This file constitutes a "to-do" and current status list for someone seeking
to support the K processor/memory boards as a software development vehicle,
should this K/FALCON/PHOENIX project ever re-start.  The items of the list
are framed within the context of the progress and achievements of my last five
months' work at GigaMos Systems, and so might not be entirely self-evident
to an outsider.  Nonetheless, this is what there is ---

1. Synchronization flipflop needs to be added to the Asynchronous Trap
    input to the Trap Request (TRAPR) PAL.

2. Synchronization flipflop needs to be added to the D input to the
    LM-NUREQ flipflop (memory cycle request from Nubus slave interface).

3. Spy mode still needs to disable writes to the instruction caches.

4. The timing of the WRITE pulse to the register memory array in the
    processor needs to be tweaked better than it has been.  The pulse
    is frequently too skinny.  That etch should also be terminated to
    reduce undershoot and reflection.

5. Most of the critical pull-down resistors on the PC boards have been
    changed from 100 ohms to 22 ohms, but check the board status file,
    angel:/lmi/dgs/k-bd-stat.txt.  This file also reflects the most
    recently stored board statuses for those boards that have been in
    service.

6. The most recent engineering changes are discussed in text files in
    the directory DJ:DGS:*.text.