> for any expression that can be evaluated numerically,
> a method for evaluating it to arbitrary precision. That is,
> N[ ...., d] evaluates to "Accuracy" d.
Well, yes, that would be nice... though problematic for discontinuous
functions, e.g. N( signum(sin(2)^2+cos(2)^2-1), 1) where you need to
calculate the inner expression to infinite precision....
Note that IS has this issue (Barton filed a bug report) -- it uses
numerical evaluation in some cases.
-s