On Tue, Oct 18, 2011 at 8:10 PM, Cantor, Scott <[email protected]> wrote:
> On 10/18/11 8:17 PM, "Sam Hartman" <[email protected]> wrote:
>>So, I'd prefer the context be fixed. I think we need it to be fixed for
>>GSS-EAP. If the use cases for ECP are going to be different then we
>>should consider that. However it would be valuable if you could plug one
>>in case of the other.
>
> I guess the way I look at it is that as a consumer of the GSS naming
> extensions, I'd be physically ill shipping code that hardcoded any names,
> so I see it all as a config time question.
>
> But then I'm also accused of designing systems that are too complex.

We should want to not hard-code mechanism OIDs, for example, but we
have little choice but to hard-code things whose semantics must be
embedded in the application.  For example: name types.

We've added mechanism attributes as a way to avoid hard-coding
mechanism OIDs: hard-code the semantics you want rather than the
mechanisms you know at that time support those semantics.  We could do
that with mechanisms, but name-types seem to basic a unit for further
decomposition.  Ditto, I think, naming attributes, though perhaps not
because they are so basic as because there's too much left to run-time
interpretation.  Fortunately, I think we'll be able to make some types
of authorization operations configurable so that naming attributes
need not be hard-coded but configurable (e.g., group-membership-like
attributes).

That said, I don't yet understand what Sam means by "context" here.

Nico
--
_______________________________________________
abfab mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/abfab

Reply via email to