On Sun, Sep 25, 2011 at 8:10 AM, R A Brown <[log in to unmask]> wrote: [...] > Yep - no one in this thread, as far as I see, has denied the > possibility of some sort of human-usable stack-based > language. After all, Forth programmers and even LISP > programmers are human ;) Taking that into consideration, I note that Forth programmers do make use of the full range of stack conjunctions, and are capable of understanding the semantic effects thereof, but only in a subset of all possible circumstances. Idiomatically, stack operators are only used in ways that are guaranteed to manipulate only the arguments to the current Forth word (if internal to a definition) or logical phrase; you don't get arbitrary rearrangements of arguments of different operations unless someone's being deliberately evil or obtuse. So, you might get away with a useable compromise between full Fith and Shallow Fith by specifying stack operators not as actual stack operators but as components of idiomatic phrases defined like Forth words, or by adding syntactic restrictions such that stack conjunctions are only permitted when they would only operate on arguments of a single predicate. -l.