NetNews Usenet Archive 1992 #27/NN_1992_27.iso/spool/comp/parallel/2593

Text File | 1992-11-23 | 6.8 KB | 129 lines

Newsgroups: comp.parallel
Path: sparky!uunet!destroyer!gatech!hubcap!fpst
From: [e-mail] (richard marlon stein)
Subject: observations of Supercomputing '92
Message-ID: <[e-mail]>
Apparently-To: [e-mail]
Sender: [e-mail]
Organization: Whole Earth 'Lectronic Link
Distribution: usa
Date: Mon, 23 Nov 1992 17:27:21 GMT
Approved: [e-mail]
Lines: 115

[This is a somewhat charged statement. I'm letting it by since it is
 an area of concern for eveyone. Steve ]


(C) Copyright 1992 Richard Marlon Stein
                        The Silicon Ceiling
                       Richard Marlon Stein
The Minneapolis Convention Center hosted Supercomputing '92 during
Nov. 16-20, a showplace for the prosperous paradoxical veneer of
masculinity that underlies the industry.  Where else could one find
the world's fastest computers, the phallic symbols of computation,
poised for display like steroid-popping body builders?  But these
macho edifices conceal a catastrophic harbinger portending the
breed's irreversible demise.
     Supercomputer manufacturers have enjoyed lavish nourishment in
the form of government subsidies.  This protection permits the
technology to evolve and procreate asexually, much to its
disadvantage.  Where economic competition is most needed,
protectionist policies shield this industry by subsidizing a few
chosen companies to almost exclusively supply the federal
bureaucracy, despite the availability of more suitable solutions
from MasPar, Convex, IBM, Fujitsu, Meiko, and others.
     The firms who do not receive federal funding aggressively
compete for the remaining market share, estimated to become US $2B
by 1995.  Yet the entire domestic industry is vulnerable from an
overdependence on foreign semiconductor suppliers not to mention
the government's failure to promote education among the aerospace,
biotechnology, and automotive industries who depend on
supercomputer-based simulations for competitive advantage.  The
government has failed to train software engineers and to transfer
the programming expertise located in national laboratories into the
public sector.
     This egregious oversight raises the cost of supercomputer
software, the codes which run vital solutions for environmental and
industrial use.  The government did not stimulate demand for
supercomputers by launching an equally aggressive policy to train
users and programmers. Furthermore, the entire High Performance
Computing and Communications Initiative does not specifically
sponsor the involvement of women and minorities who should be given
the opportunity to contribute.
     This human capital vacuum obstructs innovation, a pivotal
skill that all people possess.  The preponderance of white males in
key management areas is analogous to the glass ceiling encountered
by women and minorities.  But the ceiling that restricts the
mobility of women and minorities in the supercomputer industry
possesses an opaque surface.  It is a silicon ceiling.  One cannot
see through this monolithic barrier.
     The ceiling's composition resists conditioned feminine
attempts or equal opportunity approaches to entry.  The silicon
ceiling cannot be smashed with hammer blows, but must wait to be
etched away by a visionary rinsing of political, social, and
educational mores.
     For years, our government has failed to successfully motivate
a proportionate number of women and minorities to participate in
engineering and physical science education.  This cultural and
political failure accrues to our disadvantage in critical areas of
technology where diverse backgrounds and alternative opinions
fortuitously combine into innovative products.  Only a reformation
in the education system and social mores of society can reverse
this trend.
     The silicon ceiling conceals more than a gender barrier.  But
the macho images of Supercomputing '92 reflect the ceiling's
impermeability.  No matter where you looked in the Minneapolis
Convention Center, nearly indistinguishable supercomputers faced
each other like rutting elephants trumpeting their egos, not the
plethora of software solutions they so vitally require to survive.
     Women did attend the conference, though they were represented
as market and public relations personnel, ostensibly there to
defend the mediocre products built by all-male engineering
organizations.  Out of some 40 vendors, IBM and HP were two
noticeable exceptions.  Each presented females who served in
engineering management.
     But the `old white boys' dominate the technology hierarchy.
This anachronistic clan precludes the introduction of social
change, and facilitates an industrial complacency once found in the
consumer electronics and automobile industries.
     Although the supercomputer industry is pivotal to our economic
future, it resembles a sand pile build from hollow bravado and
poorly articulated government policy.  The technological pile grows
uniformly, expanding outward and upward, as material ego, special
interests, and stop-gap political solutions are added in lieu of
cogent industrial and economic policies.  But if one more ill-
conceived policy is dropped on top, a catastrophic cascade will
start reducing the technology citadel to a retrenched and
aesthetically weakened hill owned by foreign interests.
     The economic camel's straw of vulnerability can be forestalled
if an egalitarian glue is applied to the technological sandpile.
The formula for this glue is not sugar and spice.  It is an
adhesive agglomeration that coagulates political leadership into an
accountable body free from political action committees and attunes
them to the strategic issues that govern our collective economic
and technological welfare.  Our leadership must articulate a policy
which promotes domestic industry and competition, but also provides
incentives to liberate industries from this dependence.  No policy
can be harder to create.
     All organisms need sustenance to survive after birth.  To wean
technology-driven industries, we must apply our skill, knowledge,
and judgement to economic policy shaped with a liberal sprinkling
of egalitarian glue.  Our industries will become self-sustaining
organizations if government policies implement this dichotomy.
     Egalitarian glue engineers solutions to problems where none
existed previously.  It can act as a rinse that etches the silicon
ceiling into a polished looking glass where all people can see
equally well.  We need a remedy that permits strong minds to
discuss ideas, prevent average minds from discussing events, and
weak minds from ever gaining power.
Biography
Richard Marlon Stein is a freelance writer and software consultant
who resides in Santa Clara, CA.  He can be reached via Internet as
[e-mail].
(C) Copyright 1992 Richard Marlon Stein
--
Richard Marlon Stein, Internet: [e-mail]
To those who know what is not known;  The Chronicles of Microwave Jim!