Our encoding [of a program] is the list of [grammar] rules used in the derivation [of the program]. Each rule is represented … Note that a separate derivation is generated for … Each rule number, in conjunction with the current … [p. 66] The interpreters [for gcc, lcc, gzip, and 8q] are small: 4,029 bytes … Thus adding 9,156 bytes to the interpreter saves … The initial, uncompressed bytecode takes roughly 70 times … but trade-offs favored size at every turn. … [The compression ratio of 29% to 42% is similar to gzip.]
Google-1
Google-2
Copyright clearance needed for quotation.