Lambda (was: Refactoring in Tunes)
Thu, 13 Jan 2000 21:13:12 +0100
John Carter wrote:
> On Tue, 11 Jan 2000, Massimo Dentico wrote:
> > ... and what about rewriting (probably only a core of) Joy in assembly
> > with the usual technique of Forth? A merging with Retro/Forth OS of Tom
> > Novelli? What about the utility of such project as an infrastructure
> > for the Tunes project? A tiny, consistent, understandable, fast environment
> > with strong mathematical basis to experiment new ideas (or old ideas with
> > new point of views). It's only a very modest proposal ...
> Careful, as far as I can see Joy has one good idea and that is superb,
> but it is not an idea that will solve all the problems Tunes is
Certainly. Originally I have written:
"The idea of a system based on combinators (or Forth), as a
bootstrapping system, float around the Tunes project from the long
time (see LLL subproject)."
See here what is the intended use of an LLL (Low Level Language)
"The TUNES LLL Subproject" - http://www.tunes.org/LLL/index.html
> The heart and joy of the Joy idea, (correct me if I'm wrong), is that
> by creating a simple homomorphism between the syntatic operation of
> putting one symbol next to another and the function composition you
> create something which is trivial to create an extensive algebra on.
As a side note this abstraction maps directly to machine language also:
a composition of N instructions is given by putting one instruction
next to another in memory. In your opinion, what is the consequences of
this? (or .. please can you elaborate on "which is trivial to create an
extensive algebra on."? I'm interested.)
> The rest of the semantics of Joy is fairly stinky. (eg. only data
> types are integers, strings, 32 bit sets and lists)
> Thus I'm truly am enamoured with what I call the joy of Joy idea, the
> semantics needs a total rethink.
My suggestion is not to use Joy *as is*, but to explore the possibility
of Joy as a starting point for an LLL. However, I'm still reading the