Re Keystrokes: Possible BIOS compatibility

["Patricia M. Godfrey" PriscaMG@xxxxxxxx]

OK, I ran Robert's very helpful debugging version. It went through MV,
then displayed on the CommandLine various letter combos that, I think,
are not functions (N0, N1,...N9, N@, NA,..NE). Then it displayed on the
CL, and wrote to the file NF. NG and NH were displayed, not written, NI
was written, NJ and NK displayed, then NL displayed, and up popped a
message: Program accessed memory currently in use at 00003338 from
0A4C:24E1.
I shut down XyWrite, reopened it, them added NL to the list of items to
be omitted, saved, and ran again. This time it got through RQ--whereupon
the PC spontaneously rebooted. (I see that RQ is not a function call, so
perhaps it was RP; by then eveything had gone so swimmingly that I
mayn't have been looking as closely as I should have been.)
After sitting through scandisk of four drives (and that maddening M$
message telling one to shut down properly--as if one didn't know) I
decided to test something else. I took the original routine, cut it out
to a separate program, and limited the first SG to NQ. Sure enough, it
crashed at NI (Your program accessed memory currently in use at 000047F6
from 2020:302E). So Robert is right (of course) and it's some specific
calls, not a general memory overflow, as I had been suspecting.
Will try again, but it takes forever when it doesn't crash (I have to
set the wait to 4 or 5 seconds or I cannot jot things down and then look
back at the screen), so will send off this preliminary report.

Robert, I'm sending you the keyboard files off list.

Patricia M. Godfrey