📚 MBFA — Canterbury Corpus Benchmark

The Canterbury Corpus is the standard benchmark dataset used in academic compression research since 1997. It covers a representative spread of real-world file types — English prose, source code, HTML, binary data, executables, and spreadsheets. Results on this corpus are directly comparable to published compression research. Lower % = better compression. C = compress ms, D = decompress ms. All MBFA roundtrips must pass for results to be valid.