David(_dot_)Pawson(_at_)rnib(_dot_)org(_dot_)uk wrote:
Just picking up one point Howard;
You intimate that you need to process instances from n schema's
and hope to do it in a single stylesheet (or small group of stylesheets),
since all these schema's are derived from xbrl.
Correct. For example, suppose you (or Mr. Walsh) has invested in some
stylesheets for DocBook. You want to be able to use them for instance
documents for AcmeCorpDocBook, AcmeCorpHardwareDocBook, and MyDocBook.
If these progressive layers only have a few new elements, that's ok. You
can create new stylesheets to handle the difference. As Walsh says (as I
recall), "If you make extensions to DocBook, it's not DocBook anymore."
Being able to directly reuse DocBook stylesheets on extensions is like
saying that extensions are still DocBook. For What It's Worth, XBRL does
say that extensions are still XBRL. That's the philosophy they're
taking. I'm trying to unravel the consequences of this for XSL.
As a generality, I've always taken the starting point
that there is a one to one mapping, ...
I.e. change the schema and a new stylesheet is required.
I'm now curious if others adopt this as a starting point?
IBM did want to take the directly reusable stylesheet approach in the
documentation domain. (Their reasoning is at
http://www-106.ibm.com/developerworks/xml/library/x-dita1/index.html)
However, I don't know if this was for real use or for a "science project".
For what it's worth, my reasoning in wanting to have directly reusable
stylesheets is to allow early validation of large numbers of new
specialized schemas by reusing the XSL-based validation for the more
general schemas. (Even before anyone get's around to writing validation
suites for the unique part of the specialization.) I'd also like third
parties to write XSL-based tools (e.g., for display) based on general
schemas, and have them still work (although not be specialized) for the
thousands of more specialized schemas.
I imagine this kind of thing would be useful in a fair number of cases,
but would not be commonplace amongst the bulk of XSLT use everywhere.
E.g., I never had to do this kind of thing before, although it would
have helped me slightly (not hugely) if I could have extended DocBook
without so much cut and paste.
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list