Of course the prefix is technically irrelevant. (I have, indeed, read the
draft, even the latest one.) It would be nice, however, to use something
consistently in the XSchema specification, instead of confusing people with
random prefixes. I'd like to see XSC used in the spec.
I think you'll find in common usage that people will prefer to refer to
namespaces through common references, and that prefixes _do_ matter, whether
or not they do technically. The URI is handy, and necessary, to avoid
conflicts, but is anyone really going to talk about "namespace
urn:uuid:C2F41010-65B3-11d1-A29F-00AA00C14882"? Not unless they're showing off
their memory, in which case they might do better to memorize pi to X digits.
David Megginson wrote:
DM>it would be useful to provide a few examples with alternative
DM>prefixes, just to make this point clear.
This is a good idea, though I'd like to see it confined to the portion of the
spec covering XSchema and namespaces.
DM>You don't need a domain name, just a URL. For example,
DM>if you were to submit the XSchema spec as a note to the
DM>W3C, you could use the URL of the note; alternatively,
DM>perhaps xml.com would provide you with a stable URL that
DM>you can use.
So, any contenders for the URI? members.aol.com/simonstl/xschema/ is quite a
chunk and not likely to be stable, long term.
Simon St.Laurent
Dynamic HTML: A Primer / XML: A Primer / Cookies