We are talking about the rules for numeric vector constants, a wild and wooly sub-universe. Catenation is different. For catenation (and other functions) the table in Section II G applies.
----- Original Message ----- From: Devon McCormick <[email protected]> Date: Wednesday, February 18, 2009 8:17 Subject: Re: [Jprogramming] ill-formed number and extended precision To: Programming forum <[email protected]> > Doesn't this show that the final number is being coerced to > floating point > in the concatenated case? > > 32j3":(x: 1234567891011)*123.4567891011,x: 1234567891011 > 152415787751534.840 1524157877515348500000000.000 > (x: 1234567891011)*x: 1234567891011 > 1524157877515348374602121 > NB. Final digit is one, as expected. > > On Wed, Feb 18, 2009 at 10:43 AM, Roger Hui <[email protected]> > wrote: > > The dictionary could be clearer on this point. The example > > is running afoul because the system thinks the 2000x is > > missing some trailing stuff for the base e form: > > > > 2000x_0.5 > > 1213.06 > > 1x1 > > 2.71828 > > > > > > > > ----- Original Message ----- > > From: david alis <[email protected]> > > Date: Tuesday, February 17, 2009 22:18 > > Subject: [Jprogramming] ill-formed number and extended precision > > To: Programming forum <[email protected]> > > > > > Should the following behaviour be described in Appendix G > > > http://jsoftware.com/help/dictionary/dictg.htm > > > (G. Extended and Rational Arithmetic) > > > or is it a bug that's already been reported? > > > > > > 9.8 8 2000x > > > |ill-formed number > > > 98 8 2000x > > > 98 8 2000 > > > > > > j602/2008-03-03/16:45 ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
