Re: Minimal keywords needed for constructing a full CL system

"Rob Thorpe" <robert.thorpe@xxxxxxxxxxxx> wrote on 5 Jul 2006 10:23:
It is not really the minimum set since it does no I/O, as I mentioned
earlier. So unless your prepared to speculate about implementing the whole
universe using lambda

Why not implement the "whole universe" using lambda?

Of course it would be silly in practice, but so is the original question.
If you don't want "just LAMBDA" to be the answer, then you have to ask a
more precise question.

I think what the OP is looking for is the smallest set that meets some very
limited criteria of practicality. That was the sense I answered the
question in anyway.

Yes, of course, but what we're doing here is a kind of Socratic method of
uncovering the answer. We're demonstrating that the original question isn't
a very good one, and suggesting that maybe even the OP's original motivations
are misplaced.

The truth, as many people have already said, is that there is no single unique
canonical "minimal" set of lisp "primitives".

There is the whole spec. It would be possible to implement each function
there directly in some other language (assembly, Fortran, etc.). It's also
possible to implement some small subset of them in another language (perhaps
C), and then the rest in Lisp itself. Pretty much all CL implementations do
this latter thing.

But it is somewhat arbitrary which operations you choose to put in the subset,
and which you choose to implement in Lisp. It is arbitrary for two reasons:

1. There are different possible "minimal" sets, rather than a single
canonical one. Much like you can implement any logical operation either
just using NAND gates, or else just using AND/OR/NOT gates. Either choice
is a fine basis for building up all logical operations.

2. Unlike this theoretical exercise, in practice you need to worry about
efficiency. So an implementation might "primitively" implement many different
functions, even if it is theoretically possible for some of them to be
defined in terms of others.

Conclusion? The theoretical question "what is the minimal subset of Lisp" is
either silly or malformed or useless. "Just LAMBDA" is one possible answer,
likely of only minor interest. It's actually hard to ask a precise question
that gives the kind of answer the OP thought he was looking for.

But what he thinks he's looking for isn't as interesting as he believes it is.

-- Don
_______________________________________________________________________________
Don Geddis http://don.geddis.org/ don@xxxxxxxxxx
Instead of trying to build newer and bigger weapons of destruction, mankind
should be thinking about getting more use out of the weapons we already have.
-- Deep Thoughts, by Jack Handey
.

Relevant Pages

  • Re: Minimal keywords needed for constructing a full CL system
    ... Why not implement the "whole universe" using lambda? ... discussing abstractions like minimal sets etc its useful to try to pick ... Much like you can implement any logical operation either ...
    (comp.lang.lisp)
  • Re: The 13.7 billion light-year length denotes a fifth spatial dimension.
    ... WMAP's March 2006 polarity data is consistent with an ever constant lambda, ... The observable universe was Planck length 13.7 billion light-years ... The Inflation Theory, developed by Alan Guth, Andrei Linde, Paul ... such as symmetry breaking and phase transitions, to cosmology. ...
    (sci.physics)
  • Re: Difference between let, let* and letrec
    ... > "We have an anonymous procedure (the first lambda) having two pointers ... Now it seems that those two pointers are not ... If you don't want to bind a determined value, ... IMPORTANT NOTICE TO PURCHASERS: The entire physical universe, ...
    (comp.lang.scheme)
  • Re: quote semantics
    ... with some useful work on Lisp libraries, if you fall that way GTF ... ask for that one, too, whether it was all that related to the Lambda ... And certainly CL took influence from Scheme, ... claim that this community, which is entirely composed of people who, ...
    (comp.lang.lisp)
  • Re: reading file to list
    ... Common Lisp is the other way around. ... that many old emacs users, which includes significant portion of emacs ... 2 fundamental problems of lisp i feel this ruby ... The way the above works is that each of f, g, h is a lambda themselves ...
    (comp.lang.python)