I think all of you are missing a pretty big point here, once you have
a proper parser it doesn't matter what wikisyntax looks like by
default. You could make it spit out content from its internal parse
tree in many formats (DocBook, XHTML, POD, plaintext, LaTeX) and you
could, if you were willing to accept losing some formatting changes
(although that need not be the case) allow users to write text in any
of these formats, then parse it and save some sort of glorified parse
tree in the db.