Thu, 31 Oct 1996 02:23:07 -0500
>I don't recall seeing any references to this standard in recent years. Nor
>can I find any evidence that it has been cancelled or (due to Ada95)
>declared obsolete. The question arises as part of the list of standards to
>be included (or excluded) for compliance with the DoD's "Single Process
I worked on both the CAIS-A (KIT/KITIA) & NATO PCIS projects, and I
can assure you that, as Tucker said ("dead"), MIL-STD-1838A is
irrelevant for this initiative. My guess (and this is just a guess)
is that 1838A has reached an age where is automatically lapses if
DoD does nothing to renew or revise it; surely nothing was done.
CAIS-A is dead dead dead irregardless of whether it's lapsed or not.
The more relevant question is whether PCTE is relevant. (PCTE =
"Portable Common Tools Environment," a misnomer because it is only a
specification for the framework/infrastructure "below" the tools of an
integrated project support environment, IPSE as the Europeans call
them, or [Integrated] Software Engineering Environment, [I]SEE as has
been more popular in the U.S. the past decade. Like CAIS-A, PCTE is a
glorified virtual OS w/ modern object base or object mgmt system very
appropriate for modeling project artifacts & activities & their
relationships -- except they're not very modern any more compared to
emerging OO stuff and they provide no real GUI support. The recent
PCTE revision (completed in '95?), as Tucker said, was supposed to add
some OO stuff but what was in first drafts wasn't much; the PCIS work
turned out to be only a remote influence on the PCTE revision.)
PCTE certainly meets the current Gov't/DoD objective of being a
"commercial standard" (there's nothing more commercial than the
European Computer Manufacturers' Association, ECMA, which is an
official standards-making body in Europe -- and just so happens to be
composed of most of the computer companies whose names we know and
love so well in the U.S.) But is it viable or does it fit the initiative?
I think not & not. My personal opinion is that PCTE too is dead
because the whole concept of monolithic tools-independent,
machine-independent SEE frameworks has withered. The tools vendor
industry has gone other directions (often proprietary, thin
services-based, expanding integrated suites of development
capabilities all from one vendor, and certainly not highly decoupled).
The weak support vendors gave PCTE years ago seems to be gone. Lots
of interesting research IPSEs were built on PCTE in Europe
approx. 1988-1993 and even a couple real projects used them, but I
don't hear about that any more.
I certainly think that any SEE framework standard was at best a remote
fit for the single-process initiative. Now I would bet my career that
it's no fit. I think that expenditure of time investigating this
further would detract from the real issues of the initiative.