> > Remember it? If it's what I am thinking
of, I was doing it a couple
of
> > weeks ago. You are talking about a
'Coconut', right (explanation of
that
> > codename also on request - it has _nothing_
to do with the Tandy
CoCo).
> > Nowadays I do something similar. I create a
GROB with the right bit
> > patterns, use the SystemRPL 'Get' routine to remove the header, and
thus
> > create new objects.
No, not quite. The HP-41C used 2, 3 or 4 bytes to create the program steps.
By forcing apart the bytes and substituting new ones, new ("synthetic")
opcodes could be created.
These ranged from creating new characters to being able to access areas of
memory.
Ah, the heady days of discovery! Mother HP wouldn't officially help, but
there was plenty of behind-the-scenes help.