Re: [stack] Re: Flat concatenative basis (was: Concatenative macros?)

Robbert van Dalen <[email protected]> wrote:
> Great stuff!
> Still, I don't see the need for a flat binary base (except for
> extreme simplicity and fun).

That's a great reason :-). I'm using it to better understand perfect
flatness, in hopes that I can see whether flatness has any desirable
properties. It's also accidentally making it possible for me to
understand combinators, which I now see that I didn't fully "get".

> Why not allow 4 flat primitives that follow the 4 bases (ACGT) of
> DNA? :)
> I'm not sure but that may lead to smaller programs.

It certainly would; but it would be just that much less flat (you
could only 'cut' programs at every other bit address). Long-term, I'm
likely going to make a 256-base code, and claim that it's flat at the
8-bit byte level. But before I think about that I have to understand
this much simpler system -- 256 bases is a lot to have to choose when
one doesn't know why 2 bases are efficient.

> I also wonder how one can do arithmetics on top of {o,k} without
> resorting to Church numerals.

That's essentially correct, although actually there are other
conventions that will work as well as Church numbers do. As you know,
Enchilada uses one of them (although of course E can see into its
quotations). I haven't bothered choosing one, because I've got so much
else to do.

> > Also, Enchilada's tree data structure might be extremely useful for
> > this.

> If the flat base {o,k} settles to a definite version I would be very
> happy to implement it on top of Enchilada's tree structure.
> (or provide you guys with the algorithms).

Thank you!

> Robbert.

-Billy