“…When studying random network coding [3,4], Koetter and Kschischang [1] defined a so-called operator channel and found that an (n, M, ≥ 2δ, l) q constant dimension code C could be employed to correct errors and/or erasures over the operator channel, i.e., the errors and/or erasures could be corrected by a minimum dimension distance decoder if the sum of errors and erasures is less than δ. Some bounds on A q [n, 2δ, l], e.g., the Hamming type upper bound, the Gilbert type lower bound, and the Singleton type upper bound, were derived in [1].…”