I wonder and muse on whether, in reality, It was wise to give browsers the ability to render markup using the 'Tag Soup' rendering engine? or at least such a good ability.
Of course we know that without it more than half the web would vanish overnight, but I am constantly staggered by quite how effective it is at rendering code that is so completely wrong, code that you just stare at and honestly question how or why it has any right being parsed.
I further muse whether it may have been somehow! better to have had a mechanism by which if one wanted to make use of the power and flexibility and benefits derived from Standards and CSS that perhaps this should perhaps have been a function provided through the strict use of 'Strict DTD's only and that in using these the Tag Soup engine would be asked to be far less lenient with code parsing and rendering, thus clearly de-marking a line between good code and bad; I know that true XHTML provides for this and also that the XMl parser has been criticized for being far too strict and non fault tolerant, but I can only see that as a good thing , it says you Must write proper code or else we will return a page list of your errors, not that great a hardship or impossible goal to reach.
I cant help but feel that while a page can be so hopelessly mall-formed in markup and yet still render perfectly well, it provides yet another impediment to Standards really forcing people into good practises.
I know this isn't a serious argument, fault tolerant rendering is a necessity and one that is never going to be removed as such, just dreaming a little.