RE: [Axiom-developer] B#

[Bill Page]

On March 22, 2006 10:12 PM Tim Daly (root) wrote:
> 
> ok, lets get the confusion out of the way before this gets 
> too far along...

:o)

> 
> The "stack" of languages looks like:
> 
>  spad -- an algebra language using types 
>    spad compiler -- defines the algebra language and types
>      boot -- a non-parenthesized common lisp which is untyped
>        common lisp -- a parenthesized common lisp which is untyped
> 

Right.

> in the interpreter
> 
>  interpreter language -- spad-like, uses types
>    boot -- a non-parenthesized common lisp which is untyped
>      common lisp -- a parenthesized common lisp which is untyped
>

I think this design is a historical accident and not the way
it should/would be done today. Historically both SPAD and the
Axiom interpreter were written at the same time. But if we
were to do it again today the "stack" would probably look like
this:

  interpreter language -- spad-like, uses types
    spad compiler -- defines the algebra language and types
      boot -- a non-parenthesized common lisp which is untyped
        common lisp -- a parenthesized common lisp which is untyped

And we would replace SPAD with Aldor if/when it is legally
possible.
 
> the B-natural language sits here:
> 
>  b-natural -- single-type langauge
>    interpreter language -- spad-like, uses types
>      boot -- a non-parenthesized common lisp which is untyped
>        common lisp -- a parenthesized common lisp which is untyped
>

I think it should sit here (at the same level as the existing
Axiom interpreter):

  b-natural -- single-type langauge
    spad compiler -- defines the algebra language and types
      boot -- a non-parenthesized common lisp which is untyped
        common lisp -- a parenthesized common lisp which is untyped

B# is an alternative to the Axiom interpreter, not an extension.

Or (in the best of all possible worlds :):

  b-natural -- single-type langauge
    Aldor compiler -- defines the algebra language and types
       common lisp -- a parenthesized common lisp which is untyped

If we get rid of SPAD and the old Axiom interpreter then BOOT
becomes irrelevant. But Common Lisp as the base of this "stack"
still makes good sense to me.

> 
> Note that Bill and I disagree about boot's future.

True.

> 
> Note that B-natural and Boot are COMPLETELY unrelated.
> 

I tend to agree.

Regards,
Bill Page.