Re: Re: Ask Mathematica to sub for a long expression

*To*: mathgroup at smc.vnet.net*Subject*: [mg71953] Re: [mg71921] Re: Ask Mathematica to sub for a long expression*From*: Daniel Lichtblau <danl at wolfram.com>*Date*: Tue, 5 Dec 2006 06:04:58 -0500 (EST)*References*: <ekp4o4$2fq$1@smc.vnet.net><ekudat$86k$1@smc.vnet.net> <200612041138.GAA02860@smc.vnet.net>

carlos at colorado.edu wrote: > aaronfude at gmail.com wrote: > >>Hi, >> >>Thanks for all the responses. It doesn't quite work for me, but maybe >>with a little extra expertise it will. The problem is that that pattern >>is not always obvious w/o extra algebra. The basic problem is >>summarized in the fact that the following doesn't work the way I'd like >>it to: >> >>-a-b/.a+b->c >> >>2a+2b/.a+b->c >> >>Many thanks in advance! >> >>Aaron Fude > > > Replacing nontrivial patterns sometimes works, sometimes it doesnt. > It may depends on FullForm details. This violates a basic tenet of > computer science: user-level results should not depend on internal > representation. Here is a simple example. > > ClearAll[f,a,b,c,i]; > f[b_,c_]:=1/(c+b^2); > v=Table[{i,f[b,i]/.b^2->16},{i,-6,6}]; > ListPlot[v,PlotJoined->True]; > > Try to track the source of the error message and the plot gap. Rule-of-thumb tenets, basic or otherwise, yield to inviolable restrictions. An example of the latter is that syntactic pattern matching conforms to representation. Daniel Lichtblau Wolfram Research

**References**:**Re: Ask Mathematica to sub for a long expression***From:*carlos@colorado.edu