On Thu, Jul 18, 2013 at 10:51 AM, Wendy Roome <[email protected]>wrote:
> First, if a server provides multiple IRDs, are cost-type names local to
> each IRD? Or are they global to the set of IRDs? The examples define
> cost-type names in both IRDs, which implies the names are local. But I
> didn't see anything that said that.
>
Yes, they are local to an IRD.
>
> Second, {10.1.2.4} has the following sentence under the description of
> cost-type-names:
>
> Since an unfiltered Cost Map is requested
> via an HTTP GET that accepts no input parameters,
> an ALTO Client MUST be led towards a resource
> that has a single element in the ¹cost-type-names¹ list.
>
>
> I think that would be better expressed as
>
> Since an unfiltered Cost Map is requested
> via an HTTP GET that accepts no input parameters,
> the 'cost-type-names' capability for that resource
> can only contain one element.
>
This change sounds good to me.
>
> That section also says:
>
>
> If there is more than one Cost Type in this list,
> then the ALTO Server SHOULD return an IRD to the client
> to lead it towards the URIs for the corresponding Cost Maps.
>
> I don't understand what that means. Can anyone explain it?
>
This means that the ALTO Server may respond with an Multiple Choices (300)
status code with the body containing an IRD. If I recall correctly, the
explicit statement about the HTTP 300 status code was removed after a
discussion about there being too strong of a coupling between ALTO and the
HTTP layers. I know the WG has gone back and forth over appropriate
wording for this particular issue in the past.
>
> - Wendy Roome
>
>
> _______________________________________________
> alto mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/alto
>
_______________________________________________
alto mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/alto
