The problem of almost global consensus on the n-sphere