I think your other topic is closely related but I chose this one to reply to.

I own something well north of 10,000 technical and engineering books
so I will appoint myself as an amateur librarian.

When I was younger, I had the false notion that anything new is good.
This attitude permates a lot of society.  Including professional
libraries.  They have a lot of collection management practices around
deciding what and when to pitch something and a big one is whether the
work is still in print, while a more sophisticated collection will
also take into account circulation numbers (how often it is checked
out).  A lot of that is undoubtedly the real costs surrounding storing
and displaying something (an archived book has a marginal cost, a
publically accessible displayed book presumably has a higher
associated cost) as well as the desire to remain current and provide
value to the library's membership.

From what I have seen, there isn't much notion of retaining or
promoting a particular work unless it remains in print.  As an
example, K&R C is still in print and would be retained by most
libraries.  The whole thing becomes a bit ouroboros because that leads
to more copies being printed, and it remaining in collections, and
being read.  Obviously, this is a case of a great piece of work
benefiting from the whole ordeal.  But for more niche topics, that
kind of feedback loop doesn't happen.  So the whole thing comes down
in a house of cards... the publisher guesses how many books to print,
a glut of them are produced, they enter circulation, and then it goes
out of print in a few years.  A few years later it is purged from the
public libraries.  As an end user, one benefit to this collapse is
that used books are basically flooded into the market and you can get
many books for a fraction of their retail price used.. but it becomes
difficult to know _what_ to get if you don't have an expert guide or
somewhere to browse and select for yourself.

So why does this all matter to your more meta question of why less
great books?  There is less to no money in it nowadays for authors.
The above example of library circulation represented a large number of
guaranteed sales to wealthy institutions (academic and government =
wealth, don't let them pretend otherwise).  Except now many libraries
have downsized their physical collections to make room for multimedia
or just lower density use of space.  So there are less guaranteed
sales.

Another facet of the same coin, one reason printed books are great has
to do with the team surrounding their production.  If you look near
the colophon, you will often find a textbook will have quite a few
people involved in moving a manuscript to production.  This obviously
costs a lot of money.  As things move more to ebook and print on
demand, it's an obvious place to cut publishing expenses and throw all
the work directly onto the author.  That may result in cheaper books
and maybe(?) more revenue for the author, but it won't have the same
quality that a professional publishing team can bring to the table.

As to my deliberate decision to accumulate the dead trees and ink,
it's because although online docs are great I find my best learning is
offline while I use the online docs more like mental jogs for a
particular API or refamiliarizing myself with the problem domain.  I
have some grandeur ambitions that first involve a large scanning
project but that will have to await more self funding.

Regards,
Kevin