Re: Levenshtein distance in XSLT 2.0

Turns out saxon is better at optimising tail recursion in templates than in functions. If I make the obvious changes to mine to turn the functions back to named templates I never get the nested function call error (er, because there are no function calls:-) and can get up to length 50 or so, after that I don’t get an error but it slows down rather more than I’m prepared to wait, but your version (which isn’t passing the whole n*m array )is quicker anyway so might get further in template style.

I seem to remember Dimitre raising a similar comment on the saxon list a while ago?

David

Reply

The content of this field is kept private and will not be shown publicly.