learning really fast is that lexicon resolution is... slow, validates my prior thoughts where i just don't think it's usable for any use case without some aggressive caching or unless someone builds a service serves all the necessary schemas in one request
RE: