On Sat, 27 Jan 2007, Sven Schreiber quoted:
> Not sure I agree on that. A function may, for example, print
> the user is interested in (e.g. some sort of hypothesis test) and offer
> a return that is only of interest to experts (say, a vector of
> relatively "raw" values that could be further processed).
> Why should the GUI user be forced to assign the return, when the script
> user is free to ignore it?
On the other hand, there may be some packaged functions whose
whole point is to return something, and where it doesn't really
make sense to ignore the return -- perhaps they don't print
anything. I guess that is what you were thinking of. Maybe
(perhaps later) we should provide a way for a function author to
indicate whether the return can sensibly be ignored or not -- some
suitable inflection on the "return" statement?