Finding only the real roots

*To*: mathgroup at smc.vnet.net*Subject*: [mg113594] Finding only the real roots*From*: Equilib <petterbe at gmail.com>*Date*: Thu, 4 Nov 2010 03:58:20 -0500 (EST)

Hi all, I'm trying to numerically evaluate a function using NSolve, for example Nsolve[1/2+c+g(1-p)+root(c+p*g-1)==0,p] where my two variables g and c are both in the interval [0,1]. But, since the root is not always positive I do not get a solution to NSolve. The idea was therefore to first define the positive roots and then to do NSolve. My questions are therefore; 1) Is this the best approach, or are there more direct ways? 2) What should I do to only get the real roots? 3) If I first solve the root, and then use NSolve, will I still be able to use functions such as Manipulate on the NSolve solution? Best, Petter