On Tue, Sep 14, 2010 at 11:13 AM, Johann Hibschman <[email protected]> wrote: > Yes, but now you can't pretend that this is even a toy model of > numerical code. Continuing to machine precision in a solver is, well, > ridiculous. > > I know this is educational code, and the power-based solution is > educationally interesting and all, but it doesn't generalize to reality.
I believe, from your phrasing, that you posted this in jest. But there are contexts where a moderate epsilon can be useful. -- Raul ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
