An encoder structure for an error correcting code with arbitrary parity positions is presented. The invention is effective for all error correcting codes whose parity check matrix is of the Vandermonde type. In contrast to conventional encoder circuits, the parity symbols produced by this encoder are not restricted to form a block of consecutive parity symbols at the beginning or at the end of a codeword, but may be spread arbitrarily within the codeword. A general structure of the parity check matrix for such a code is found by exploiting the special Vandermonde structure of matrices. From this general parity check matrix, an expression for the evaluation of the parity symbols in terms of a polynomial with limited degree is derived. An efficient hardware implementation of the proposed encoder is suggested.