RSS Advisory Board

As part of the research for the RSS Profile, I compiled statistics on how frequently RSS core elements and namespace elements appear in feeds.

The full report reveals that the most popular namespace element is dc:language from the Dublin Core namespace, which appeared in 36 percent of the feeds.

The second-most common namespace element is atom:link from the Atom syndication format, appearing in 15 percent.

Four core elements were found in fewer than 1 percent of the feeds surveyed: textInput (0.31 percent), rating (0.26), skipDays (0.10) and the skipDays element day (also 0.10).

The following channel elements were found in 1,933 RSS 2.0 feeds, which were chosen because they appear in OPML subscription lists published on the web. The quantity and percentage columns refer to how many feeds contained the element at least once. Parent-child relationships are indicated by hyphens ("-") in the element names. Namespace elements are indicated by colons (":").

Another report is available for item elements.

As a syndication framework developer, I have always had a strong interest in doing this type of data mining as a means of understanding the syndication feed landscape. Thanks for these statistics!

Is there any way you could provide how you went about compiling this information (data sources, methods, algorithms, etc.) so that those who are interested can do their own analysis?

Sure. I'll write about it on my blog, because I'd like to see people help improve the methodology and pull in a larger number of feeds. It was tough to compile.