and W3C XML Schema's approach has issues

Extension A wide variety of limitations on how to combine content models if you want a new type to have additional content, varying with the nature of the declaration of the original type. Also not symmetrical with restriction.
Restriction Really a declaration of intent, not a formal process. Derivation by restriction often produces content models that are more verbose than the original they override. Also has conflict between different notions of restriction in structures vs. restriction in datatypes, though terms and vocabulary overlap.
block/abstract/final W3C XML Schema purports to let you control how types are used with these keywords, but they've proven difficult, even mind-numbing in the case of block, as they apply both to type extension and restriction and to substitution groups.
Type identification in documents In addition to interacting with the rest of the schema, these features need to interact with xsi:type attributes in the instance documents. As the same element may be of different types in different contexts depending on how they were declared, this can get remarkably complicated.