0j_40 ": x: ". 40$'1234567890' 1.2345678901234568469964621180726096691200e39 0j_40 ": ". 'x',~ 40$'1234567890' 1.2345678901234567890123456789012345678900e39
Extended precision behaviour is explicit because they consume an order of magnitude more resource, and it is undesirable for users to stumble upon it by accident. On Sun, Sep 15, 2013 at 2:28 PM, Dan Bron <[email protected]> wrote: > Does it? Even for integral arguments (i.e. where the internal datatype > guarantees that the numbers are whole)? > > Any idea why? How did you discover this conversion? > > -Dan > > Please excuse typos; composed on a handheld device. > > On Sep 15, 2013, at 4:47 PM, Raul Miller <[email protected]> wrote: > > > On Sun, Sep 15, 2013 at 10:42 AM, Dan Bron <[email protected]> wrote: > >> Yep. Put an 'x' behind or a 'x:' in front. > > > > Note, also, that x: in front generates a floating point representation > > of the number as an intermediate result. This would be good in > > contexts where too much precision was undesirable (where we think of > > it as noise), and bad in contexts where the precision represented > > meaningful information. > > > > -- > > Raul > > ---------------------------------------------------------------------- > > 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
