Re: [Maxima-discuss] bfloat issue

[Dimiter P. <dim...@gm...>]

Hi Barton,

Your approach reminds me about the suggestion of Richard re handling of the
infinities.

I had something else in mind actually:

( "definitions and predicates",
declare(infinitesimal, feature),
declare(unlimited, feature),
infimp(x):= featurep (x, infinitesimal),
unlimp(x):= featurep (x, unlimited) or not freeof(inf, x) or not
freeof(minf, x) or not freeof(infinity, x),
finitep(x):= not (unlimp(x) and infimp(x))
)$

(
declare(zeroa, infinitesimal),
declare(zerob, infinitesimal)
 )$

declare(ST, linear)$
declare(ST, multiplicative)$

(
matchdeclare(v, infimp),
defrule(simpist1, ST(v), 0)
)$

and then write a simplification and evaluation function.

In this way all infinitesimal values can handled based on pattern matching
rules.
BTW csign is not documented in the help.


best regards,

Dimiter