Please, please do not scrape the HTML pages. The XKCD admin has put a JSON API (http://xkcd.com/json.html) in place for people who want to do things like this. Thanks.