BISC Seminar, 10 April 1997, 4-5:00pm, 310 Soda Hall

Michael Lee (
Thu, 10 Apr 1997 09:29:21 +0200

This is a multi-part message in MIME format.

Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Michael A. Lee
Berkeley Initiative in Soft Computing
387 Soda Hall                                      Tel: +1-510-642-9827
Computer Science Division                          Fax: +1-510-642-5775
University of California                    Email:
Berkeley, CA 94720-1776 USA       WWW:

--------------20188CD1AEA Content-Type: text/plain; charset=us-ascii; name="guberman.4.10.97" Content-Transfer-Encoding: 7bit Content-Disposition: inline; filename="guberman.4.10.97"


BISC Seminar

Dr. Shelia Guberman Chief Scientist Paragraph Int.

Thursday, April 10, 1997 310 Soda Hall 4:00 p.m. - 5:00 p.m.


Soft computing does not mean the usage of words only as variables, but the use of soft natural text-like constructions. Many programmers note that the logic of the text's construction (i.e. the sequence of statements in the text) in a natural language is exactly opposite to those in computer languages. This difference is crucial : not the complexity of grammatical rules, the number of terms, nor the context dependency. The sequence of statements in a natural language is influenced by the need to assure understanding by another person. Thus the main requirement in creating text is to ensure the UNDERSTANDING of any statement at the time it appears. On the other hand the sequence of statements in a programming language is influenced by the needs of the computer: the focus is on the performance of the computer. The main requirement in program development is to ensure the properly EXECUTION of a given statement at the time it appears.

We find that there exist a small number of rules that can transform a sequence of statements written using the logic of natural language into a sequence of statements that will be correct for a programming language. Therefore, it is not necessary to duplicate the effort in order to develop an algorithm for the human style logic (to understand) and transform it into computer-style logic (to perform). One can use any programming language. Each line of the program will be legal in its language but the sequence of the statements will be constructed according human logic (ensuring the understanding by another person). Then the preprocessing will be done and the statements will be arranged in a different sequence. The resulting text is a legal program in that programming language and represents an algorithm and can be executed by the computer.

So, we propose a poor (as poor as a programming language is) but a natural-like language. (And of course this will help in understanding the programming work of others).