I disagree on epub3; Standard Ebooks is using it as a base format. While it certainly is a sometimes-strange mix of XHTML and HTML5, and there are things I'd personally have done differently, as a means of presenting static documents it rivals the power that HTML5 on desktop browsers has. (Because it's basically just HTML5!) Whatever formatting limitations there are are generally are on the client rendering side--for example, Kindles are notoriously bad at rendering tables, but that's because the Kindle renderer sucks, not because any of the competing standards don't have good <table> tag support.

Plus, since it's basically HTML in a zip file, anybody with basic web page production knowledge can jump in to producing an epub book. Sure there's some epub-specific cruft here or there but the core HTML markup has been well-understood for a decade.

Additionally, the power of HTML5 and epub's semantic inflection standard lets us mark books up in fascinating new ways that aren't easily possible in simpler formats. Whether or not that ends up going anywhere we have yet to see; another thing Standard Ebooks is trying to do is add semantic inflection to the books we produce.

Yes, it's not great for more complex stuff like interactive ebooks or javascript and so on, but no format is yet, and ereader technology isn't really there yet either.

Sorry, wasn't clear there: EPUB3 as a standard is theoretically OK, but real-world support of it on ebook readers just isn't there yet. And the update cycle is a lot slower than browsers or even mobile phones.

Also, do you know what is a great platform for "javascript and interactive stuff"? The web browser!