Google crawled in the year 2005 around a billion pages to find out what Tags and Tag attributes are used within them.
So this Study shows for example that within the <html> the attribute xml:lang is quite often used, but it’s totally senseless within a HTML Document as no HTML processor does look at it.
It also shows that many people aren’t able to write values within quotes, so mostly wrong attributes and values for them would be the answer.
As you also can find out is that the meta name revisit-after is useless as just one Spider ever looked at this attribute.
Tags: Google, HTML, Mark-up, SEO, Spider