Responding to some problems I had with levin Stavros made a small patch to
simplifya which implements a flag which makes maxima convert floating
point numbers to bfloats
at the current precision before doing other simplifications. Here is a
use case
(%i1) form:sin(2*x)-2*sin(x)*cos(x)$
(%i2) fpprec:100$
(%i3) subst([x=0.1],form);
(%o3) - 2.7755575615628914E-17 <=== note low precision
(%i4) numer_bfloat:true$ <=== the "flag"
(%i5) subst([x=0.1],form);
(%o5) -
3.57183559775710931942432358845419049828721873382124581629726826008635\
1987208556861485775068313037593b-102 <=== precision ok now
I would like to propose this flag, or something with similar
functionality, for inclusion in maxima.
Michel