SGML and other issues in parser for web browser

Hello all,
I am trying to write a parser for web browser called "w3" it is
distributed with emacsen, but the development work on it is now moribund.
While looking at the parser's code(it was last modified in 1996) I saw
lot of hacks for dealing with SGML.
I felt like removing them but before doing that I thought it would be
wise to draw upon the wisdom of other people who are involved in
development of browsers.
Are there still lot of legacy web pages out there which uses SGML?
Should I retain support for SGML or remove it(doing so will help the
parser in speed as well as keeping the code simple). And speed is a big
concern for this browser.
At one point of time I was also contemplating making a parser just good
enough for eating XHTML 2.0, but then thought XHTML 2.0 is hardly used
by anyone, and so I should support HTML 4.1. Any opinions on this are
also welcome.

Also some suggestions on design of the parser are welcome.

Thanks.
--
Surendra Singhi
www.public.asu.edu/~sksinghi/

"O thou my friend! The prosperity of Crime is like unto the lightning,
whose traitorous brilliancies embellish the atmosphere but for an
instant, in order to hurl into death's very depths the luckless one they
have dazzled." -- Marquis de Sade