RE: Namespaces, schemas, Simon's filters.

[Evan Lenz <elenz@xyzfind.com>]

Ron Bourret wrote:
> What confuses me about all of this is that, if you don't tie a name in a
> namespace to a single semantic concept -- which people seem loathe to do
> -- what is the point of having namespaces? That is, if a URI + local
> name can identify two different concepts, why did we add the URI?

I think of namespaces more in terms of ownership. I can own two elements
with the same expanded name and process them differently based on their
context. Another person is also free to use "local element types" in their
namespace. Likewise, another part of my application can use a different
namespace. Granted, there is a pretty fine distinction between "another part
of my application" and "another part of my document." This concept of
"ownership" is a flimsy one, but it has served me well.

The real lesson to be learned is that Namespaces can be used in the way you
describe and Namespaces can be used in other ways, such as the way I
describe. Technically, since we have well-formedness, we should embrace the
flexibility of the specification, not for the sake of mere flexibility but
because of the variety of existing practice.

Evan Lenz
XYZFind Corp.