Lists Home |
Date Index |
On May 6, 2005, at 12:29 PM, Uche Ogbuji wrote:
> What standard? Many parsers do not use SAX. There is no standard
> processing model. Users do seek a standard processing model. That's
> the problem.
Right, and it is logically a separate problem than defining a standard
Different applications have different processing requirements, so
having the syntax specification define any one model would be wrong.
For example, editing applications could well require a different
API/InfoSet than XML databases. IMHO. This was one of the most
significant issues in the early DOM specification process: the tension
between different application domain requirements.
> I understand your basic point. If there were a standard processing
> model, that would be one thing. The problem is that there isn't, and
> it's well demonstrated that this is a source of confusion.
I don't in any way disagree that there is a problem, or that there
should be a solution. I'm just saying it'd be better to do it outside
the scope of the XML specification itself.
There is no canonical processing model... and there probably cannot
ultimately be a single canonical processing model, but there are coarse
classes of processing styles (pull, event-driven, tree-traversal,
iterator-based, etc.), some of which logically subsume the other. Some
tighter specification around those would help solve the problem IMHO.