Lossless Data Compression Testbed: ExCom and Prague Corpus
Jan Holub, Jakub Reznicek, Filip Simek · 2011
Designers of new compression algorithms have to compare their algorithms with the existing ones. They can test over one of standard corpora and compare the achieved compression ratio with published compression ratios made on the same corpus. The corpora are static collections of files so they get older and do not contain new file types that appear. Moreover one can compare only the compression ratio. The times of compression and decompression are very important as well but harder to compare (get the implementation, adjust for the system, run the tests on the same machine).