Adding a "Safe" option?

[Jason Clark]

Okay, I know that the current goal is the 1.0 release, and I know that 
a stated goal of MD is simplicity (no feature creep).  However, I'm 
thinking of modifying my MD for my own use, and I'd like to think out 
loud here a bit, and maybe get some feedback.  If you're a purist, 
please avert your eyes ;-).

I'd like to use Markdown on my site to format comments.  (FWIW, mine is 
a Blosxom blog, not MT.)  Additionally, I do not want to allow any raw 
HTML to beentered in comments.  Only Markup-generated XHTML will be 
allowed.  However, I can't just strip anything that looks like html 
before passing it to markup, since I want to allow code samples to 
contain HTML.  Since Markup already understands code blocks, It makes 
sense to me to put the scrubbing code inside Markdown and make it a 
configurable option.  If I do this, I have a choice of two ways to 
handle user-entered HTML:  I can either strip it, or `treat it like a 
code span`.  Couple this with a mandatory preview before posting a 
comment, and I think my comment system will be simple yet powerful.

For lack of a better term, I've been calling this a "safe" option.  
Does this sound useful?  If I get this to work, I'll of course publish 
the patch.  Am I overlooking anything, functionality-wise?  I've only 
browsed the full source, so I haven't quite determined how I'll 
accomplish this, but I think that MD's methodical parsing strategy 
should work to my advantage.

Jason Clark <jason at jclark.org>
http://jclark.org/weblog/