[Bullard, Claude L (Len)]
> Obtuse. The real world works as I have described it. The
> fantasy world of URIs as unique ids of unique resources
> doesn't work as long as their is an HTTP in the morph and
> the syntax for naming and resolution are identical.
>
Yes, and I found out the hard way that some "Web Services" tookits will not
accept arbitrary URI schemes - had to stick with http: URLs. I think I
recall that XML Spy did not like non-http: schemes for schema namespaces
either, but i am a bit hazy - put that at about 80% sure. So you cannot in
actual practice try to evade the issue by using "uri:xxx:yyy" instead of
some non-responding http: URI, depending on what tools you work with. Ugh.
Cheers,
Tom P
