Many of our pages don't validate as valid html.  For example:
http://validator.w3.org/check?uri=http%3A%2F%2Fcreativecommons.org%2Flicenses%2F&charset=%28detect+automatically%29&doctype=Inline&group=0

Some of these errors might be unavoidable, since these pages use the HTML5
doctype but also need to have some rdfa present.  Some of the validation errors
should be fixed, such as this embarassing one:

"Line 208, Column 172: Stray end tag (embed)."

So, two things need to be done:
1) fix the broken html
2) adjust unit tests to ignore rdfa related errors


I'm wondering if the easiest way to do part 2 is to have the unit test call the
validator overriding the doctype to be xhtml+rdfa as well as calling it with the
correct doctype (html5) and showing only the errors from the html5 validation
that also appear in the xhtml+rdfa.  Not sure if this would require the document
to be polyglot html5 (valid html5 that is also valid xml) to work.

This should also probably be a different unit test than the one for the deeds,
since in the case of the deeds, we *don't* want to ignore rdfa errors.