On 4/13/07, Raymond Toy <toy.raymond at gmail.com> wrote:
>
> Jay Belanger wrote:
> > Raymond Toy <toy.raymond at gmail.com> writes:
> >> I also don't understand why the defaults for maxima should cater to the
> >> least experienced and knowledgeable users.
> >
> > Because they're the ones we want to start using Maxima, because
> > they're the ones who are least able to change the defaults ...
> >
> Except that they're only a beginner once. Tell them how to change it,
> point out the numerous web pages that explain floating-point etc. Once
> they're past that stage, the defaults may be annoying.
Yes, they are only a beginner once. Once they are no longer beginners, they
can edit their init files etc. (Though actually it would be nicer if we
could provide a way *within Maxima* to change defaults, e.g.
set_init(fpprintprec,6) )
And what about printing of bfloats? Bfloats are now printed with the
> specified precision. When I use bfloats, I want to see all the extra
> digits. Otherwise, double floats probably would have sufficed.
I agree that using fpprintprec to specify both the printing of floats and
the printing of bfloats is a problem, especially since different bfloat
objects may have different precisions. Again, thinking of the naive user,
if I say bfloat(%pi),fpprec:100, I expect 100 digits to come out.
if you can't
> understand some points about floating-point, and can't understand
> scientific notation (as someone else mentioned), then I would think most
> of the symbolic stuff would be incomprehensible.
>
Actually, I think that understanding floating point etc. comes long after
understanding basic algebra. A beginner in algebra (9th grade, say), can
understand factor(x^3-1) and bfloat(%pi),fpprec:100 and
trigsimp(sin(x)^2-cos(x)^2), but I wouldn't expect him or her to understand
rounding and cancellation errors.
-s