Thank you for your answers...
>>>I consider normalizing to be hoisting
all
definitions/declarations to the top level and replacing recursive structures with references to the hoisted ones. <<< I like to search all elementes with references (ref, type, etc.) in the
schema an replace them with the referenced structure (so make them
anonymous). I like to have a schema without references, so that the
application which uses the normalized schema have not to jump around but
has already a "flattened" schema to work with in a more easier
way. I have similar intentions as posted here (documentation etc.),
but I want to generate also another output with it.
We have considered the recursive-problem, and our program already
recognizes and deals with recursive structures.
Another problem is how to normalize the other more complicated
things like list, union, simpleTypes with extension, restriction and so on. Is
there something specified?
>>>Take a look at http://sourceforge.net/projects/xsdcomp/
and see if it is along the lines of what you were trying to do (especially the
message forums). It has been inactive for over a year and stuck
on a working draft syntax. Basically, the idea was to use XSLT to resolve all the references in a schema and produce a "compiled" version of the schema that could be used to produce documentation, class definitions, etc.<<< Yes, on the first view it comes very near to what I mean! How did you fix
the "compilation"-structure? Your own ideas or have you refered to something
official?
*greets* Thilo
|