So, have you got some specs for exactly the way IE and Gecko handle every single case of non-standard code? Including cases where it’s clear the code is broken, but it’s not clear what the author meant, and multiple interpretations are equally valid?
No? There’s no specification? They’ll have to reverse-engineer it by visiting every page on the internet with IE and Firefox and seeing what those browsers do with them? Gee, that sounds workable!