I don't think rounding is desirable, here. Instead, what is desired is a mechanism analogous to the display precision where the precision of the collection is based on its largest members.
But all that's needed here, is a loss of a few unnecessary bits of precision. So what we do is add a number (which drops a few bits of precision) and then subtract it again. FYI, -- Raul On Wed, May 2, 2012 at 11:31 AM, Edward Mokurai Cherlin <moku...@sugarlabs.org> wrote: > On Wed, May 2, 2012 6:11 am, Bo Jacoby wrote: >> Hello J'ers >> >> Even if a complex number is actually real it may display nonzero imaginary >> part because of round-off errors. Adding and subtracting 1e6j1e6 removes >> the round-off error, but is there a better or more general solution? > > J uses the APL definition of complex floor, so normal rounding works. > > <.1j0.00000001+0.5 > 1 > _1+<.1j0.00000001+0.5 > 0 > > -- > Edward Mokurai (默雷/निशब्दगर्ज/نشبدگرج) Cherlin > Silent Thunder is my name, and Children are my nation. > The Cosmos is my dwelling place, the Truth my destination. > http://wiki.sugarlabs.org/go/Replacing_Textbooks > > > -- > Edward Mokurai (默雷/निशब्दगर्ज/نشبدگرج) Cherlin > Silent Thunder is my name, and Children are my nation. > The Cosmos is my dwelling place, the Truth my destination. > http://wiki.sugarlabs.org/go/Replacing_Textbooks > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
