Re: Yet another Version 4.2 Integration howler
- To: mathgroup at smc.vnet.net
- Subject: [mg35301] Re: Yet another Version 4.2 Integration howler
- From: "Kevin J. McCann" <kjm at KevinMcCann.com>
- Date: Sat, 6 Jul 2002 06:10:31 -0400 (EDT)
- References: <ag3eou$73f$1@smc.vnet.net>
- Reply-to: kjm at KevinMcCann.com
- Sender: owner-wri-mathgroup at wolfram.com
Vladimir Bondarenko wrote:
>
...............................................................................
>
> BUG # 3617
>
> For an expression having a concrete numeric value, Indeterminate is
> reported
>
> BUG HISTORY: PRESENT 4.2 for Microsoft Windows (February 28, 2002)
> ABSENT 4.1 for Microsoft Windows (November 2, 2000)
> ABSENT 4.0 for Microsoft Windows (April 21, 1999)
> ABSENT Microsoft Windows 3.0 (April 25, 1997)
> ABSENT Windows 387 2.2 (April 9, 1993)
>
> Reproducible Always
> BuildNumber 156656
> CPU ID AuthenticAMD AMD Athlon(tm) XP 1600+
> RAM size 512 Mb
> Free HDD size 11 Gb
> OS ID Microsoft Windows 98 4.10.2222 A
In fact, the integration from 1 to infinity is Catalan's constant, and, of
course, the integral from 0 to 1 of the integrand is the negative of it.
Nice upgrade ;(
Kevin
>
> The value of the integral is defined.
> However, Mathematica reports Indeterminate which is invalid.
>
> Integrate[Log[z]/(1 + z^2), {z, 0, Infinity}]
>
> ACTUAL: Indeterminate
>
> EXPECTED: 0
>
> CHECK-UP: NIntegrate[Log[z]/(1 + z^2), {z, 0, Infinity}] // Chop
>
> 0
>
>
...............................................................................
>
>
> Best,
>
> Vladimir Bondarenko