Lists Home |
Date Index |
"Bullard, Claude L (Len)" wrote:
> No, I'm not rejecting them in their entirety. They
> are too useful. I am saying they don't belong in the
> core given definitions that leave their 'resolvability'
> an unpredictable property.
The resolvability issue strikes me as entirely artificial and an
unfortunate artifact of having used URLs instead of URNs for namespaces.
We don't talk about the "resolvability" of element type names or PI
targets or notation names. Why should we talk about the "resolvability"
of namespace names? These are *names* not *locations*, and its
unfortunate that they look like locations.
(That said, the human factor does count. If we'd left it to machines,
this whole mess never would have evolved :(