Matthew Petroff, an undergraduate at Johns Hopkins University, wrote in with his cool Kindle Weather Display:
The system consists of two parts, one running on the wall-mounted Kindle and one running on a server. The server fetches and processes weather data to create an image, and the Kindle downloads and displays it. While the Kindle is capable of handling both parts, I already had a server, and utilizing it saved me the time and hassle of setting up a build environment for the Kindle.
The server side of the system uses shell and Python scripts to convert weather forecast data into an image for the Kindle. The scripts first download and parse forecast data from NOAA via the National Digital Forecast Database XML/SOAP Service. After parsing the data, the data then needs to be converted into an image. This is accomplished by preprocessing a specially crafted SVG file to insert temperatures, forecast symbols, and days of the week. This SVG is then rendered as a PNG using rsvg-convert and converted to a grayscale, no transparency color space as required by the Kindle using pngcrush. Finally, it is copied to a public location on the web server.
Also check out Matthew’s guide to modding a point-and-shoot to take IR photos.