I want to transform an XML document into HTML. Some XML elements have links to others documents like:
In the H
XSLT is a functional languages which means for a given input it will always product the same output, so by definition a guid method or any other random generator would not be part of the design spec. Your best bet if you're client bound is to use a time-related method as part of a pseudo-random seed for generate-id, however as your goal appears to be strong decaching you should abandon this and just focus on applying the correct anti-cache headers to the resources you're trying to protect.