Markdown's complexities reside almost entirely within its syntax, sure. There would have to be some kind of formal grammar, or a very, very deep suite of test cases. Without a formal grammar, you introduce risk through ambiguity. But, Markdown's scope is relatively limited, so that risk is unlikely to result in catastrophic failures (as long as you disregard that whole "embedded HTML" thing).
generating Markdown is tricky, but that's fine. In this scenario, Markdown would be hand-written and machine-read. There are plenty of use cases where Markdown wouldn't (and shouldn't) replace HTML, and "computer-generated web pages" is one of those use cases.
I think the biggest hitch would be metadata. A lot of Markdown parsers have support for a YAML header, but that behavior is heavily implementation-dependent (even relative to other markdown features). Coming to a consensus on this could get ugly.
generating Markdown is tricky, but that's fine. In this scenario, Markdown would be hand-written and machine-read. There are plenty of use cases where Markdown wouldn't (and shouldn't) replace HTML, and "computer-generated web pages" is one of those use cases.
I think the biggest hitch would be metadata. A lot of Markdown parsers have support for a YAML header, but that behavior is heavily implementation-dependent (even relative to other markdown features). Coming to a consensus on this could get ugly.