Ah, I stand corrected.  There _are_ rational infinities, but you have to
make them deliberately using x: (or some other means).  e.g.

   (x: _) <.
83121579216557378445487899878180864668798711284981320763518679672151497189239x
83121579216557378445487899878180864668798711284981320763518679672151497189239

   (x: __) >.
83121579216557378445487899878180864668798711284981320763518679672151497189239x
83121579216557378445487899878180864668798711284981320763518679672151497189239

I should have thought more before I wrote my first reply.  For % to be
closed on rationals rational infinities are required.

   % 0 3r4
_ 4r3
   3!:0 % 0 3r4
128

Sorry.




On Tue, Jan 26, 2021 at 10:33 AM 'Sergey Kamenev' via Beta <
[email protected]> wrote:

> I make calculations on elliptic curves on finite fields.
> In elliptic curves exists so-called "point in infinity".
>
> This is how I noticed this problem.
>
> 26.01.2021 18:52, Roger Hui пишет:
> > The reason for these results, dating from very early days and old old
> > versions, is that there are no rational or extended integer infinities.
> In
> > float F rat (or rat F float), the rule is to convert the rat into a
> float,
> > and then apply the F.  Therefore the results in your message are
> "correct".
> >
> >
> > On Tue, Jan 26, 2021 at 3:16 AM 'Sergey Kamenev' via Beta <
> > [email protected]> wrote:
> >
> >> Hi!
> >>
> >> Min and Max function makes unnecessary conversion of type when compared
> to
> >> infinity and loses significant information.
> >>
> >> Example:
> >>
> >>
> >> _ <.
> >>
> 83121579216557378445487899878180864668798711284981320763518679672151497189239x
> >> 8.31216e76
> >>
> >> Should be:
> >>
> 83121579216557378445487899878180864668798711284981320763518679672151497189239
> >>
> >> __ >.
> >>
> 83121579216557378445487899878180864668798711284981320763518679672151497189239x
> >>
> >> 8.31216e76
> >>
> >> Should be:
> >>
> 83121579216557378445487899878180864668798711284981320763518679672151497189239
> >>
> >> Nice day!
> >> Sergey
> >> ----------------------------------------------------------------------
> >> 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
>
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to