Re: Expansion of an exponential expression

• To: mathgroup at smc.vnet.net
• Subject: [mg47303] Re: Expansion of an exponential expression
• From: carlos at colorado.edu (Carlos Felippa)
• Date: Fri, 2 Apr 2004 03:31:59 -0500 (EST)
• References: <200403300902.EAA06956@smc.vnet.net> <c4gg30\$3fh\$1@smc.vnet.net>
• Sender: owner-wri-mathgroup at wolfram.com

```Andrzej Kozlowski <akoz at mimuw.edu.pl> wrote in message news:<c4gg30\$3fh\$1 at smc.vnet.net>...
> On 30 Mar 2004, at 18:02, Carlos Felippa wrote:
>
> > As a result of some calculations I have
> > (a snipet of a more complex expression)
> >
> > anlog= Log[ ((2 + Sqrt[1 + 2*h])^(-1))^(t/h)   ]
> >
> > I would like to simplify this to
> >
> > anlog= (t/h) Log[((2 + Sqrt[1 + 2*h])^(-1))   ]
> >
> > so I can then take the Taylor series in h.
> > But FullSimplify[anlog,h>0]  doesnt do it.
> > Do i need a ComplexityFunction?
>
> There is always PowerExpand.

I avoid PowerExpand in complicated expressions (remember that what
I showed is a tiny part of something much longer).  Very often the
answer is wrong, and then I have to trace back a needle
in a haystack.

In my opinion PowerExpand should be removed from Mathematica.
Paraphrasing Dijkstra, "PowerExpand considered dangerous"

```

• Prev by Date: Re: Expanding \$Path
• Next by Date: Re: Consolidated method for solution of strange character in
• Previous by thread: Re: Expansion of an exponential expression
• Next by thread: webMathematica question