Aucbvax.5495
net.applic
utzoo!decvax!ucbvax!baden
Wed Dec 16 19:21:43 1981
What are they for?
I have a running FP interpreter (modelled after Backus's Turing award
Lecture) written in Franz LISP. I have received an FFP
system from Purdue, also an FP interpeter from UCLA, both written
in C.
My interpreter contains no statistics gathering hooks as
do the other two systems; I intend to add these later.
What might these languages be used for? I don't see
FP or FFP proper as a higher level programming language; rather I
see it as a research tool. One idea that has interested me
is a 'Functional APL', the chief intent here is to permit
easy subdivision of problems (e.g. it might be easier to break an array in half
than a linked list in half), and to reduce the amount of indirection
(a related problem). We may of course be a little optimistic since
the subdivision of problems, and the resultant information migration
and coalescing is not an easy problem to solve.
I see a functional system as being embedded in a more traditional
one in the sense that state must be maintained, albeit much
less often then in say a non-applicative system. One way might be to
allow single assignments.
On another note I see an applicative architecture as well suited to
the VLSI design space: reduction of global state to help reduce the
dependence upon global interconnects. If we can the size of interconnects
down then we might get a faster running chip, and can increase the
amount of concurrency.
-----------------------------------------------------------------
gopher://quux.org/ conversion by John Goerzen <
[email protected]>
of
http://communication.ucsd.edu/A-News/
This Usenet Oldnews Archive
article may be copied and distributed freely, provided:
1. There is no money collected for the text(s) of the articles.
2. The following notice remains appended to each copy:
The Usenet Oldnews Archive: Compilation Copyright (C) 1981, 1996
Bruce Jones, Henry Spencer, David Wiseman.