Hi Tom,
Thanks for taking the time to review the draft.
On 08/11/2023 18:01, tom petch wrote:
<tp>
I think that the Abstract is misleading. I see this as a data model defining
the structures to be used to specify a BGP community which happens to use YANG
as a DDL but could use all sorts of languages. It might (should?) have
appeared as an Information Model first. This becomes clearer from the
Introduction but that Abstract needs to lead into that.
Thanks. The abstract could indeed use some more words to make this clearer.
And the Introduction needs to come before the Terminology.
Will fix this in 01. Thanks!
Technically, I find it flawed because it makes extensive use of strings with no
constraints. In several, cases, RFC have defined data types that could be used
for such fields - IMHO strings are not a suitable alternative.
Where there is no data type, then string allows a length of
18446744073709551615. I have argued for a restriction thereon, as we had in
SMI, but YANG Doctors always resist. The type t also allows the full range of
pictogram languages which I find a hindrance; I note that recent I-D have
imposed restrictions thereon, typically A-Z 0-9 and a selection of
punctuation varying according to the need.
Fair point.
I'm not convinced a custom datatype is the right solution everywhere
(for example 5.3 and 5.6.2), but in those cases at least a restriction
on length and allowed characters would indeed make sense.
Reference clauses I find helpful - in some places they are required in a YANG
module.
Ok!
Kind regards,
Martin
_______________________________________________
GROW mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/grow
