driver, prolog. Some typos.

[markus.gylling]

XHTML subset identity.
Is there an implicit or explicit reason for not providing a driver that formally defines the OPS XHTML 1.1 subset? See [1]. If this is not provided, how is conformance going to be assured? Such a driver is optimally something that, if existing, also should/could be referred from the NVDL schema.

Similarly, how does the prolog of an XHTML member of an OPS fileset look? There is a risk of authors providing the misnomer classic XHTML DOCTYPE, which will sometimes confusingly be DTD valid, since that is a superset of the OPS explicitly supported modules. Because of the namespace compoundness we have introduced, a classic XHTML doctype will sometimes also be concretely disvalidating. I think it should be spelled out that an OPS XHTML document does not carry a DOCTYPE; that its identity is inferred from the OPS context (as opposed to its prolog, as in the good old days), and that the namespace composition is formalized in the NVDL schema.

Typo. OPF Section 2.4, first example.
Wrong media-type="application/html+xml" />
Should be xhtml+xml

Typo. Section 2.2, example 2.
xmlns="http://www.idpf.og/2007/opf" Should be .org

OPS Section 2.4.2 DTBook Usage Requirements
"This DTD version, all previous versions, and all future versions will be available on the IDPF website."
Is this true? The canonical locus of these DTDs is http://www.daisy.org/z3986/2005/

[1] http://www.w3.org/TR/2004/WD-xhtml-modularization-20040218/dtd_developing.html#sec_E.4.