except that percent doesn't really seem like a routing-specific data type!

(perhaps the "right" thing to do is to deprecate, and eventually obsolete,
the routing one and define it in a core netmod module?)

On Thu, 30 Jul 2020 at 14:59, Benoit Claise <bclaise=
[email protected]> wrote:

> On 30/07/2020 15:25, Juergen Schoenwaelder wrote:
>
> On Thu, Jul 30, 2020 at 02:58:22PM +0200, Benoit Claise wrote:
>
> On 20/07/2020 11:19, Ladislav Lhotka wrote:
>
> Juergen Schoenwaelder <[email protected]> 
> <[email protected]> writes:
>
>
>    - Percentages are frequently used in YANG models but usages differ a
>      lot in precision and range. It is not clear what the proper
>      generic definition of a percentage type would be and whether it is
>      worth having it.
>
>      RFC 7950 example:
>
>           typedef percent { type uint8 { range "0 .. 100"; } }
>
>      RFC 8294:
>
>           typedef percentage { type uint8 { range "0..100"; } }
>
>      I-Ds:
>           typedef percentage { type decimal64 { fraction-digits 5; } }
>           typedef percentile { type decimal64 { fraction-digits 2; } }
>
>      The yang catalogue seems to be down. :-(
>
>    - Proposal: do not add a percentage type since it is trivial to
>      define a context specific percentage type that matches range and
>      precision requirements (and there is already a definition in RFC
>      8294 for those who need exactly that definition).
>
> I agree with this proposal. It is also possible to use
>
>     units percent;
>
> where necessary.
>
> On the other hand, when I look at the numerous percent/percentage
> occurrences in YANG model, it doesn't hurt to define that typedef.
> https://yangcatalog.org/yang-search/ => search on "node name" and typedef
> only
> We can find 56 entries from IETF, IEEE, BBF, OC, MEF, vendors
> Most of them points to:
>
>    *typedef*  percent {
>       *type*  uint8 {
>               *range*  "0 .. 100";
>
>       }
>    }
>
>
> But that one is already defined in RFC 8294 in ietf-routing-types.
> Does it make sense to define it again in yang-types?
>
>
>
> My point was taht it makes sense to group typedefs in a few documents:
> RFC6991, 6991bis (hopefully published soon) and .... my bad,  I forgot that
> RFC 8294 is "Common YANG *data types* for the routing area"
>
> So we're good. Thanks.
>
> Regards, Benoit
>
> /js
>
>
>
> _______________________________________________
> netmod mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/netmod
>
_______________________________________________
netmod mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/netmod

Reply via email to