HTMLParser is still simpler than html5lib, but if/when possible we are following the HTML5 standard rather than taking arbitrary decisions (like we used to do before HTML5).  HTMLParser doesn't claim to be a fully compliant HTML5 parser (and probably never will) -- its goal is rather being able to parse real world HTML in the same way browsers do.  There are also a couple of corner cases that are not implemented in HTMLParser because it would make the code too complicated for a little gain.