Define explicit handling rules for HTML 4 elements

[martintoreilly]

At the moment we unwrap all elements not defined in the HTML5 standard by default. However we cannot tell if an unknown element will be treated as a block or inline element by the browser. Therefore we risk merging two blocks of text that would be rendered as separate visual "paragraphs" into a single visual "paragraph". By explicitly define parsing behaviour for "knowable" pre HTML5 tags we can minimise the risk of treating a "knowable" block level element as inline - see issue