Cart
[Home] [By Thread] [By Date] [Recent Entries]
I
don't know how far along anyone is, but a combination of the Schematron approach
and
the Schema approach as has been demonstrated by Rick Jeliffe looked very
powerful. Also, I think there was an announcement
that Clark, Jeliffe, Holman et al
were
initiating ISO work to create a combined specification using RelaxNG.
Please any of the above
correct me if I am wrong.
I
tested the Topologi implementation. Regardless of the marketshare, I was
impressed with
the
capabilities when working with a complex schema. It actually caught errors
that were
not
caught by systems with greater market share and was simple to use. I found
that
combining it with a schema editor (XML Spy) was very
productive. I did not pursue
the
combinations of schematron assertions and schema productions very far, but it
appeared to be a good way to go if co-occurrence
constraints were a concern (and
they
usually are when the original model is a relational db).
len
|
