Ok... I'll go on.
Just tell me if I get out of line.
Firstly, I sympathise with the issue you bring up about preservation. However I am not certain it is as dire as you suggest it. The fact that you or I cannot in under a minute recreate a gorgeous PDF in HTML or whatever else doesn't really have any worse implications than the same going for printed books.
If the eBook needs to be recreated (and, keep in mind, by those days conversion galore ought no longer be the only way to enjoy a variety of books--only exceptional circumstance ought to require it) in a worst case scenario it can always be retyped or OCR'd and corrected and then typeset anew.
Secondly, like both Sonist and I stress now and then: the rigidity is a feature, not a bug. The same document cannot work on a myriad devices with nary a thing in common without considerable degradation (or, my notion of the same PDF holding multiple versions for different anticipated screen sizes and font sizes). HTML works and is good enough... but without a typesetter it can never move beyond good enough with anything typographically complex. And, as noted in my last message, it hasn't in over a decade of websites.
Third, I don't necessarily think that currently copyrighted materials (so long as they are attentively typeset for a reasonable variety of devices) ought to be readily extractable. Piracy happens in no small part due to its ease and convenience. DRM is trivial to disable in comparison to the task of 100% validly/correctly converting PDFs. If publishers do their part for the consumer, PDFs are the ideal format. Assuming you can get over the irritation of not having control over your books that you never head prior to your eBook device, and are willing to instead just sit back and enjoy your beautiul professional quality eBook--truly the equal of your neighbour's heavy and dusty hardcover.