Project 13: Communication languages

Oxford 1975:
============

The final topic discussed on Wednesday was communication languages,
and it was clear that there were many different opinions. Dr Battiste 
favoured amendment to the existing languages whilst Dr Lawson thought 
that languages such as BABEL were worthy of investigation. Dr 
Reinsch observed that a language from which both Algol 60 and FORTRAN 
code could be automatically generated would probably contain the 
disadvantages of both.	He cited the lack of recursion in FORTRAN and
of DATA statements in Algol 60.	Dr Pool suggested that they consider 
establishing a group, interested in numerical aspects of Algol 60, as a
long term project. Professor Dekker felt that Algol 68 would be a
better choice. It was agreed that Dr Lawson act as co-ordinator for 
the Group in the area of communication languages.

Amsterdam 1977:
===============

To continue in view of the interest of Professor Yanenko.

Baden 1978:
===========

Professor Gear said that the idea was for something above computer languages 
to describe algorithms precisely. He acknowledged that it would be hard to 
get agreement; any proposals needed to be accepted by TC-2. To compound the 
difficulties, details had to be agreed first, for instance [..] might demand 
that the enclosed expression be evaluated first.

Professor Dekker mentioned the virtues of Algol 60 and Algol 68. Dr.
Lawson told the meeting that his group had produced a structured
FORTRAN pre-processor and a lexical analyser for FORTRAN, written in
FORTRAN and Dr. Battiste drew attention to the work of Boyle.

Dr. Reid, as acting Chairman, said that Professor Hoare had asked for
a statement on the facilities required from language designers. He
asked whether Professor Gear could produce such a document. In reply
Professor Gear promised to produce something for the Novosibirsk
meeting.

Novosibirsk 1979:
=================

Gear introduced his ideas on communication languages (see IFIP/WG 2.5
(Novosibirsk-5) 605). The aim is not to design another computer
language, but to permit communication of algorithms in a form easily
understood by the recipient and easily edited to a computer language
(essentially using hand editing, but with a machine editor or word
processor to help).  It should include the possibility of natural
language sections.  Rice stated that TOMS encouraged communication of
this sort, but only one author had so far done it. Reinsch said that
Algol 60 had been designed also with this aim. Brown suggested lifting
the whole control mechanism of a computer language and embedding
natural languages within it. Rice felt that emphasis should be on
human communication rather than machine interaction. Gear promised to
prepare a new working paper for discussion at Harwell.

Harwell 1980:
=============

Gear explained his ideas on communication languages (see IFIP/WG 2.5
(Novosibirsk-5)605 and (Harwell-39)739).

Discussion

Curtis: I dislike the practice of including all documentation as
comments in the code. It leads to huge listings on every compilation
and is quite unnecessary in any software (e.g. Subroutine libraries)
for which there is a formal arrangement for associated documentation.