Re: procedural variables

[sampson@cod.NOSC.MIL (Charles H. Sampson)]

In article <24590@neptune.inf.ethz.ch> mneerach@iiic.ethz.ch writes:
>
>Could anybody explain what the rationale for not including procedural
>variables in Ada is (I have heard it has been discussed and dismissed again
>in the past few years) ?

     I've been searching for some pertinent ancient documents, but without
success, so I'm going to have to rely on memory.  Probably everybody remem-
bers that Ada is the result of an effort by the U. S. Department of Defense.
This effort had a limited scope: to define a single computer language that
could be used for all embedded computer applications within the DoD.  (An
embedded computer is one that is a part of a larger system, such as an air-
craft or a smart torpedo.)  This was a well executed project.  It began with
a survey of many of the groups withing the DoD, to determine what capabili-
ties were required, followed by a set of successively refined requirements
documents.  As I remember it, one result of this survey was that there was
no need at all for procedural variables (in DoD embedded computer applica-
tions).  I also remember that the first requirements document (called Straw-
man; this is the one I can't find) explicitly stated that the language would
not have such a capability.  In the successors to Strawman the requirement
was stated positively in terms of the current three formal parameter modes.

     Incidentally, at one stage of this process the capability that ulti-
mately became exceptions was defined as a fourth formal parameter mode.

                                 Charlie