Dr. Iverson disavowed operator overloading. Since otherwise this seems
like a great case for overloading I propose a new feature. In this, the
rational precision is specified as maximum either denominator or number
of digits therein, and set as a global parameter. Currently it exists
as _ . When R0 u R1 has a larger denominator the result would be
approximated using for example
https://en.wikipedia.org/wiki/Continued_fraction . Implementation of
this proposal would make rational arithmetic wieldy. With our infinite
precision denominators can become ridiculous, and we wouldn't need to
write messy
R1 sub_MyRat_ R0
normalize R0 + R1
Thoughts? Would you be more likely to exploit j's rationals?
Stay well, Dave.
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
