Wednesday January 25, 2006

Google has a study of how HTML is really being used out in the wild. They've posted their results, Web Authoring Statistics

December 2005 we did an analysis of a sample of slightly over a billion documents, extracting information about popular class names, elements, attributes, and related metadata. The Good, The Bad and The Ugly: it's all in there.

I think this is a demonstration of Google's expanding interest in grokking the semantics of that are latent in document structures. The results are broken down by

• Pages and elements
• Elements and attributes
• Classes (class="")
• HTTP headers
• Page headers, head element contents
• Metadata, meta tag contents
• The body element
• Text elements
• Table elements
• Link relationships (rel/rev)
• The a element
• The img element
• Scripting: The <script> element
• Editors and their custom markup

I'd love to see how the data changes over time. I suspect parts of the web are becoming more orderly (web 2.0 applications are likely using well formed document structures) while the web as a whole is probably atrophying (the vast installation base of crappy or misconfigured tools are likely the preponderant generators of markup). I'm anticipating a lot of interesting data emerging as the ascendance of microformats continues. Goog, looking forward to follow-up surveys!

google microformats html