Re: [nmrg] commands on sming snmp mapping document

[Frank Strauß <[e-mail]>]

Juergen Schoenwaelder wrote:

> Here are some comments on <draft-irtf-nmrg-sming-snmp-02.txt>. I will
> not list the editorial issues here which I will forward to Frank on
> paper but rather focus on the technical issues.

Ok. I've done another round of edits that address the editorial minor
issues you gave me in paper form and a number of issues in your email
upon which this is a reply. Below are the remaining issues that I did
not yet work on.

> a) There is no section on revising SNMP mapping definitions. Not sure
>    why we overlooked that until now.

> b) [...] we [probably should] change TimeTicks to TimeTicks32,
>    [then] we [would] have to update this document as well.

> d) The description of the node statement should probably say something
>    regarding the Pointer type.

> f) The fact that the scalars' implements statement assigns
>    sub-identifier starting at 1 causes problems with existing scalar
>    groups that have all sort of numbering gaps. I guess it is needed
>    to be able to override it.
[Of course, this also applies to the table's implements statement.]

> k) Perhaps the NMRG-SMING-SNMP-EXT module should be renamed to
>    NMRG-SMING-EXT-SNMP so that all SMIng language extension are
>    grouped together and to reduce confusion by module authors (at
>    least I got sometimes confused).



Slight comments on two of the other issues:

 > e) The description of zeroDotZero should probably be:
 >
 >    "A value used as null value for identifiers and pointers."

I changed it simply to "A null value used for pointers." Does
that make sense?

> l) The References section must be split. I suggest to have normative
>    references [SMIng] and [RFC2119] and to make all other references
>    informative.

Ok. I also moved the ABNF RFC to the normative references, since
implementors have to be able to read the SMIng grammar.