comparison current update - HTML

Update on comparison technique for HTML method. 

For HTML we're trying to reduce it to comparison of 2 XML files. There
is tidy tool [1] that cleans and transforms HTML by closing all unclosed
tags and fixing quotes around attribute values. Works for me, has COM
interface, so nicely scriptable.

The problem remains 
a) HTML allows to have attributes without values. Ex: <INPUT disabled />

b) HTML has different rules for white space handling then XML - less
strict.

For a) I don't have solution - it is something that is rare and can be
manually fixed though.
For b) I have solution for windows platform - I can use IE classes to
serialize HTML is a consistent "canonical" way  - this is something some
of our test teams do for html comparison. If we find similar solution
for non-windows platforms, then it's ok. 

[1]  http://www.w3.org/People/Raggett/tidy/