Re: [Cbor] Review of draft-bormann-cbor-cddl-control-00

[Carsten Bormann <[e-mail]>]

Hi Jim,

thank you for your fast review!

On 2020-06-16, at 23:25, Jim Schaad <[e-mail]> wrote:
> 
> * 2.1 - para 3 - I think you can state this easier as "Target can controller
> MUST be strings.  The result of the operation has the type of the target.
> The concatenation is performed on the bytes in both strings."

Done, thanks.

> * 2.1 - para 3 - I just did a really fast search on CDDL and I am not sure
> that it states anyplace that strings are treated as UTF-8.  I think this is
> probably implicit but saying you are doing the concatenation on the bytes
> has the potential to cause a problem.

Added "If the target is a text string, the result of that concatenation MUST
be valid UTF-8."

> * 2.2 - para 2 - The rule for dealing w/ mixed floating point and integer is
> ambiguous.   This is not a good idea.  It would seem that an equally valid
> interpretation is to use ceiling rather than floor.

Fixed.

> * 3 - nit s/can directly be added/are added/

Done (is).

> * 3 - need to have some rules on the controls, targets and controller in
> terms of types. Can I use .abnf with a binary target and controller?  

I added some text:

> * 3 - I assume that there is an implicit rule that the ABNF must be
> validated and the CDDL is invalid if it is not valid.

Well, we never exactly specified what “valid” means for ABNF.
If you mean that a tool should throw an error upon encountering a controller string that is not “element CRLF *rule”.

> * 4 - I need a much more detailed explanation of what this control is
> supposed to be doing.  

I added one more way to further muddy the waters (but also adding explanation).

> * 4 - I find the following very hard to parse: Extensions that are known at
> the time this definition is known

Fixed (I think).

Thanks again; I have submitted the update as -01.

Grüße, Carsten