Re: Integration bug
- To: mathgroup at smc.vnet.net
- Subject: [mg38484] Re: Integration bug
- From: adam.smith at hillsdale.edu (Adam Smith)
- Date: Fri, 20 Dec 2002 04:24:27 -0500 (EST)
- References: <atp6v0$909$1@smc.vnet.net>
- Sender: owner-wri-mathgroup at wolfram.com
It gets even stranger:
Indeed replacing "d" by "z" does give the result Pi.
I quickly checked through a few letters and it seems that you get the
correct result only for the letters "z" and "y". All others return
zero.
This is just a guess, but maybe the code "assumes" that this integral
may be part of a multiple integral in dx dy dz (i.e. a volume
integral) since these types of terms often arise (particularly in
physics) and has the correct result of Pi hardcodes in for variables
x, y and z.
Adam Smith
"Dana DeLouis" <delouis at bellsouth.net> wrote in message news:<atp6v0$909$1 at smc.vnet.net>...
> I am new to Mathematica also, but I seem to recall a bug in the use of
> the variable “d" before also. This was the first thing I thought of.
> Changing “d" to something like “z” gives the answer of Pi.
>
>
> Removeall
>
> All Global` variables Removed!
>
> $Version
>
> 4.2 for Microsoft Windows (June 5, 2002)
>
> Integrate[Sin[z + x]/(z + x), {x, -Infinity, Infinity}]
>
> Pi
>
> Integrate[Sin[d + x]/(d + x), {x, -Infinity, Infinity}]
>
> 0
>
> (Note: Removeall is my own function that removes all global variables
> and is included to show that all variables are cleared.)
>
> Dana DeLouis
> Windows XP
> $VersionNumber -> 4.2
> = = = = = = = = = = = = = = = = =
>
>
> "David W. Cantrell" <DWCantrell at sigmaxi.org> wrote in message
> news:at4cv9$eta$1 at smc.vnet.net...
> > Using version 4.1,
> >
> > Integral[Sin[x+d]/(x+d),{x,-Infinity,Infinity}] yields 0 .
> >
> > This is, of course, incorrect. (Does version 4.2 make this error
> also?)
> >
> > Mathematica does Integral[Sin[x]/x,{x,-Infinity,Infinity}] correctly
> > however, yielding Pi, which should also be the answer for the original
> > integral (regardless of the value of d).
> >
> > Does anyone have an idea why Mathematica gets the original integral
> wrong?
> >
> > David Cantrell
> >
> > --
> > -------------------- http://NewsReader.Com/ --------------------
> > Usenet Newsgroup Service New Rate! $9.95/Month 50GB