Because Internet and the non limited number of blogs is a nice source of words, I use to crawl the web to fill the Reverse database. So, sorry to:

• ha.ckers.org (an amazing website!)
• yaronet.com (for french words)
• myspace.com ^^

but it was the origins of my scan.

The crawler is a very simple 100-lines Python script, even if it's not the fastest, it's nice enough to be used here.