The article titled "Scholars and Dollars," excerpted from Randall
Stross's book Steve Jobs and the NeXT Big Thing in the May/June issue of
Educom Review, is a purely market-oriented analysis of the history of
NeXT computers in higher education. As a result, the account leaves us
with the profoundly misleading impression that Steve Jobs and his NeXT
software and hardware engineers offered little if anything to colleges
and universities. The story of the ideas behind NeXT, their impact, and
their implications for the future of instructional technology is quite
different when not so narrowly construed. Unfortunate as it may be,
market performance and intellectual quality are not always linked.
Fundamental Questions
The history of NeXT and NeXTSTEP in educational computing raises some
fundamental questions, as follows, about the current values and
methodologies of those of us in this field.
What are and what should be the objectives of technology in higher
education? Are our aims now limited to providing networking for personal
computers owned by students? Given our checkered history, should we now
jettison larger pedagogical and research goals of the kind that require
substantial institutional and faculty investment, experimentation, and
risk in favor of a fiscally conservative basic service orientation?
What responsibilities should faculty, administrators, and educational
technologists shoulder in making computing technology work for
education? Should we, as Stross seems to, expect the technology industry
to provide off-the-shelf solutions to educational problems, or will
success measured in teaching terms require significant efforts to
customize computer-based tools to the pedagogical designs of individual
faculty?
Is there still a place for risk taking in educational technology? Has
the era of risk taking in the form of developing innovative
instructional technology been eclipsed by a conservatism that favors
acquiring hardware and software from so-called reliable vendors? If so,
is that shift an overreaction to past failures and one that will
undermine the progress we have already made?
These are issues that concern the soul of technological support for
education. The history of NeXT and NeXTSTEP to this point suggests some
sobering answers.
Stross recites a litany of explanations (based on marketing criteria)
for NeXT's failure as a workstation vendor for higher education. He
validates the thinking of those who from the beginning pronounced NeXT
an unwise risk, claiming that the company's problems were an inevitable
outcome of Jobs's failure to recognize that academic computing was
moving away from instructional projects like PLATO and ATHENA and toward
$1000-a-pop personal computers for use by students ensconced in their
dorm rooms.
Betting Heavily on Pedagogical Success
Stross claims that the only institution to bet heavily on NeXT was
Allegheny College, regularly referred to in articles about NeXT as "a
small liberal arts college in Meadville, Pennsylvania." As Allegheny was
bold enough to take such a risk four years ago and is not a large
university with the resources or high profile of an MIT or a University
of Illinois, it seems appropriate, in evaluating Jobs's contribution
more substantively, to ask how the institution's bet has paid off. As
former director of educational computing at Allegheny, I hope to offer a
broader perspective on the educational value of the technological
support program designed and implemented by faculty, administrators,
students, and staff at that institution.
Stross's flippant attitude toward Allegheny's motivation in making a
commitment to NeXT in the first place is indicative of his narrowly
market-minded cynicism. He describes Allegheny's commitment to NeXT as a
"ploy to attract national attention in the same calculating way that
Drexel University had bet early on the Macintosh." Nothing could be
further from the truth, and in fact Allegheny's reasons for choosing
NeXT illustrate the very problem with taking a singularly market-based
perspective, especially one that favors lower-end $1000-a-pop
alternatives.
Allegheny's choice was the result of a joint faculty and administration
effort to create a computing environment that would support teaching and
learning in the diverse environs of a liberal arts college. The bottom
line at most liberal arts colleges concerns improving teaching and
learning for this year's class, not gaining national attention. At the
instigation of Dr. Edward Barboni, then Allegheny's vice president for
planning and information, the Faculty Computing Committee at Allegheny
undertook a yearlong study to identify the instructional issues the
faculty perceived as the most important and those areas they saw as in
need of improvement.
The design of this study was as unique as the application of technology
that emerged from it. The committee engaged in extensive discussions
about pedagogy at the college--not computing and pedagogy--just
pedagogy. The idea was to evaluate the potential for computing as a tool
to solve pedagogical problems only after a comprehensive map of teaching
objectives, methodologies, and perspectives of Allegheny's faculty had
been developed. A mad dash to keep up with the Joneses (or Carnegie
Mellons in this case) in computing was, as it should have been,
supplanted by the prior consideration of real educational needs.
One thing that became clear in the process was that, with the exception
of extremely general purpose applications like electronic mail and word
processing, professors would usually not use computer-based
instructional tools that did not conform to the specific, even
idiosyncratic needs of their classes. This is true nationally, and it
partly explains why turnkey educational computing applications have made
little impact. Even excellent instructional software often wins Educom
awards and is used successfully and creatively by its inventors, but
then dies away, with little or no migration to other institutions.
Faculty are unwilling to adapt their courses to software that doesn't
quite fit their pedagogical needs, and typically they will not invest
even small amounts of time in learning and coordinating someone else's
software. (Little wonder--we don't even like to use each other's
textbooks!) Some faculty will, however, invest time and energy in
developing their own courseware. There is a consensus among faculty who
have developed courseware that such an activity not only addresses real
educational needs but also produces new ideas about teaching and
research, and in fact routinely causes substantial rethinking of
traditional subjects and how they are taught.
The Allegheny committee agreed that if educational computer programs
beyond general-level applications (i.e., word processing and electronic
mail) were really to fill their needs, faculty were, for the most part,
going to have to design the software themselves. Thus, their main
criterion for choosing a computing platform became the ease and speed
with which high-quality custom applications could be developed on that
platform. NeXTSTEP provided then--as it does now (having been ported to
Intel processor and, soon, to Hewlett-Packard and Sun workstations)--an
object-oriented and highly visual environment for courseware development
that is an order of magnitude better than any of its competitors.
Allegheny's decision was risky but unambiguous, if the institution were
to hold to its educational goals.
Allegheny's NeXT Steps
Allegheny is a college with approximately 1,800 students and 140
faculty. Four years after committing to NeXT, they now have more than
250 networked workstations running NeXTSTEP. A year from now, every
faculty member who wants one will have a NeXTSTEP workstation, and there
will be one public computer running NeXTSTEP for every eight students.
Outside evaluations of Allegheny's efforts have yielded surprising
results. Of those responding to a recent survey, the faculty believe by
a ratio of 20:1 that the college has improved educationally as a result
of its initiative in educational computing. Five times as many students
report that learning is more enjoyable because of the computer than
report it is less so, and the number who agree that learning NeXTSTEP
was easy outnumber those who disagree by more than 10:1.
Besides WordPerfect, Improv, Mathematica, Mosaic, Gopher, and a host of
other high-end general tools, there are now close to 150 custom
applications being used in classes in mathematics, computer science,
physics, chemistry, biology, geology, philosophy, psychology, modern
languages, English, and economics. The applications range in complexity
from a simple discussion group coordinator that is used in several
different humanities classes to a sophisticated program for modeling
continental compression that is used for several weeks of class and for
upper-level undergraduate research projects in geology. By this fall,
every lab in a two-semester sequence of Introductory Biology will be
accompanied by a dry-lab computer application that either introduces
students to complicated lab equipment, or complements wet-lab content
with simulated experiments, or provides customized tools with which to
analyze data collected in the lab. It is perhaps worth noting that some
of these are very close to the kind of "virtual DNA wet-labs that Steve
Jobs envisaged" and that Stross implies are simply not a reality in
contemporary instructional computing.
This was all done with a half-time faculty director; two full-time
support staff who were responsible for training faculty and students and
for training, coordinating, and supervising all student-user
consultants, not to mention all other administrative details; a head of
technical services; one networking technician; one and a half developers
for the first two years and three full-time developers for the second
two years; and, most important, an energetic and open-minded faculty.
The vast majority of all money spent on personnel and on hardware came
from grants: The Pew Charitable Trusts, the National Science Foundation,
the Culpepper Foundation, the Vira Heinz Endowments, and the George I.
Alden trusts all found Allegheny's vision for using NeXTSTEP
sufficiently compelling to underwrite some risk taking for the sake of
better education. The same initiative on market-popular DOS/Windows PCs
or Macintosh computers would not have produced even a remote
approximation of the environment that students now have at Allegheny.
Objects of Success
Creating a computing environment that facilitates the development of
high-quality educational applications is what Steve Jobs and his
engineers have truly done right. Four years and hundreds of high-
quality, specifically targeted educational applications later, Allegheny
and a host of other NeXT sites around the country have succeeded in
actually getting what they paid for. It is worth noting that, contrary
to the impression given by Stross's article, besides Allegheny, both
Rose-Hulman Institute of Technology and Colorado School of Mines made
campuswide commitments to NeXT and that individual departments and
scholars at California State University at Los Angeles, MIT, Indiana
University, the University of Michigan, Princeton University, Columbia
University, McGill University, and many other higher education
institutions supported major instructional and research computing work
with NeXTs. That NeXTSTEP has not carved a major niche next to DOS, Mac
OS, and standard UNIX is certainly not due to the quality of the
product.
The NeXTSTEP application development environment combines true object-
oriented programming with visual representations of the objects and
menu-based control of the messaging processes that drive events in
object-oriented applications. This combination is a conceptual tour de
force that has not yet found the education market it should. Part of
the reason is that for a long time NeXT itself did not seem to
understand just how powerful a development environment its engineers had
created. It focused its education marketing for NeXT on its publication
tools, its value as a Mathematica platform, and its seamless integration
of standard pieces of software. There is good reason to believe that
neither Jobs nor most of his educational marketing staff had realized
what individuals at institutions like Allegheny were already focusing
on: there were tools to support significant new initiatives in
educational computing hiding behind the Interface Builder icon on the
NeXTSTEP dock.
NeXTSTEP comes with a library of objects (e.g., a window object, a view
object that controls all PostScript drawing, and a menu object) that
program developers can reuse and even customize individually in
constructing their applications. Allegheny College and other sites have
added to that library their own reusable objects tailored for
educational applications (e.g., a Cartesian graph object, a petri dish
object, a text-searching object, and a data entry object). Careful
object design and development for discipline-specific applications can
result in substantial easing of the burdens of creating courseware.
There are two subtle but truly revolutionary technical advances in
NeXTSTEP that make all the difference in simplifying programming. First,
any object can be given a visual representation that can then be
manipulated by drag-and-drop, point-and-click tools--known in earlier
versions of NeXTSTEP as Interface Builder--to produce complete
applications. This architecture parlays the object-oriented paradigm
into a visually articulated but largely unrestricted programming
environment. Second, a clever object designer can create objects for use
with these tools such that messaging between objects can be programmed
by using a mouse to choose from a list of messages an object will
recognize. This means that with significant effort in object design and
an eye ever toward the needs of other developers in a particular
discipline, an object engineer can create a rich set of building blocks
with which others can invent new applications literally without writing
any code. It should be noted that a great deal of work still needs to be
done on object engineering in order to realize this very real potential
of NeXTSTEP.
Similar ideas have been used in development tools such as Visual Basic
and Toolbook for Windows, Prograph for Macintoshes, and various sets of
widgets for X-Windows, but all of these are pale shadows of the tools
provided in developer versions of NeXTSTEP. Even if Taligent becomes a
reality, there is the open question of whether Microsoft and its
partners will pay as close attention to a development environment as
NeXT has.
The new ideas in application development that underlie NeXTSTEP have
profound implications for educational computing. First, application
development has opened to a wide population of both professors and
students a new class of what I choose to call layman-developers. I think
it is not an exaggeration to call this aspect of NeXTSTEP part of the
intellectual democratization of computing. After four years of steering
this course, there were forty faculty who had engaged to some degree in
instructional application design and development at Allegheny College.
Faculty at Rose-Hulman and the University of Houston have provided large
numbers of students opportunities to create sophisticated instructional
software--an experience that undoubtedly contributes far more to their
education than the use of a dozen off-the-shelf educational
applications. Most important, the object-oriented, visually articulated
programming environment allows layman-developers to focus more on the
ideas they are exploring rather than on programming details.
Second, such an operating system and application development environment
frees academic computing departments to hire the kinds of programmers
they have long needed but have been unable to hire because of the
technical demands of the job. A notorious problem in instructional
application development is that many professional programmers lack
pedagogical sensibilities. Such programmers often become absorbed in the
technical details of programming to the detriment of the instructional
value of projects (to say nothing of actually completing the software
for use by students). Object-oriented, visually realized programming
means that, in addition to object-engineers, discipline experts can be
hired and quickly trained to do functionally and graphically
sophisticated programming for instruction in their discipline. The
application developers working in Educational Computing at Allegheny at
the beginning of the project were actually three professors: a chemist,
a mathematician, and a philosopher.
Not for Liberal Arts Only
Even in more research-oriented settings, such a development environment
can change the balance of intellectual energy between that spent on
programming and that spent on problem solution. During 1993, I helped
guide one of the consulting clinics conducted regularly in collaboration
with commercial clients by the mathematics department at Claremont
Graduate School. A group of students was given the task of developing a
genetic algorithm to sort out radar ghosts from true targets. NeXTSTEP
development tools allowed that clinic both to write the computational
engine for this software in standard ANSI C for export to more
mainstream operating systems at Hughes Aircraft Corporation and to
create a sophisticated graphical interface in just two days of
programming time for display and testing of the results. Unfortunately
for Hughes, even though it wanted the entire package, it could take only
the computational kernel because it had no NeXTSTEP computers on which
to run our valuable visualization interface of the entire problem and
set of solutions.
A graduate student in the Center for Politics and Policy at Claremont
has struggled with me for the past six months to program with Visual
Basics--a genetic algorithm with a rich input and output interface--to
let social scientists experiment with his technique for solving game
theory problems. He could have concentrated far less on programming and
far more on game theory if he had used NeXTSTEP tools instead. Such
differences in ease and power of programming can have a significant
impact on the quality of results that emerge. It is a difference that
can affect the projects, goals, and lives of both scholars and their
students.
The Bottom Line
Jobs-bashing is comparatively easy. Jobs does appear arrogant, and he
did make many marketing mistakes with NeXT. Nevertheless, despite
hardware problems and lack of software vendor support, Jobs and his
colleagues at NeXT have presented higher education with a remarkable set
of tools with which to create graphically, textually, and functionally
rich applications while significantly reducing the need for programming
expertise and large blocks of programming time. In short, they have
given us the tools to create a new class of programmers--the layman-
programmer. Or, for academe, a better designation might be the scholar-
programmer.
Allegheny College has developed that paradigm to a very large degree and
offers the rest of us the interesting vision of a possible future for
instructional computing. One clear implication of the Stross excerpt is
that Allegheny made a big mistake in betting on NeXT as a primary
computing platform because of NeXT's ultimate inability to be
competitive as a hardware vendor. That conclusion can rest only on the
assumption that colleges ought to choose computer hardware, operating
systems, and application environments by those objects' likelihood of
being solid commercial successes.
No one would deny that corporate viability has to be a consideration for
academic institutions when institutions choose computing resources. But
if we, as the educational computing community, have in fact adopted
Stross's implicit but clear message that an institution's instructional
technology strategies should be judged by whether those strategies back
the right commercial horse, then the innovations and risk taking that
must precede transformations in any aspect of education are unlikely to
take place in instructional technology.
Did Steve Jobs let higher education down? Was his promise to build the
next generation of computers for higher education itself just a
marketing ploy? Stross observes that there has been a long succession of
failures to produce dramatic educational transformation through
application of instructional technology, of which NeXT is only the
latest example. Such a succession of misadventures should perhaps lead
us to look not to failures in our technologies but failures in ourselves
as educators. Are we expecting technology vendors to hand us turnkey
educational solutions? Do we expect to plug in a new machine like the
NeXT and have it instantly solve problems in higher education? If so, we
will always be disappointed and thus will pronounce the new technology a
failure.
If we, as professors, administrators, and educational technologists, are
unwilling to invest the money, time, and imagination necessary to find
the pedagogical transformations made possible by technological
innovations, we cannot blame the computer industry for resulting
failures of technology to have the hoped-for impact. Steve Jobs has
presented us with dramatically new possibilities for instructional
computing. He could not be expected to also use those possibilities to
solve educational problems: that is a task for educators, not
technologists.
Professors, administrators, and educational technologists at Allegheny
College and in many academic departments around the country have used
NeXTSTEP to solve educational problems and extend intellectual vision.
Our measure of their successes and failures should be in intellectual
and educational, not marketing, terms: Did more students learn better?
Were new perspectives on subject matters introduced by the creation of
software? Was there excitement about the projects and was there
improvement in the overall learning environment?
By these criteria, the institutions that invested in NeXTSTEP were and
continue to be successful. That NeXTSTEP will remain an option as an
operating system on Intel processors and soon on Hewlett-Packard and Sun
platforms (and even within the Solaris operating system in the form of
the OpenStep alliance between NeXT and Sun) should be welcomed and
encouraged. If we as educators allow ourselves to judge instructional
technology companies by their marketing departments and year-end
dividends and not by the intellectual quality and value of the products
they produce, then we are in serious trouble.