Re: Mike Feldman, meet Archie

[cis.ohio-state.edu!zaphod.mps.ohio-state.edu!usc!sol.ctr.columbia.edu!eff !world!srctran@ucbvax.Berkeley.EDU (Gregory Aharonian)]

    As many have pointed out, and as I pointed out in my original posting,
the Archie search I performed was not very thorough.  I did not intend it to
be, as it was more of a lark prompted by Mike's periodic posting of Ada
projects.
    As someone pointed out, searches for "C++" strings don't return much,
even though there is over ten gigabytes of C/C++ code scattered across the
ftp-sites of the world.  Though, given the logic of assuming a search of
"ada" or "Ada" is what most people would do using the various Internet
searching protocols, you would figure that people with Ada codes available
across the nets would take the initiative and rename their files to be
more obviously identifiable.  If Verdix uses ".a", why can't such sites
use a file-name-redirect to create a variant filename that points to the
original file.  It's these little initiatives that separate real software
reuse practioners from the amatuers at places like ASSET and DSRO, that is,
trying to anticipate and support your customers (don't you hate that word)
needs.

    The point I weakly tried to make with my posting, and more strongly in
other postings, is that the Ada world has no validated measure of Ada
acceptance in the non-Mandated world, that is, no one knows the actual
market share currently, or during the past ten years.  No one knows if
Ada as a new language is being more widely accepted than other new
languages, such as C++ (probably), Smalltalk (maybe), Forth (doubtful) 
or Prolog (doubtful), if it is gaining or losing compared to these
languages.

    The stick-our-heads-in-the-ground attitude, as epitomized in the
million dollar waste of the Mosemann studies, is that the Ada community
has no interest in the demographics of Ada in the non-Mandated world, and
as the GAO pointed out and the DoD not addressed, even in the Mandated world,
the data collected is incomplete and sparse.

    There is nothing more silly than the whole Ada-9X project, for the
simple reason that Ada83 is a pretty good language already, whose major
problems was that no one (DoD, compiler vendors, contractors) cared to
really evangelize for the language and help it win market share.  Instead
of focussing on that rather complex socioeconomic problem, the DoD decided
to focus on a technical problem, that of upgrading Ada83.  Fine, I guess,
but the language is going to be even less accepted than Ada83 because some
of the windows of opportunity open to Ada83 have been shut by C++ and
Smalltalk (partially thanks to IBM).

    Now if people in the DoD were serious about making Ada an accepted
language outside of the Mandated world, in the formal business sense of
launching a product, the first business thing to do would be to do a
comprehensive survey of Ada use in the non-Mandated world, for example,
by doing methodical Archie and WAIS surveys, contacting software
development companies, scanning thousands of technical reports with
source code listings, searches of universities theses, commercial CDROMS
with source code, counting job requirements in help wanted ads for languages
listed, and other areas to find out usage trends for programming
languages for the past ten years to see how well Ada is doing, where it
is doing well, and what are the perceptions of Ada in the non-Mandated
world.

    Such information is crucial for Ada83 and Ada9X survival, yet no one
seems to care, and the one chance to answer such questions, the Mosemann
studies, were flawed beyond conception.  All anyone ever hears about Ada
in the non-Mandated world are ancedotal information, like Mike's posting.
There are no hard facts, more specifically, no systematic collection of
information that would satisfy someone interested in investing in an Ada
venture.

    I would love to do such a study, if for nothing else than to give the
rest of you a chance to flame me for something I have done at taxpayer
expense (and if anyone knows a DoD office I can contact, let me know).

   In any event, my Archie search, Mike's posting, and general knowledge
about Ada usage, are all pretty much equally twentieth-assed, even though
many DoD decisions are made assuming such data exists.


Gregory Aharonian
Source Translation & Optimization
-- 
**************************************************************************
Greg Aharonian
Source Translation & Optimiztion
P.O. Box 404, Belmont, MA 02178