Re: Possible Mathematica 4.0 bug
- To: mathgroup at smc.vnet.net
- Subject: [mg30836] Re: Possible Mathematica 4.0 bug
- From: Ioan Alexandre Romoscanu <romoscanu at imes.mavt.ethz.ch>
- Date: Thu, 20 Sep 2001 03:51:30 -0400 (EDT)
- Organization: Swiss Federal Institute of Technology (ETHZ)
- References: <9o97d9$cg2$1@smc.vnet.net>
- Sender: owner-wri-mathgroup at wolfram.com
I have seen lots of these things. 1. is not equal to 1 according to mathematica. Check the attached picture to see an explicit example. [Contact the author to get this - moderator/ Cheers I. Christian Mikkelsen wrote: > The following problem cost most of a mornings work and > I ended up attributing it to a bug but before I send > it off to Wolfram Research I'd like some comments. > > I have a polynomial defined as > > H[z_]= (1/z^6) > ((0.17897972007618376+0.1832666067995157 > I)((0.07806788263644916+0.9898486037537829 I)+I > z)((-1.0304811754750547-1.1012670896195078 > I)+z)((-0.4716966687204682-0.6843778680006037 > I)+z)((-0.4537703518350889+0.7354674436955 > I)+z)((-0.32047121654063226+0.18872779388498095 > I)+z)((0.7411819790476893+0.15538114678579676 I)+z)); > > with some auxiliary definitions > > f[w_] = Arg[H[Exp[I w]]]; > fmm[w_] = (f[w+h]-2f[w]+f[w-h])/h^2 /.h->10^(-6) // > Simplify; > d = 1 10^8; > > The following results in an erroneous(?) > simplification > > disp[l_]=fmm[d/l] //Simplify > > but defining d = 1.0 10^8 gives the correct(?) result. > A different result at any rate. Ok, so in principle I > have a workaround but I would be happy to know if the > issue also exists in version 4.1. I don't currently > have access to 4.1 but I am working on that. > > Finally, it is of course possible that I have missed > something so I would be most happy to get feedback on > that as well. > > Cheers, > > Christian Mikkelsen > > ______________________________________________________ > Hold styr på vennerne, familien eller studiegruppen > med Yahoo! Grupper. Du finder dem på www.yahoo.dk/grupper > Yahoo! Grupper - et sjovt og praktisk forum for interessefællesskaber. -- _________________________________________ ioan alexandre romoscanu - institut für mechanik CLA G31, eth zentrum, 8092 zürich, schweiz tel. +41 1 632 77 54, +41 76 323 63 05 fax.+41 1 632 11 45, romoscanu at imes.mavt.ethz.ch temporarily nestlé research center H 212 vers-chez-les-blanc 1000 lausanne 26 tel. +41 21 785 83 65 fax. +41 21 785 85 53 _________________________________________