RE: Another Idea for Ada 20XX

["Robert C. Leif, Ph.D." <rleif@rleif.com>]

From: Bob Leif
To: James Rogers et al.
A package to do this, A Broad-Based Environment for Test (ABBET), was
created by William A Whitaker, Ph.D. There is also a collection of packages
by Dmitry A. Kazakov.

I believe that it might be possible to have either the compiler optimize
away the unit checking for objects where the units are constant and/or
create a class of assertions that are only checked at compile time. Fresh
ideas on this subject should have a high priority for Ada 200X.

-----Original Message-----
From: comp.lang.ada-admin@ada.eu.org
[mailto:comp.lang.ada-admin@ada.eu.org]On Behalf Of James Rogers
Sent: Sunday, December 02, 2001 8:01 AM
To: comp.lang.ada@ada.eu.org
Subject: Another Idea for Ada 20XX


In scientific terms a measurement is done in terms of some unit.
For instance, distance is in meters, microns, feet or furlongs.
Speed is then calculated as a distance per unit time. Acceleration
is the second derivative of speed.

I would like to have this notion of measurement units and their
relationships supported in a language. This would require the language
to perform unit analysis while performing static or run-time
calculations. Because of the added computational overhead, there must
be a way to choose or reject unit analysis for defined types.

Perhaps a new keyword for unit-analyzed types could be introduced,
along the lines of the keyword "tagged" added to Ada 95. A unit-
analyzed type would carry its own unit description tag to
facilitate unit analysis. Any numeric type without such a tag
would not be subject to unit analysis.

Possible declarative syntax might be:

type Meters is unit range (0.0..1.0E380);
type Hours is unit range (0.0..9.99E30);
type Speed is unit(Meters / Hours);

The advantage for the software developer is in the usage:

Vehicle_Speed : Speed;
Track_Distance : Meters;
Lap_Time : Hours;

Speed := Track_Distance / Lap_Time;

This approach would allow the appropriate mixing of data types, which
is currently forbidden by Ada, in a meaningful and correct manner.

I know this proposal places a heavy burden upon compiler writers.
I also suspect it would be very useful for ensuring the correctness
of programs, particularly in embedded systems.

Note that a lot of the unit analysis could be performed statically.
This would allow minimal performance penalty at run time.

Jim Rogers
Colorado Springs, Colorado USA