Jim,
I like how you figured out how to cancel y' and y^2. Very clever!
I think I can shed some light on the IC issue. The first order problem you posed only needs one IC, but when you convert it to a linear 2nd order DE, that in general needs two ICs for a unique solution. However, the...