Path: utzoo!mnetor!uunet!husc6!think!ames!aurora!labrea!decwrl!hplabs!hplabsz!taylor
From: spaf@purdue.edu (Gene Spafford)
Newsgroups: comp.society
Subject: Re: University Education and Industry Needs
Message-ID: <1450@hplabsz.HPL.HP.COM>
Date: 20 Jan 88 08:09:57 GMT
Sender: taylor@hplabsz.HPL.HP.COM
Organization: Department of Computer Science, Purdue University
Lines: 94
Approved: taylor@hplabs

I am coming into this discussion in the middle, so pardon me if I
cover something already said.

Basically, I see this discussion as over the question as one of how to
get University research in cooperation with Industry (both funding and
expertise).

Well, such a beast (actually, many) exists; I'm part of one.  Dave's
diagram illustrates our goal nicely:

>   Universities ===================+  --> Applied Research
>     |  + - - - - - - - - - - - - -|- - - - - - - - - - +
>     |  |              +===========|=======================+
>     +=================|===========+                    |  |
>        + - - - - - - -|- - - - - - - - - - - - - - - - +  |
>       /               +================= Industry ========+
>  Areas of Research

Here's the story: 

The NSF has a program to establish cooperative Industry/University
research centers in critical technology areas.  These centers are
funded by industrial partners and research is done (largely) by the
university component.  The direction of the center and the projects is
shared between the industry affiliates and the university coordinator.
In the last few years the NSF has established over 30 of these centers
in critical technology areas like polymer research, recombinant DNA,
materials handling, and so on.  All but one or two have been
successful, with a few developing patentable discoveries worth millions
to those centers and their affiliates.

The NSF starts the centers with an initial 5 years of administrative
funding and support, and helps develop the structure of each center
according to its mission.  They also provide administrative expertise 
and assistance, and help to ensure stability.

The Software Engineering Research Center is one such center, founded in
late 1986 jointly at Purdue and at the University of Florida,
Gainesville.  The SERC is charged with research in all aspects of
software engineering -- including specification, design aids, metrics,
testing technology, maintenance aids, verification techniques,
debugging, interfaces, and many other topics.  We have about 15 active
projects at the moment, with at least that many faculty involved,
including such SE luminaries as Sam Conte, Harlan Mills, Rich DeMillo,
Stephen Yau, Buster Dunsmore (and myself :-), plus many other faculty and
students.

The SERC currently has over a dozen industry affiliates and more in the
hopper, plus support from some state and federal research
organizations.  Each affiliate pays a small yearly fee ($30K, less a
federal tax *credit* -- less than 1/3 the cost of capitalizing a new
researcher), and gains a voting membership on the advisory board.  They
also get first crack at proprietary rights in our research results,
along with the usual kinds of perks like personnel exchanges, short
courses, etc.  The advisory board meets twice a year and hears research
reports and proposals from the university research folks.  The IAB
(advisory board) decides what to fund, and at what level.  They also
provide feedback about things they'd like to see included in the
research, and about internal projects that may have a tie-in to the
research.  They also appoint a subcommittee of members to monitor each
project.  Thus, they get some definite say in the projects done, and
they closely observe and guide the research that is done.

>From the standpoint of the researchers, we get to research "real-world"
problems and place prototypes for evaluation.  We get to cooperate with
researchers in corporate labs, we get access to coroporate data and
experience, and we have access to reasonable funding without the
additional hassles involved in going to funding agencies.  For the
affiliates, they get access to over (currently) $600,000 (annual) of
university research by top people in the field, the potential for
proprietary commercial use of the research results, and interaction
with the university and industry research community for their people

That's really a thumbnail sketch of the SERC, and the other NSF centers
too.  The idea *does* work, and works well, and it doesn't require the
companies to put up millions of dollars on research they have little
say in.  For small investments they get a great deal in return,
specifically targeted to be available in a few years' time.  It's a
*great* partnership.

If you'd like more info on the SERC, send me your *US MAIL* address and
I'll mail you a packet or indicate someone at your company who already
has that information.  Our current affiliates are:  ATTIS, CSC Corp,
DEC, HP, IBM, GTE Harris, Magnavox, ModComp, Racal-Milgo, Arthur
Anderson, and US AIRMICS.  BNR is about to join.  Groups at Kodak,
Bellcore, Gould, Sun, Apple, Xerox, Pyramid, TI, Contel, and Covex are
all expressing interest, although none have made any commitment to
join.  Our goal is to limit affiliates to 20 or 25.

If you're interested in info on our available technical reports or
specific projects, send me e-mail and indicate that...I'll see what
I can provide.

Gene Spafford