On Jun 10, 2009, at 3:00 AM, bodrato at mail.dm.unipi.it wrote: > An addsub/2 can be useful here, currently the effect of > (a,b) <- ( (a+b)/2, (a-b)/2 ) > is given by the sequence: > a <- (a+b)/2 > b <- a-b > which implies one shift only. Ah yes of course, this is already better than what I suggested. david