I had 0/0 so I derived 1/x^3 and (1/sqrt[x^2+1] - (1+ax)/(1+bx) separately, simplified the equation and reached 1/3 lim x=>0 (b-a)/x^2, but if L is finite I don't really know where to go from there. Did I do something wrong?
it doesn't look possible. I looked at the Taylor expansion of . Even with the best choice of a,b the expansion includes an term and therefore the limit would not exist