I constantly learn from this listserv, but almost never comment because I'm no authority on the xml stack, just a devoted user and serious student of it. But a parasitic "schema" thread has emerged and I speak up (mouth off?) because I feel so strongly about this.
There is a well-intentioned cadre out there (never on this listserv!) who are not looking for a simpler xml, they are looking for a simpler domain modeling language. They mistakenly think that domain modeling is an appropriate use for an xml schema language.
It is not. (The conflation is nourished egregiously by the syntactic and conceptual centrality of xsd:complexType in the xsd spec.)
An xml schema language is for document modeling, not for domain modeling.
For document modeling (in any xml, including "micro"), we have rng (dsdl). If anybody thinks rng (dsdl) is conceptually too complex, they might reconsider working with xml.
For a different approach to expressing a document model: sure, something like examplotron.
And doubtless we need richer, more concise and more systematic ways of asking for various kinds of "lax" (less "draconian") validation.
But in any case, the schema discussion has nothing to do with microxml per se, and needs to be completely separated from it.