Jan-Pieter, Thanks, that's very clear, I had tried variations on 1e23x to no avail.
Regards, Rob. > On 26 May 2015, at 10:42, Jan-Pieter Jacobs <[email protected]> > wrote: > > Hey Rob, > > x: converts from standard precision to extended precision. > The problem lies in the fact that 1e23 is a float, and as such can not > exactly represent the number 10^23, and the precision is already lost there. > > The solution is using extended precision values to start with: > 10x^23 > > 100000000000000000000000 > > Jan-Pieter > > > 2015-05-26 11:34 GMT+02:00 Rob B <[email protected]>: > >> Hi, >> >> No doubt this has come up before but I cannot find any reference >> >> Could someone explain this please? >> >> x: 1e22 >> 10000000000000000000000 >> x: 1e23 >> 99999999999999991611392 >> >> Regards Rob. >> ---------------------------------------------------------------------- >> For information about J forums see http://www.jsoftware.com/forums.htm > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
