firstly i'd like to say that the numerical methods textbook is easily the best thing about going to unisa for me! it's
excellent, typos aside.
i haven't properly looked at the shooting method, but from the name and briefly reading the description, it sounds like you pick an arbitrary-ish starting value, follow the differential equation to the x-value where your boundary condition is defined. most likely you won't hit the exact y-value for which the boundary condition is given, so the whole process becomes a search problem: which initial value gets me (by following the DE) to the right value where the boundary condition is defined?
this kind of makes sense to me, but again i haven't properly looked at it yet (suffering mat307 atmo, just squizzed the abstract description of the method) so what i wrote could well be wrong. have a look online (eg. wikipedia), i'll look there eventually too...
edit: typing this out actually cleared it up for me a bit, so i looked around online (too lazy to get my book out the cupboard, lol) and it seems my guess wasn't too far off . some resources:
http://en.wikipedia.org/wiki/Shooting_method
http://numericalmethods.eng.usf.edu/topics/shooting_method.html