[Octave-bug-tracker] [bug #34251] When I use fzero to determine the roots around 0.5 for tan (x) - sqrt ((1 / x) - 1) I get a non zero complex part

[Jordi Gutiérrez Hermoso]

Follow-up Comment #1, bug #34251 (project octave):

I can't reproduce this in Debian GNU/Linux with the current stable or
development branch.

You have a sqrt in your function, so it's conceivable that fzero attempts at
least one negative argument for fun which is why you're getting imaginary
parts.

Would you agree to close this bug as not a bug? Numerical error is a fact of
life.

Can you reproduce this with fsolve instead of fzero?

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?34251>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/