Lists Home |
Date Index |
The fault is not, in this case, with the "Namespaces" recommendation,
but with the abstraction levels of the parsers and document models. If
all they do is model text, each application needs to work its way up to
names on its own.
Paul Prescod wrote:
> People keep responding to me as if I am implicitly approving of Tim
> B-L's opinion. I don't! Namespaces in general, local namespaces in
> particular and QNames in attribute values in really, really, particular
> have done nothing but give me headaches since day 1. The namespaces
> spec. is short, concise and precise but it leaves so much to the
> application that it is not applied consistently in practice.