My latest phone bill informs me in small print that "This bill is protected by one or more of the following U.S. Patents: Des 385,298; 390,599; 5,845,942; and 5,951,052", which apparently cover innovations such as "Bill having one or more information panels and a perpendicularly oriented remittance panel".  I know, I know, Qwest gains substantial competitive advantage from these patents.  Next they'll be patenting the idea of invoicing their customers by mail.

James Robertson has been losing sleep over the future of Atom. In the interest of public service, I've created an XSLT stylesheet to create an Atom feed from a result set from the Amazon.com web service API. For example, top selling DVDs. This stylesheet should work for any Amazon search, on the inductive principle that it works for this one - I'm leaving the other 2 steps in the proof by induction up to the reader ;-). I've released the stylesheet into the public domain, so maybe Amazon can add an Atom option to their syndication page.

This excercise did bring up a question about how to handle the <issued> and <modified> subelements of <entry>. These are required subelemets, according to the docs I've seen (largely Tim Bray's RNG schema, though the wiki seems to bear this out). However, I'm at a loss to determine how to create values for these. It seems like logically, the modified element should be the date that the request was made. Issued could be the same date time, or could be the date/time of the original request, I suppose. However, none of that information is available from the raw Amazon response, and I don't know of a way to find out the current date and time from straight XSLT. If you knew what XSLT processor Amazon was using, you might be able to use an extension function (actually, I checked and though it's using Xalan-C, no extension library appears to be available). I noted that Amazon is using dates in a format like "26 August, 2003", which isn't compatible with the xsd:DateTime format that Atom is using, so you can't easily use that date for any of the date elements, though I'm not sure you'd want to. Anyway, the upshot is that technically, this XSLT won't produce a valid Atom feed, though you could get close, though without dates, I wonder how useful the RSS feed would be. I'll have to try plugging Amazon's RSS into an aggregator and see how it handles changes to the feed.