Cas bugs in the evaluation of the logarithm?

[ilario980]

hi,

can you evaluate ln(f(x)) in your favorite cas (where x is a zero of f(x) )? what is the result?

 

[tiny-tim]

Welcome to PF!

Hi ilario980! Welcome to PF!

hi,

can you evaluate ln(f(x)) in your favorite cas (where x is a zero of f(x) )? what is the result?

If f(x) < 0, then ln(f(x)) isn't defined.

Other than that, the limit of ln(f(x)) will be -∞.

 

[ilario980]

yes, i know; but what seems to me strange is that i tryed in some cas to evaluate that value and the ouput is not -inf and is different depending on the function f(x)

after this experience i believe that trust cas may be dangerous...

I.M.

 

[tiny-tim]

yes, i know; but what seems to me strange is that i tryed in some cas to evaluate that value and the ouput is not -inf and is different depending on the function f(x)

Hi I.M.!

(btw, it's case, not cas )

Did you evaluate it using a calculator?

When you get near 0 or ∞, a calculator can give rounding errors that make a lot of difference!

 

[chroot]

He means "computer algebra system," or CAS.

- Warren

 

[tiny-tim]

He means "computer algebra system," or CAS.

- Warren

ah!

thanks, Warren! ​

 

[ilario980]

hi Tiny-Tim,
i was not playing with my calculator keys
i was interpreting (superficially, i admit) some data and this behavior was leading me to an erroneous conclusion


Ilario M.

 

[Borek]

Another reason to not trust computers too much.