Project 26: Review of Ada

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

Drs. Brown and Smith would review the ADA language.

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

Lawson reported that Whitaker had been succeeded by his deputy, Fisher.
Field testing to take place this summer. Tutorials to be held, then
trial applications in the language. There is some kind of language
processor available now. A production compiler should be available
next year. Reid mentioned a problem with generic functions. The actual
function used depends on both the arguments and the result.
Straightforward for a simple assignment but several passes over a
complicated expression may be needed to determine what actual
functions are needed. Dekker remarked that the designers of Algol 68
had chosen not to permit such generic functions. Ford said that he
understood that the fixed-point proposals had been withdrawn. Smith
remarked that he thought using throughout an expression the highest
precision mentioned in it was a good rule and would avoid problems
with generics.

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

Smith recommended the book "Programming with Ada" by Wegner, published
by Prentice-Hall. It is not precise or totally reliable but is easy to
read and is a great help towards reading the reference manual (Sigplan
notices, June 79). Reid asked how the group should act following
Wichmann's talk. Fosdick remarked that the spirit of the Brown model
had been lost because thresholds were not checked properly. Reid
remarked that Wichmann is aware of this difficulty.  Lawson was
unhappy over the use of base 2 everywhere. Hull has asked Wichmann
what the group can do and he said he would like help with handling
underflow. Rice sought to reject the criticism that the numerical
community had failed to react; the organizers had hired many language
people to work on the project but no numerical analysts. It was agreed
that a subcommittee consisting of Reid (chairman), Smith and Lawson
should consider the numerical - features in Ada and communicate with
Wichmann.

Argonne 1986:
=============

Document: IFIP/WG2.5 (Argonne-26) 1326, 4 pages.

D. Dekker presented the status of the work on the "Pilot Implementation of base 
modules for large portable numerical libraries in Ada" (PIA). Participants in the 
project are Numerical Algorithms Group (U.K.), National Physical Laboratory (U.K.), 
Centrum voor Wiskunde end Informatica (The Netherlands), National Institute for 
Higher Education (Ireland), University of Liverpool (U.K.).

Discussion followed. The information about the PIA project was to be presented at 
several meetings (SigAda included), and a book on Scientific Ada is in preparation  
(Ford).

Standardization and accuracy of elementary functions in Ada was discussed. B. Ford 
offered to share the specifications he had with C. Lawson and G. Paul (Ford, 
Lawson, Dekker, Reid, Paul).