Comment on `Decomposition of pure states of a quantum register'

I. Raptis and R. Zapatrin in the quant-ph/0010104 show possibility to express general state of $l$-qubits quantum register as sum at most $2^l-l$ product states. In the comment is suggested more simple construction with possibility of generalization for decomposition of tensor product of Hilbert spaces with arbitrary dimension $n$ (here simplicial complexes used in the article mentioned above would not be applied directly). In this case it is decomposition with $n^l-(n^2-n)l/2$ product states.