Re: Bug of Integrate
- To: mathgroup at smc.vnet.net
- Subject: [mg82770] Re: Bug of Integrate
- From: Thomas E Burton <tburton at brahea.com>
- Date: Tue, 30 Oct 2007 05:46:44 -0500 (EST)
Sort of a bug, more of a misunderstanding. The argument has two
roots, +/-3. You took one value; version 6.0.1 took another. Version
5 takes a mixture of the two roots, apparently, and yields 0. Both
versions yield
3*t*Cos[t]*Sqrt[Sec[t]^2]
for the indefinite integral; that is, neither version hazards a guess
as to which sign of the square root you want. The "bug" occurs when
you ask for various definite integrals. Versions 5 and 6 differ from
each other and also vary by what limits you choose. For instance, if
you integrate from 0 to Pi/2, version 6 chooses the positive root,
and if you integrate from 0 to Pi, you get zero! Given the ambiguity
of the integrand, none of these answers are wrong, exactly. Beware.
> When I try to calculate the integral
>
> Integrate[Sqrt[1/Cos[t]^2]*3*Cos[t],{t,0,2Pi}] Mathematica 6.0.1
> yields -6*Pi.