2023-01-22 00:06:05 <KipIngram   > I don't really do much with it right now.  I just implemented it far enough along to prove it worked.
2023-01-22 00:06:32 <KipIngram   > I eventually plan to set up an easy system for having a ring of background tasks (which promise not to hold control too long).
2023-01-22 00:06:46 <KipIngram   > Just write a definition and poke it into that list, and it'll start getting time slices.
2023-01-22 00:10:46 <KipIngram   > Oh, and they'll al so promise not to change the stack, so I won't have to worry about any kind of "real" context switch (though I could, I guess).
2023-01-22 00:12:18 <KipIngram   > You know, in this system I've got now I used a register to mark the base of the system, but I also used one to mark the base of the header section.
2023-01-22 00:12:25 <KipIngram   > I'm  not going to do that again - that's wasteful.
2023-01-22 00:12:30 <KipIngram   > That can go in a variable.
2023-01-22 00:12:49 <KipIngram   > Anything mucking about with the headers isn't high performance code.
2023-01-22 00:13:20 <KipIngram   > I'm wondering now how I failed to see that before.  I guess at the time I had a whole bunch of free registers.
2023-01-22 00:13:46 <KipIngram   > And I guess it had an appealing "symmetry" aobout it.
2023-01-22 00:15:27 <KipIngram   > Also, this time I'm going to put docol immediately before next.
2023-01-22 00:15:40 <KipIngram   > Avoid the jmp next at the end.  Since a Forth system runs docol a LOT.
2023-01-22 00:15:54 <KipIngram   > Small benefit, but many times.