Clarify language around public / private extension conflicts

[jhoyla]

Previous language could be interpreted to mean that public and private extensions are allowed to conflict.

[tgeoghegan]

I don't understand the purpose of this parenthetical statement. As jhoyla notes, the old version was ambiguous. The new version is more correct, but now is redundant and I think we should delete it.

Suggested change

	type. (For the purposes of this check, a public extension MUST NOT conflict
	with a private extension.) If so, the Aggregator MUST mark the input share as
	type. If so, the Aggregator MUST mark the input share as

@cjpatton do you remember why that sentence in parentheses was ever there?

[cjpatton]

I don't recall discussing this paranthetical in particular.

I think we're attempting to define the following algorithm: let $S$ be the multiset of extension points among the public and private extension fields: if any value occurs twice in $S$, then reject.

I would probably clarify this as follows: "if any extension type occurs twice across the public and private extension fields, then reject".