Lists Home |
Date Index |
Michael Kay wrote:
>>While this approach "works", in practice wouldn't the documentation
>>value of the schema (including documentation that can be
>>the schema) suffer?
>Do you have any better way of handling the fact that elements are sometimes
>optional and sometimes mandatory depending on the circumstances? I would
>have thought that writing it in this way is as good a way of documenting
>this dependency as any.
I'm inclined to think that the approach described by Bob DuCharme later
in this thread, and shown by example by Jeni Tennison, would produce
better documentation and be more manageable.
Conditional compilation has a way of getting out of hand, even with
simple preprocessors. While your example of a schema snippet written as
an inverted transform is simple enough, an XSLT programmer could too
easily produce a "schema" that would confound analysis.