Some backend updates to KatrinaFloodNews.com should make the site somewhat better.
<p>I added a custom feed download from <a href="http://news.google.com">Google News</a> which increases the story count by about 20%, towards the bottom of the list. </p> <p>Next, the Google News descriptions are being filtered through <a href="http://tidy.sourceforge.net">Tidy</a> so they aren’t so horrible to work with or debug. Google sure puts out some awful HTML. Chance of <a href="http://validator.w3.org">validation</a>: 0%. HTML Tidy does character set translation which fixed up some display glyph problems. And has its own unique configuration file format…</p> <p>Finally, since I stand a chance of parsing the news summaries after they go though Tidy, I can exclude the links from the text the heuristics match, from which I’ve seen some false positives. Currently all stories are relevant.</p> <p>So, overall it’s incrementally better, with the worst problems zapped. Next Up: Database backend so I can generate an RSS feed.</p> <p>Oh, and I did some <a href="http://news.yahoo.com/news?tmpl=story&u=/prweb/20050907/bs_prweb/prweb281183_1">PR</a> on it the other day. Interesting process.</p> <p>Now, for a survey: Google has some good filters in place to prevent people from abusing the AdWords system for Katrina-related searches. In fact, the ad links are downright <i>useful</i>. So, would it be crass to add AdWords to the site under the Red Cross banner? Leave a comment if you have an opinion.