Monday, 26 September 2011

Auto-generated books

The front page of today's Wikipedia had a link to an article on Philip M. Parker. His claim to fame is that he is the author of 107,000 books (according to Amazon), although he claims it is more like 200,000 books. He achieved this seemingly impossible feat using computers, databases and web searches to generate the content of each book.
Of course this approach could be taken to write chess books, and in fact such books already exists. I have a few books, from the 1980's/90's, which could be considered 'database dumps', although such books are not looked upon highly.
Then while searching for something else this evening, I came across a book which is much closer to a 'generated' book. The provocatively titled 'The Final Theory of Chess'  was written with a bank of computers running Fritz 7 (plus later versions). It is essentially an opening manual with a vast number of lines that end with a computer evaluation. There has been some human input, but as the author admits, this 'was kept to a minimum'.
Nonetheless, what the book is trying to achieve (an attempt to push the boundaries of opening analysis) is made redundant by the very technology it employs. Why use a book, when you can achieve the same outcome with your own computer. Of course you don't get an understanding of the opening, but the book did not offer one either.

1 comment:

Bill and Jean Egan said...

Have a look here to see a sample of a computer generated chess book:
This is Australian oriented but there are dozens more. They are created by a straight lift off of material on Wikipedia with no editing, which is legal as long as ackowledged but why pay excessive sums for what's free for all?

The same, or similar, company is churning out print-on-demand copies of lots of public Domain out-of-print books, which may be a partly useful public service though much of this material is available free as e-books from sites like Gutenberg anyway.