Book Generation

In ReadwriteWeb:  More detail on Phil Parker's econometric book generation algorithms.  That are said to be generating millons of books that are sold on Amazon.  This reminds me of a methodology called 'content analysis', which was used starting in the 60s to generate analyses of verbatim descriptions.  Then produce scores and then use the scores to create interaction.

Based on the article, this kind of knowledge =book generation works best in highly specialized, narrow domains of knowledge, where the components of the knowledge already exists.  A little like taking a list of URLs that you might use as an introduction to a topic, then converting that to more sculptured prose.