Re: [abfab] Direction Forward for aaa-saml

[Leif Johansson <[e-mail]>]

On 2015-07-22 18:26, Sam Hartman wrote:
>>>>>> "Leif" == Leif Johansson <[e-mail]> writes:
> 
>     Leif> On 2015-07-22 18:20, Sam Hartman wrote:
>     >>>>>>> "Leif" == Leif Johansson <[e-mail]> writes:
> 
>     Leif> (I am and was speaking entirely wo any chair hats on btw)
> 
>     >> 
>     >> 
>     Leif> Right but we _could_ add Endpoint and leave the work of
>     Leif> specifying the URL format of radius radsec servers to whomever
>     Leif> wanted to deploy it
>     >> 
>     >> I'm very against that.  It's not guaranteed to be interoperable
>     >> without the URI and I don't think we'd have confidence in the
>     >> semantics without going through the URI spec.
> 
>     Leif> Thats why we have the Binding parameter! If you don't
>     Leif> understand the Binding then you can't use the Endpoint.
> 
> No, my point is that until the URI is specified, it seems unlikely that
> two implementations would both work with this endpoint.

Sure but that is no worse than any SAML deployment. Until you agree on
the binding you can't talk.

> I absolutely agree that it wouldn't break other bindings.
> But for example if one implementation wanted radsec://... and one wanted
> radius+tls://... then they wouldn't both be able to consume the same
> metadata.
> 

If they invented the same binding then yeah life would suck for them.

Again: if somebody wants to use Location then have them define the
binding and URL format. Done.

	Cheers Leif