In classical crypto, homophonic substitution attempts to mitigate risks of frequency analysis via employing one-to-many mappings of plaintext characters to ciphertext characters instead of one-to-one mappings. However, constraints of manual processing naturally very severely limit the extent to which the flattening of the frequency distributions of the ciphertext characters could be achieved, which obviously is on the other hand a non-issue for computer processing. I have written a Python code HOMOPHONE-SP for performing homophonic substitution (in combination with transpositions) where, for obtaining higher security, the processing of the individual plaintext characters is chained via borrowing an idea from modern block ciphers. The software is available at http://mok-kong-shen.de M. K. Shen