Technographer.net

As part of the Technographer.net site is a mirror of the Ogg version of the Linux Void podcast. For all the podcatching people out there, there is an RSS feed for that section of the site, which is an independent Wordpress installation on the subdomain.
Until now, I have used Feedburner - a very useful service which will generate an RSS feed from a site with sound links; this is practical and very easy to use. However, the feeds they create are huge. The automated process simply scrapes every last scrap it finds, and this gives some unfortunate results. Here is an example from the podcast client Gpodder:

So you see, there is a point in doing it yourself. It is, however, not too difficult - once you know how.
Examine the code below. This is a valid RSS feed file; it is an excerpt from the file placed at http://linuxvoid.technographer.net/soundfeed.xml:

<?xml version=”1.0″?>
<rss version=”2.0″ xmlns:atom=”http://www.w3.org/2005/Atom”>

<channel>
<title>Linux Void - Ogg</title>
<link>http://linuxvoid.technographer.net</link>
<description>The Linux Void podcast by Greg and Peter</description>
<language>en-us</language>
<copyright>2008 Licensed under CC-by-sa: Attribution-ShareAlike 3.0</copyright>
<lastBuildDate>Sat, 29 Nov 2008 21:22:00 +0100</lastBuildDate>
<webMaster>morten@technographer.net (Morten Juhl-Johansen Zölde-Fejér)</webMaster>
<atom:link href=”http://linuxvoid.technographer.net/soundfeed.xml” rel=”self” type=”application/rss+xml” />

<item>
<title>Episode 12 - Progress?</title>
<link>http://linuxvoid.technographer.net/?p=85</link>
<guid>http://linuxvoid.technographer.net/?p=85</guid>
<description>* Review: TinyME * News: Songbird 1.0rc1, performance comparison: MacOS X 10.5.5 and Ubuntu 08.10, Creative X-Fi Linux driver, MPD v0.14~alpha1, new Ogg feed address * GNU Screen, Eterm</description>
<pubDate>Mon, 10 Nov 2008 19:53:00 +0100</pubDate>
<enclosure url=”http://linuxvoid.technographer.net/files/lvoid12.ogg” length=”19420285″ type=”application/ogg”/>
</item>

<item>
<title>Episode 11 - Hawking Pumpkins</title>
<link>http://linuxvoid.technographer.net/?p=25</link>
<guid>http://linuxvoid.technographer.net/?p=25</guid>
<description>* Review: ALTLinux    * News: Banshee 1.3.3, Access Linux, new Linux Action Show format, Damn Small Linux 4.4.7, Windows 7, Ubuntu 8.10 Intrepid Ibex * Tips and apps: WICD, Netcfg</description>
<pubDate>Tue, 04 Nov 2008 11:22:00 +0100</pubDate>
<enclosure url=”http://technographer.net/files/lvoid11.ogg” length=”49301458″ type=”application/ogg”/>
</item>

</channel>
</rss>

As one can see, this is an XML file with all values encapsulated in tags, with some nested within others. The doctype and the RSS version are defined.
There is a preamble: A section describing metadata covering the channel feed - and after that the data for the individual episodes. Not all the preamble data is necessary, but much is certainly useful. Not everyone will be interested in aspects like the language, but with the concept of the semantic web, you should be able to interact through accessible metadata - through podcast aggregators, sites and clients. A free software and free documentation enthusiast will most likely want to include a licensing comment. The date for the last time you updated the feed is useful for yourself as well as the one using it - and a webmaster address, in case something goes south.

The individual <item>s hold the podcast episodes. As I have set a Wordpress up for linuxvoid.technographer.net, I have included the shownote link in the individual episodes. Note that this link is not the link to the file, but to a URL describing the episode. Also, note that this is not intended for the site URL, as it is is defined in the preamble. It is intended for if you have a text referring to this particular item.
The <description> is just that. There is a small detail about the formatting in this field: I experimented with using line breaks in the file. Gpodder would respect it, Firefox would not.
The <pubDate> - publication date - formatting is a bit tricky, but using the short format layout as indicated works across the board.
And then, there is the part which sets this feed apart from a regular RSS feed: The embedded multimedia link, the <enclosure url>. There is a link to the file location, the full size in bytes and the MIME type for it to be handled correctly.

This is, of course, just a short introduction based on the experiences with writing the feed. More on this can be learned from the W3schools article on RSS, which goes into a more general introduction to RSS feeds.

edit Dec 1st 2008: Added ATOM and GUID tags.