Charles P: The problem is that > >the present INT function rounds the result up if the fraction is >= 0.5. > > This is almost never desired. I'll bet that almost 100 percent of the > >time the INT function is used it introduces unintentional errors, just > >like the rounding examples in this thread, and for that reason alone it > >should be eliminated. If you want to round, you need two parameters: > >ROUND(number, places). > >Trying to include rounding in a single-parameter function is not a > >feature, it's a bug. > > > >Regards, > > > >Charles P. > > Robert Covington: > Charles, is your beef with INT that > > INT(1.49) returns 1 in FB 3 > > but > > INT(1.5 ) returns 2 in FB 3 > > ? > > I can see your point, as I had a bit of adjustment time taking to the FB 3 > way of INT over the FB II version, but prefer it now. I would lose data > during convolutions if it was to return a 1 for INT(1.9) > > You can still have your INT(1.5) = 1 by using FIX(1.5) > > ? > > Robert Covington I converted all my INT's to FIX. Charles P.