[mdx] MDX & expressing communites-of-interest

[Scott Cantor]

Josh Howlett wrote on 2009-05-20:
> Isn't an aggregate representing the members of a community GET'd from a
> well-known location more likely to match current metadata distribution
> practices? It seems more scalable also, in that entities can be divided
> into smaller distinct aggregates, rather than published within a single
> mega-aggregate.

Strictly speaking, no, neither the IdP nor the SP support the notion of
"groups" based on the fact that two entities live in the same metadata file.
All of the metadata sources you supply are just pulled in and used
collectively, and if you want to define rules based on "community", the only
way to do that currently is by referencing EntitiesDescriptor Names.

In terms of scaling, I think an externalized list of entities bound to a
group name is clearly better because it gets around the problem of
non-hierarchical and overlapping groups.

The sense in which you're correct is that if you wanted to stand up a
service that only supported "InCommon", then yes, you can rely on the fact
that there's a WKL for that.

-- Scott