Michael Torrie wrote: > And I posit that most efficient interpreters don't use switch/case at > all, but rather jump tables. Not always true. If the number of cases is small enough, a tree of branches could be faster due to cacheing and pipelining issues. -- Greg