I see the EPUB 3.3 standard as being largely irrelevant from a practical point of view. It would be one thing if they were codifying the actual standards that a publisher needs to follow in order to produce a usable book. But the stakeholders who matter the most, the major e-book reading platforms (Kindle, Kobo, Nook, and Apple Books), are not part of the standard making process and have no real incentive participate. Unfortunately those players can dictate to publishers what they will accept and having requirements that do not quite match each other is an advantage if it makes it more costly for publishers to format books that work with their competitors.
The EPUB 3.3 specification explicitly defers to other standards for the definition of supported HTML and CSS. It even mentions that as those standards evolve existing EPUB documents may become unreadable. I don't think that is a good thing.
Instead of specifying things that a reading system might support in my opinion it would be better to focus on a core minimal subset that EPUB readers already support. Then get agreement from the major players that they will continue to do so in the future. That way we would have something that has the possibility of being used to publish books that will still be readable in twenty or fifty years. Something equivalent to
PDF/A which specifies a subset of PDF suitable for the long-term archival of documents.