I was commenting based on the Boot Procedures man page on pg 5 of the pdf at:
https://www.bell-labs.com/usr/dmr/www/pdfs/man71.pdf
on dmr's "Unix Manual, first edition" page:
https://www.bell-labs.com/usr/dmr/www/1stEdman.html
which makes no reference to the rk disk, but does talk about loading up
initial files.
The page for init (pdf p.11) mentions that the usr directory is
mounted from the rk disk.
The rf0 page (pdf p3) at
https://www.bell-labs.com/usr/dmr/www/pdfs/man41.pdf
says
"writing is inherently very dangerous since a file system resides there"
The description of seek on the special files really slapped me in the
face for the first time!
The tty page in the man41 file ever so slightly advances my theory
that the interrupt character on the PDP-7 console (0176 or rather 0376)
could have been generated by the "ALT MODE" key on the console Teletype.
But regarding memory management, I foolishly put my foot in the gopher
hole of assumption that the earliest kernel on hand required the
mystery MMU, without verifying by reading the code, and forgetting
that "editions" refer to dates the manual was published, and that the
fragmentary fossil record that is the earliest surviving code (which
was in a constant state of evolution) doesn't align with the manuals.
Mea culpa, for adding any confusion!