Define tags/profiles/slices of the language as separate TCK suites
AsciiDoc is a large language which makes it daunting to approach a compliant parser. We could "gamify" the process by having subsets of the language defined as their own suites.
This would provide more of a roadmap for people to follow in their implementations and help people avoid known gnarly parts of the specification until they have more of the basics in place. This could also help people communicate meaningful subsets of the language that they support.