<hr><i>anonymous claiming to be kkinnell added on 2009-05-27 15:46:17:</i><br>
Actually, I had added an 'annotate' tag in the kkinnell-exp branch.  What it did
was to strip out the text between the opening and closing tags.  Completely strip it out, so that a "view source" from the browser wouldn't see it.

That was perhaps a bit of creeping featurism (as Richard pointed out at the time) and I was going to make html tags work instead, but <i>I thought</i> someone beat me to it.

If not, the code in that branch is a decent start--take a look at what was added for the annotate tag and treat the text recognized between start and end comment tags the same way.  It's recognizing the comment tags that's the hard part.  &lt;!-- and --> are asymmetrical (and not orthogonal w/ respect to the rest of html) and parsing them may lead to grief, madness, and atonal music.

Or you could wait until I shake loose at work and have some time to work on it again.