RSA-OAEP vs. RSA-OAEP-MGF1P

[Cantor, Scott]

> I guess I'd respond by saying that if it advertises those two algorithms -- with
> no mask function parameters -- then if I understand what you're saying it's
> just saying the same thing twice.

Functionally yes, but they're discrete algorithm URIs in the spec and at runtime, so saying you support one definitely doesn't imply the other.

> * If we don't want the SP to advertise the ability to use other digest functions
> (I assume it has that ability, and note that I'm not saying that it's necessary at
> this point, AFAIK MGF1+SHA-1 is fine unless if you just don't like the musty
> smell of SHA-1 anywhere in your system) then there's no need to advertise
> the unparameterised generic form as well as the MGF1P form as they are
> semantically identical.

No, they're really not in terms of the XML. The SP does support consuming all of the digests, and I think it may support the MGF1 variants, but I'm less certain of that. It can't currently encrypt with any of the variants.

> * If we *do* want the SP to advertise the ability to use other digest
> functions, we'd need to advertise each one and take the size hit. People or
> federations can always edit them down again.

It may not be trivial to actually get it doing that, but you can certainly file an issue.

-- Scott