Re: comments on PR-xml-infoset-20010810

> But RFC 2606 recommends the use of .example for documentations.

I take the statement in RFC 2606 as meaning "of the four domains
test, example, invalid, and localhost, example is the one for use
in documentation".  I don't see anything deprecating the use of the
existing example.* name, and the real point of 2606 seems to be to
provide for the case where a top-level domain is needed (which it
isn't in this case).  In the absence of anything prohibiting the
use of example.org, I prefer it on aesthetic grounds.

About your other points:

I will change the RFC reference to point to IETF.

The phrase "absolute URI with an optional fragment identifier" does
not imply that a URI can contain a fragment identifier, any more than
"ham with eggs" implies that ham can contain eggs.  But there's
another point which needs to be resolved - should namespaces have to
be URI references?  If so, they can't contain non-ascii characters
which is inconsistent with (for example) system identifiers which are
defined to be escaped if necessary by the processor.  Should they
perhaps be IURIs (or some similar term)?

-- Richard