From: Alina Sternberg
Subject: INCOL - INcremental COmputation Language for incomplete data - incol1.txt [1/2]
Date:
Message-ID: <4kce47$2fa2@news.doit.wisc.edu>
BEGIN --- CUT HERE --- Cut Here --- cut here ---
Incremental Computation Language for calculation with incomplete data
Genrikh Babich, March 27, 1996
(for contacts use the following e-mail address:
···················@spool.cs.wisc.edu
Subject: for Genrikh Babich)
For 15 years in Moscow we conducted research in the field of calculations with
incomplete data. The result is a software realization of the mixed computation
principle formulated earlier by the Russian academician A.P.Ershov.
The central idea for calculation with incomplete data follows. The user
defines only the input data known by the time the calculation begins. The
defined input data set may be incomplete, i.e. some data necessary to reach
the final result may be missing. In addition to missing data the computational
algorithm may include calls to some not-yet-defined functions.
During the computational process all possible calculations are performed and
the so called incomplete result is presented in the form of a function of
unknown inputs. Some inputs may be omitted if they do not affect the output.
The incomplete result may also include references to the undefined functions.
The user is then prompted to provide some additional data, after which the
calculation is advanced to a more specific result. Then the man-machine
interaction repeats in cycles until the user is satisfied with the last
reached incomplete result which may happen to be complete.
The application of the general idea described above in the development of a
working software product was divided into three stages:
a) Creation of a specialized algorithmic language INCOL (INcremental
COmputation Language), semantically oriented to computations with incomplete
input data. (DONE)
b) Development of interpretation algorithms for the INCOL-programs and, on
the basis of these algorithms, -- development of a computational engine with
the ability to handle incomplete input data. (Implemented on a IBM/360-like
Russian computer; needs to be reworked)
c) Development of software products in the area of expert systems,
decision making systems, (what else?). These will benefit from use of the
INCOL-engine as a kernel surrounded by specific user interface shells, data
bases, and other utilities.
We would be glad to communicate with interested parties, to provide a
description of the INCOL-language and accompanying interpretation algorithms,
to discuss our accumulated experience and ways to promote these ideas which,
in our view, are quite promising.