Brian's Arrow System

Thu, 27 May 1999 12:28:06 -0700 (PDT)

On Wed, 26 May 1999, Maneesh Yadav wrote:

> <Tril, forige the negativness, but I sense a real lack of reality in the
> above>
> You have no idea what the components you mentioned are going to look like;
> yet you have the exact way they're are going to be integrated together?

Does anybody have a better idea?  I don't have THE answer, but I know what
I am going to do.  After I wrote that, I realized that the stuff I was
talking about may be completely unnecessary.  For this reason: If there
are two systems that are fully expressible, i.e. can express anything, you
don't need to find a higher model that includes both of them, because they
are equivalent.  If something is in one, it can exist natively in the
other.  What a fully expressible system means is uncertain.  Maybe Fare's
(newest idea for a)  paper will help. 

> Now why are you going to bother bootstrapping to a new system when there
> are plenty of existing (better) ones?  

Obviously they are not what we need.  If tunes already existed, we would
know it, because it would start integrating other projects by the handful.

> Your goign to tie all your code
> down into some un-advanced Retro API and then make the effort to write in

Not at all.  There is no advantage to coding at the retro level compared
to coding at the posix level or LISP VM level or Smalltalk VM level.  If
we run our HLL on retro it's just to be able to say we wrote the whole
system, not for any practical reason.  I don't care which lower layer is
used to bootstrap.  If my post suggested otherwise I apologize, that was
just one possible approach.

> You have not been at all definite in terms of your framework, as
> some of our conversations of IRC have indicated (that may have changed by
> now, but somehow I doubt that)...and I would be surprised if you really
> understood the Arrow system (I certainly don't, but I know what really
> understanding it (assuming it's worth anything at all) entails); how can
> you possilby map out a chronology for integration when you have no idea
> what these things are?

Brian is working so hard to understand his own system, why would I claim
to know it better than him?  I have given up trying to explain my ideas in
english and have started programming them.  I feel I am a much better
programmer than a writer.  However, I do see a lot of negative comments
from you.  How do these actually benefit the project?

David Manifold <>
This message is placed in the public domain.