[net.lang.prolog] POPLOG

SHardy%SRI-KL@sri-unix.UUCP (11/01/83)

A recent message said that programming environments offering more
than one language, such as POPLOG or LM-PROLOG, are no solution to
difficulties in using pure Prolog, such as wanting to use ASSERT or
RPLACA.  The message said such systems in fact make things worse
since they make it necessary to learn two or more programming
languages.

I can't speak for the LM-PROLOG people, but as a one time member
of the POPLOG team, I can say that our goals for POPLOG were user
convenience and efficiency - not finding solutions to the theoretical
problems of mixing meta level knowledge with base level knowledge
(E.g. using ASSERT).

We wanted, for example, to interact with our programs via a user
modifiable screen editor.  Writing a screen editor in pure Prolog is
impossible.  We wanted to process arrays of floating point numbers.
We wanted to be able to use HEARSAY-like agenda mechanisms.  We
wanted to be able to write efficient compilers containing a minimum of
machine code.  We wanted to do lots of things easier in LISP or POP-11
than Prolog.

We felt that professional programmers could be fluent in several
languages (E.g. Prolog, POP or LISP, C or FORTRAN, and assembler) and
would appreciate the choice of appropriate level for any particular
module of a large system.

-- Steve Hardy,
   Teknowledge