans to read. That indicates ASCII string.
Yes, it needs to be language-independent. Even a structure in C
wouldn't properly define the signature as there may be padding bytes
in some implementations (and the size itself may have padding bits).
--
Vincent Lefèvre <vincent@xxxxxxxxxx> - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)