I think that's probably a cleaner approach. It also means you can pick up
the required type (from the "as" attribute) easily.

Michael Kay 
http://www.saxonica.com/

> 
> We ended up implementing this in oXygen by looking in the 
> XSLT documents as XML documents and extracting the 
> information directly from there. You just have to be careful 
> that parameters in included files can be rewritten in the 
> including files but otherwise there are no difficulties. And 
> while you do that you can get also whatever comment or 
> documentation element exists for that parameter, if you need 
> that as well. oXygen extracts also that and presents the 
> parameters and their documentation in the parameters dialog 
> when configuring a transformation scenario.