Wendy, Hans,

Thanks for clarifying this. Please see below.

On Thu, Sep 10, 2015 at 12:11 PM, Wendy Roome <[email protected]>
wrote:

> Hans,
>
> Thanks for pointing that out. I never noticed that paragraph before, and I
> do not see any reason for ECS to give costs between prefixes.
>
> I believe that paragraph is what is politely called a "blooper". (There
> are other less polite descriptions.)
>
> My guess is that someone added that in an early draft, thinking it was a
> good idea, but later abandoned that approach. Subsequent reviews
> concentrated on the detailed protocol descriptions and examples, and we
> glossed over the section headers. So no one noticed that dangling thread.
>
> I suggest creating an errata sheet to delete that paragraph, and to fix
> any other, ah, *bloopers* we discover.
>
>         - Wendy Roome
>
> On 09/10/2015, 11:30, "Hans Seidel" <[email protected]> wrote:
>
> >Hi All,
> >
> >Contrary to my knowledge I just found out that the endpoint cost service
> >is not limited to endpoint addresses but also permits prefixes as stated
> >in section 11.5 of the RFC7285.
> >However, I took a more detailed look at this section and recognized that
> >prefixes are no accepted input parameter.
> >
> >Here are some more details:
> >
> >11.5.   Endpoint Cost Service
> >[...]
> >In particular, this service allows lists of endpoint prefixes (and
> >addresses, as a special case) to be ranked (ordered) by an ALTO
> >server.
>

If we want to limit to no prefixes, the change will be a simple revision:
Sec. 11.5:
  "... this service allows lists of endpoint addresses to be ranked
(ordered) ...

On the other hand, if we want to include prefixes in the grammar, we need
to use the definition in Sec. 10.4.4. (Endpoint Prefixes).

I am fine with an errata to do the first change. Although I do see an value
in using prefixes, although the client and the server will need the context
for the semantics: is it the average of between two sets?

Richard


> >
> >    ...
> >
> > From my point of view it is not clear whether prefixes are allowed as
> >ECS endpoints or not.
> >Can someone clarify? Maybe I overlooked something.
> >
> >Thanks
> >Hans
>
>
> _______________________________________________
> alto mailing list
> [email protected]
>
> https://urldefense.proofpoint.com/v2/url?u=https-3A__www.ietf.org_mailman_listinfo_alto&d=AwICAg&c=-dg2m7zWuuDZ0MUcV7Sdqw&r=4G36iiEVb2m_v-0RnP2gx9KZJjYQgfvrOCE3789JGIA&m=fmGqUyxnrMIiZIP7xlORkFFSgwmz0j_EootHahvpbJY&s=hFY7zMi06b8zMk2zsYfC11SnbvFYHChU5nvoTWUDGC0&e=
_______________________________________________
alto mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/alto

Reply via email to