I'm reading through this proof that any homogeneous equation $\displaystyle y' = f(x,y)$ can be rewritten as $\displaystyle y' = g(y/x)$. It says that, by homogeneity, $\displaystyle f(x,y) = f(tx, ty)$ for all real $\displaystyle t$ (I think it should have the proviso that $\displaystyle t \ne 0$, but it doesn't.). It then claims--and here's the part I don't get--that an appropriate choice of $\displaystyle t$ may be $\displaystyle 1/x$. I thought $\displaystyle t$ was some real number, not a real-valued function. Am I missing something important here, or should I just take homogeneity to mean something like $\displaystyle f(x, y) = f(tx, ty)$ for any function $\displaystyle t(x,y)$?