Re: URIEquivalence-15 and IRIs

Roy T. Fielding wrote:
> I think that, in general, it is far more effective to require that
> a namespace identifier be in a restricted, canonical form, than it
> is to require all parsers to perform URI-equivalence operations
> when comparing two namespaces.  Much more efficient as well.

This makes all sorts of sense to me.  We could take the literalist 
interpretation of the namespace rec (char-for-char means bit-for-bit) 
and then issue an architectural finding that in URIs

- %-encoding MUST NOT be used unless it's required
- %-encoding MUST use lower-case a-f

Does this work?  -Tim