PDP-11:

Apparently, DEC's V7m USG's System III totally puke on my kl11 driver.
RSTS/E, all the BSD's, and the Bostic distribution of V7 work just fine.
Thus, if you have this problem, you can displace the first tty-line of
my 4-line driver with the original Supnik driver by setting -DUSEKL 
in the DEFINES of the PDP-11 Makefile.  

The odd thing about all this is that the code is based the same, 
except for adding driver maintenance register functionality and a bit 
more hardware-like responsiveness to other registers.  Without these 
changes to the original KL code (when used for the extra lines), RSTS 
and 2.1BSD wouldn't see the 2nd to 4th lines as operational at all.  

My guess is that the V7m and S3 systems do some other kind of recognition
of the hardware lines.   I've searched, but have yet to figure out what
could be causing those OSes to bail on even the first tty line.

DL-11:

o For FreeBSD, I use popen(3) to connect to external programs (e.g. netline).
  On others, I have to use socketpair(2).  Socketpair is MUCH slower
  and tty thruput is quite noticeable.

o Due to the limited platform access available, no doubt a lot of
  portability is being lost in the PDP11 simulator.  However, I've been
  testing on every platform I can get my hands on, and have yet to find
  one I've caused a problem on.  See PLATFORMS for those which have 
  verification with my modifications.

VMS:

o The DL11 code doesn't work on VMS.  Make sure you use -DUSEKL to compile
  the original KL11 console code.  Someday I'll make an MMK file for VMS.

magtape support:

o Endian-independance was lost in my shuffles.