I just came to a section of RR’s book LSS where he talks about lookup tables as a faster algorithm. Here is a proposal. I think there is a sense in which all algorithms are the same. There are all just counting. But they are counting in different spaces. Each algorithm is just a map from here to there, but here is no more valid, natural, useful, then there.
![[info]](http://l-stat.livejournal.com/img/userinfo.gif){kind=small}
compiseverythng's journal