Re: Attribute uniqueness test: a radical proposal

> With namespaces and URIs, the simple option is to just quote the URI
spec.

Tim, that's the simple option IF AND ONLY IF we grant the new assumption
that the namespace's identity is a URI.

The as-published Namespaces design -- that the namespace's identity is a
string, with URI syntax suggested for convenience of management --  is
simpler. That seems to be precisely what offends folks about it.

>To then add baggae such as "a URI *except* don't use this form or that
form"

The Forbid option can equally well be described as removing baggage. "An
absolute URI plus a locator. Not a URI Reference, hence no relative
syntax."


I think the past month's discussion clearly shows that it _IS_ taking more
than 90% of the effort to close this last corner. How many man-hours did
the Namespace REC take? How many have been spent debating it since then?
______________________________________
Joe Kesselman  / IBM Research

Received on Wednesday, 14 June 2000 16:59:55 UTC