Hi,
My main reason for proposing two books is that many people may be more
likely to download or buy a shorter, less expensive, and less
formidable-seeming book.
I see the downloaders and the book buyers as being different targets,
probably with different habits.
Speaking about myself when I download documentation. I prefer to get
one large PDF containing all the info and being easily navigable with
links in the TOC and cross-references in the text, where appropriate.
I don't particularly want to download and open several docs when
researching the same subject.
My perception of the book buyer market is that they'd be likely to be
tempted either by very advanced reference works containing material
that might not get into the normal user's guide.
Alternatively, they'd perhaps be up for things like printed memory
joggers and "quick reference material" to have at hand or on a wall
nearby. You can probably image the kind of material I'm thinking of -
the kind of stuff it's really most convenient to have on paper rather
than on-screen.
Or perhaps tackle the real beginners market with kind-of "Dummy's
Guide" type books or text books suitable for children and students?
But, for me at least, it would fall into a niche apart from our
standard documentation such as we're working up at the moment.
Anyway, those were just a few additional reflections...
Of course, given that most people won't read a book anyway, I suppose it
doesn't really matter. 
Well, I'm not sure that's necessarily entirely true... as I said
above, perhaps it's more a question of identifying user bases who
prefer to work with printed material, and material that is more
meaningful, practical and useful when it's on paper, on your wall,
etc...
