I should clarify that I use the rescue disk repeatedly
and go through the initialize-swap and mount-partitions
steps each time, so the list of mounted volumes in
my previous message is accurate.

Also, I tried doing this at the boot prompt:
  linux root=/dev/hda2

... which refers to my root partition.  That partition
in turn contains /boot as a mount-point, but /dev/hda1
would not be mounted at that point so I guess that is
bound to fail.

--

It might help if someone could explain some of the mechanics
of this process.  A lot of us Linux newbies would
benefit.  For example:

* What program actually copies the boot code into the MBR?
  Is it lilo that does this?

* What goes into the MBR?  Just a jump into lilo code
  residing elsewhere in the boot partition?
  Maybe you can point to a discussion somewhere of
  how the MBR and boot partition are are laid out?

* Back to the floppy attempt, what program actually does
  the "make bzdisk"?  Is this something I could hack
  by hand?

* When booting from a floppy onto the hard disk, does
  the floppy just contain an MBR pointing onto the hard
  disk, or does it have to first execute code residing
  on the floppy?  If the latter, then what code would
  that be, would it be a kernel, or lilo, or something
  else?

* I made a copy of the Debian rescue disk so that I could
  mount it and hack in some changes to bootinto the hard
  disk environment.  I see files that apparently control
  the process, but the whole install mechanism is in
  there and it muddies the water.  Can I get an explanation
  of how a boot-into-hard-disk would operate at the
  grubby mechanical level?

I have an operational RedHat system, dual-booted against my
Windows environment on a different box.  I've been going there
to download files and create my install disks.  I could do
various kinds of "make" there, but I would need the results
to be Debian-compatible.