If you are doing anything in terms of publishing RSS feeds, from blogs, for podcasts, etc, keep posted in front of you a reminder to start and continually running your published feeds through a feed validator.
Problems may not be visible, as many News Readers are forgiving on RSS miscues (like web browsers are with sloppy HTML), but with more and more “services” built on content shared by RSS, we might keep forgetting that this is XML under the hood, and XML is pretty specific (for good reason) on structure.
At least 85% of the problems people write to me about Feed2JS are problems where RSS feeds fail validation. This is critical, since RSS feeds are parsed by MagpieRSS library and it will bonk out quite often if a feed is not valid (tight structure is needed to parse feeds).
Common problems are weird characters in feed content (like unescaped & and odd-ball characters that find their way there via Word cut and paste), and more often than not, invalid formats for item publishing date and time.
And this is not a problem with Mom and Pop little blog sites; just today I was trying to help someone diagnose a problem with a CNET feed that had a lot of structural problems and an invalid pubdate format:
They have the channel’s image tags outside the channel tag, which munges up all further validation, and the pubDate format does not match the RSS 2.0 spec.
Maybe you think this is petty or technical ranting, but the small pieces only join well when there are some established rules followed that enable the joining. It’s not only C-Net, I’ve found bad feeds from Wired, and other firms that have armies of technical staff who should know better.
And this is just the beginning as I begin to ponder all the extra pieces of the iTunes RSS extensions.
Validation is easy to get. Do it now.
The post "Three Rules of RSS Publishing." was originally rescued from the bottom of a stangant pond at CogDogBlog (http://cogdogblog.com/2006/01/three-rules-of-rss-publishing/) on January 21, 2006.