>But the question is: how can you tell 'intentional' html from
>something that just looks like HTML?  (and, what impact does that have
>on the software used to spit out PRs?)

Maybe the best option is to require HTML in PR's to be in a HTML comment
and that the filter removes one level of comments. Thus normal < can be
translated, but <-- < --> would become < .

Does that sound reasonable ?

>cgd

Marc B.