The eventual goal is a mind of sorts. There are pieces of data with lots of different paths to them.

The data I want to start with is the Hugo and Nebula Award winners and nominees. Once upon a time, [this program](http://dhappy.org/.../lib/calibre/download_bot.py) downloaded the title and author information from Freebase, then connected to an irc channel, searched for the books, unzipped the contents, and put them in a particularly structured filesystem.

I have, from that endeavor perhaps 150 works from a total I can no longer calculate because Freebase is gone.

The first step is to get back a dataset of works to search for. Fortunately, <a href="http://www.thehugoawards.org">the Hugo Awards</a> has a regular set of pages for the years.

So, I ran <code>wget -mk http://www.thehugoawards.org/hugo-history/</code> and wrote [a ruby script to consume those files](slurp/rb) and dump the info to Postgresql using ActiveRecord from Rails.

Next I generated a set of scaffolds for a basic HTML interface and edited them to display various subsets of the data.

My previous program drawing from Wikidata only downloaded winners, so there are lots of holes.

In looking to fill those I discovered the file lists from the OmwnServe bots in #ebooks. The old program searched for each title then downloaded from the search result.

The new one will search from the lists of available files. Both massively more efficient and I can see all the options.

I'm still enamored with the bicameral nerual network.