[xsl] XSLT Processor caching and keys

["Fraser Goffin" <goffinf@xxxxxxxxxxx>]

We receive XML messages which contain a fairly high number of codified 
values (from schema enumerations). I have beed asked to provide an 
'enrichment' of these messages by looking up each of the code values and 
adding its description to the output. I am considering doing this in XSLT 
(comments welcome) since we do possess (or could create) XML based 'lookup' 
documents.

My question is this though. If it is found that using 'keys' is a helpful 
optimisation, if I keep a copy of the XSLT processor in a cache (the 
messages are requests to a web service), does this include any internal 
structure/indexes created for said keys (I don't know how these work 
internally so I could be completely wrong ?).

Fraser.