Re: Hashed URIs

comments:

there are too many ambiguities, and too many choices.

if this is going to be useful then there needs to be an unambiguous
specification for canonicalizing URIs that is part of, or referenced
by, this spec.  unfortunately this has to vary from one URI type to
another, but at least you could define algorithms for specific types.

it would also help to list a couple of specific use cases for this.
we have too many URI types that have dubious utility as it is.

Keith