Also, wouldn't it be better to call it a hash table or a lookup-table rather than a compression algorithm. The key is swift and appropriate recall. Example: Compare a long-time practicing theoretical physicist with a physics grad student. Both know most of basic quantum mechanics. But the experienced physicist would know when to whip out which equation in which situation. So, the knowledge content is not necessarily compressed (I'm sure there is some compression) as much as the usability of the knowledge is much greater.