Re: Replacement Rule with Sqrt in denominator

*To*: mathgroup at smc.vnet.net*Subject*: [mg114374] Re: Replacement Rule with Sqrt in denominator*From*: Daniel Lichtblau <danl at wolfram.com>*Date*: Thu, 2 Dec 2010 05:41:25 -0500 (EST)

Themis Matsoukas wrote: > I will have to agree with kj and Richard Fateman that this is both > annoying and borderline bug. It certainly can be a nuisance. It's not a bug. It might be a feature, if you define feature much differently than I do. > Of course there is an explanation as to why this happens but > this doesn't make it correct. I know that Mathematical looks for > literal occurrences when it applies replacement rules, and so I > cut-and-pasted the denominator to create the replacement. Obviously > what I see is not what there is. Nonetheless, when I manipulate an > expression I go by what I see, and Mathematica in this case was deceitful. > > tm The alternative would be to output everything in FullForm. That's not something many people would want. So for better or worse we go with the form you refer to as "deceitful". Changing the semantics of rule replacement to what amounts to "read my mind" is not an alternative. Daniel Lichtblau Wolfram Research