scaling in SVM is always necessary?

Question

I\'ve already red some posts here about SVM scaling and why it is such important. Now I\'m wondering if scale=TRUE in the ksvm function is always n