Heureka! In byte mode (treating a byte as a unit, not a text line), LINECOMP (or BYTECOMP?) compresses enwik8 to 33 MB - as opposed to the 35M original .zip file. So the algorithm is actually a "zip-enhancer", if you will (we use zip for the final compression internally). And this is with the same really, really crude intermediate encoding shown above (mostly just numbers in decimal format).
(BTW I think Matt's web server is moody. Couldn't download enwik8 from multiple machines.) ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/Tb2cf064c700f181c-M46f7c9ecc6cae0c529d70843 Delivery options: https://agi.topicbox.com/groups/agi/subscription
