debugging cables!

[Pandora B. Berman <CENT at MIT-AI>]

the massive (and long overdue -- my fault) debugging cable changeover
has finally (almost) finished. now the following machines are cabled
back-to-back:
1-18	2-8	3-5	15-16	19-20
to debug any of these machines, simply go to the other; the cables are
in place.

cadrs 4-6-9 share two cables; if any of them need debugging you must
check that the master/slave sense of the cable between them is correct.

cadrs 10-11-13-14-17 have three debugging cables among them. the eecs
lispm people (mostly jis and eric) are responsible for switching the
cables around to allow debugging of these machines.

cadrs in 936 have one debugging cable; these are cadr7 and whichever
cadrs are being debugged prior to release for service (currently cadr23).
they must also have the master/slave sense of the cable checked before
trying to debug one machine from the other.

cadr12 does not have the new heavy-duty connectors (we are out of them;
more are being ordered). it and the debugging urchin will get the new
connectors when they arrive.

the appearance and moving around of cadrs will modify this scheme, but in
general wherever possible cadrs will be cabled back-to-back to
prevent users form having to play with cables.