The feeds are aggregated every thirty minutes using Perlanet, a Perl program that I wrote. The list of sources to aggregate is kept in a YAML file. Every thirty minutes, the source feeds are read and combined in a single feed. The most recent entries from this combined feed are republished as a new feed and as an HTML page. The feed handling is all done with XML::Feed and the HTML page is generated using the Template Toolkit.
Actually, there are two sets of feeds that are aggregated. The longer feeds are in one set and the shorter feeds are in the other. The two web pages that are created are published on the front page of this site using Server Side Includes.
If you’re interested in even more details, here are the files used to build the two streams:
Another similar process, combining all of the feeds into one page and one feed, is used hourly to construct planet davorg.