Hi Wendy,

On Tue, Mar 19, 2013 at 10:01 AM, Wendy Roome <[email protected]>wrote:

> Whoops -- I meant [9.1.2] should say that an ALTO server must provide
> 'numerical' OR 'ordinal' mode, not that a server must support both!
>
> Related question: Do we really need to require that all ALTO servers
> support a Cost Type named 'routingcost'? Suppose I have an ALTO server that
> provides a custom Cost Type to a select group of authorized clients. None
> of those clients need 'routingcost'. Why must my server support a Cost Type
> that no client will ever use?
>

This argument can apply to other components of the spec as well. For
example, the current spec includes that the "pid" Endpoint property must be
supported and one can come with a customized deployment case where pid is
not used.

My understanding of the decision that an ALTO Server must provide
"routingcost" is also related with the design of allowing default. For
example, CostMapCapability can be empty and implies "routingcost". One
analogy one might or may not like is that TLS (RFC5246) defines that
TLS_RSA_WITH_AES_128_CBC_SHA
is mandatory. But we can imagine that there is no mandatory and hence
allows TLS negotiation to fail. But I agree with you that removing the so
far MUST supported cost type and endpoint property will not impact interop
in the sense that the protocol still works, and a conforming ALTO Server
may just not returning any information. What do you and others think?

Thanks for your always good comments!

Richard



> So I'd be happy to drop that requirement altogether.
>
> Aside: My design philosophy is Libertarian, so I tend to challenge "MUST"
> requirements. I think they should exist only if they are absolutely
> necessary for interoperability.
>
> - Wendy Roome
>
> From: "Y. Richard Yang" <[email protected]>
> Date: Mon, March 18, 2013 15:12
> Subject: Re: [alto] minor inconsistency in draft 14
>
> Hi Wendy,
>
> So you think each implementation MUST provide both numerical and ordinal
> modes for the 'routingcost' metric. I believe that this is true for all
> existing implementations, and hence a useful requirement.
>
> If no objection from others, we should add this to ensure interop.
>
> Thanks!
>
> Richard
>
> On Mon, Mar 18, 2013 at 10:38 AM, Wendy Roome 
> <[email protected]>wrote:
>
>> We've worked hard to allow ALTO servers to use ordinal mode instead of
>> numerical mode. However, Section "9.1.2 Cost Maps" (first paragraph, last
>> sentence) says:
>>
>>         This resource MUST be provided for at least the
>>         'routingcost' Cost Type and 'numerical' Cost Mode.
>>
>>
>> I think that should read
>>
>>         ... Cost Type and either 'numerical' or 'ordinal' Cost Mode.
>>
>>         - 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

Reply via email to