RE: Dump XPATHs

Bruce wrote:

> what's the actual purpose for an "XPATH dump"? 
> Maybe if we knew the 'requirements' better, 
> we could propose a better solution.

i've been trying (on and off) to build such an animal
for some time as an attempt to report on schema test 
case coverage:

1) generate a set of possible canonical XPaths from 
   a schema (this problem, AIUI)

2) generate canonical XPaths from a set of XML documents

3) sort | comm results to find out which cases from #1
   are covered by documents in #2, and which cases are
   missing. 

This approach obviously doesn't cover combinations, but
is good enough for building a taxonomy of schema features
exhibited in test case instance documents.

The best i've come up with so far for step#1 is to 
experiment with the Sun and XMLSpy 'document instance 
generators' which have controls over stuff like depth of
recursion.

Unfortunately i've yet to find one that meets my needs or
is Open Source (hackable), so became bogged down developing
my own..

Paul