(To appear in the October'95 issue of the IEEE SP Magazine.)
  
****************************************************************************
(From Jack Deller who normally writes this column.)

Editor's Note: The editor wants to acknowledge and express sincere
appreciation to associate editor Joe Picone and the contributing
authors for a very timely, informative, and interesting issue on SP
education. Joe has not only done an extraordinary job of organizing
this special issue, but below he has set a new high standard in
"Editor's Message" writing that will be difficult for the editor to
live up to in future issues.  I take solace in the fact that our past
record of achievement in this area will not lead the reader to expect
such.
  
In his message, Joe mentions having taken a graduate course from me in
1980. He says I made it look easy.  It wasn't.  We were both in our
twenties, I a shiny new assistant professor, and, incredibly enough, I
knew even less about signal processing than I do now.  Nevertheless,
I'm flattered that Joe thought so.  Over the years, I have learned
much more about speech processing from Joe than he could ever have
learned from me.
  
One more thing.  I did chair the board meeting as Joe suggests below,
but didn't dare make the wine selection.  I'm afraid I don't know a
Chardonnay from a Chebyshev.
****************************************************************************
  
At ICASSP'95, I had the pleasure of attending the first Signal
Processing Magazine Editorial Board meeting, which was chaired by the
distinguished editor of this magazine (he got to choose the wine).  I
found myself in the middle of simultaneous discussions of two
seemingly unrelated topics: how Jack Deller finds inspiration for his
columns, and undergraduate curriculum overhaul.  Unbeknownst to me,
life would take one of its strange turns, and I now find myself faced
with the daunting task of explaining what these topics have to do with
digital signal processing (DSP) education, and why Jack Deller isn't
writing this column (you will have to wait to the end for an answer to
the latter).
  
Having just relocated to the peaceful and serene confines of academia
over a year ago, I figured it was time to kick back and enjoy the
stability and security of a public-sector institution. Visions of
teaching speech processing and introductory signal processing into the
next century were dancing through my head. I had, after all, survived
the turbulent times of corporate down-sizing of the early 90's, spent
two years starting a research lab in Japan, and generally seen 10
years vanish into the thin air.  Time for a change. What could be so
hard about teaching the same courses I took as a student for the next
20 years. (In fact, one of my graduate courses was taught by the
Editor, and he didn't seem to have too much trouble.)  With blissfully
naive anticipation, I packed my bags and headed to the deep south.
  
My image of a serene, insulated academic environment where I would, at
last, be free to pursue a life of leisurely scholarship, quickly
disintegrated.  I find that I am once again in the middle of a
turbulent storm.  Across the nation, engineering enrollment is down,
research funding is down, the job market is tight, and students don't
seem to be getting any more motivated or well-prepared.  Not
surprisingly, curriculum overhaul is the hot item.  In fact, it is a
movement sweeping the country.
  
At my institution, for example, after just having completed our most
successful ABET evaluation ever, we are now discussing dismantling our
entire undergraduate program (in steps of course).  There are good
reasons for this nationwide curriculum-revision frenzy. the average
engineering students at a public university now takes over five years
to receive a b.s. degree.  The number of semester credit hours
required at some schools has grown to close to 140 - which means some
semesters will require 18 hours or more of engineering coursework.
And most engineering students are still only required to take one
programming course. clearly, something has to give.
  				   
One of the major driving forces for this overhaul is the need to
integrate computing technologies into the classroom in a much more
significant and extensive way than has been traditional. "War" stories
abound concerning battles between the "computer advocates" and
"traditionalists" whose view of updating the electrical engineering
curriculum means adding a fourth physics course involving solid state
electronics, introducing relevant courses like "220V Power And You",
and solving Smith charts on a PC.  But the change is, of course, more
fundamental, and is slowly graying the lines between computer science,
computer engineering, and electrical engineering.  (How many of the
readers remember how to bias a transistor?)
  
Thus, it was no surprise that at ICASSP'95, I found all my colleagues
in academia grappling with the same issues - how to cram computer
software and hardware courses into an already overloaded curriculum.
During this infamous dinner meeting, I did manage to finagle a copy of
a report about curriculum overhaul from the person sitting to my right
who just completed such a study at his institution. In fact, I
collected several of these from colleagues at ICASSP, and they all
look remarkably similar - fewer hours, more design courses, but still
very little required computer science.  DSP is inevitably one of the
central disciplines in this modern undergraduate curriculum
debate. With this special issue, SP magazine weighs in on these
important developments with three timely articles that are describe
below.
  
Concurrently with the curriculum discussion, a second conversation
broke out during the board meeting dinner concerning exercise cycles
and the story behind Jack Deller's column (May, 1995 Editor's Message)
on this topic. Weeks later, while sitting though another seemingly
interminable curriculum meeting, I found my mind wandering to memories
of the enjoyable ICASSP dinner, the topic of the educational process
(someone at my meeting was talking about the importance of
fundamentals again), and one's favorite home improvement power
tools. All good engineers, of course, are closet home remodelers
(though I have told my wife many times that I am retired!).
  
The previous weekend, I had finally got around to cleaning my most
prized power tool possession - my Sears Craftsman 2.0 HP Shop-Vac. I
have the super-duper unit: 2 HP and 16 gal capacity.  It was the most
powerful unit made in 1985 - the main reason I bought it.  Large
shop-vac's are one of the engineering marvels of the 20th century. Ask
anyone who has owned a long-haired dog and a car with cloth interior.
  
I clean this unit about four times a year while regularly ignoring the
manufacturer's recommendations that an empty canister and clean filter
are crucial elements in vacuum science.  (I know it is time to clean
it when more stuff comes out than goes in.)  But this past weekend was
special, I replaced the air filter, a paper cartridge, for the first
time since I purchased the unit. Why? The cartridge costs $15 a pop,
and like all good engineers, I was convinced this was a money-making
ploy by the manufacturer. I had renovated two houses with this unit,
cleaned pools, cars, and spas many times, even done yard work with
it. Guilt had caught up with me - it was time for a major cleaning.
  
The 20th Century shop-vac is an amazing invention (visit the Ford
Museum in Detroit - sight of ICASSP'95 - if you don't believe me).
First of all, it is the type of technology that simply doesn't
scale. I have used many vacuum cleaners in my lifetime. Very simply,
this is a technology where bigger is better.  The large units will
suck up anything, even dreaded dog hairs on a car's upholstery -
something no normal vacuum cleaner will do. I have been to Hades and
back with this unit. My wife, who was very skeptical of this tank when
I brought it home, has even become a believer in it.
  
Even Deller would be hard-pressed to find a connection between
shop-vac's and DSP education, but curriculum meetings are fertile
ground for bizarre free mental association.  It was at this meeting
that the shop vac as a metaphor for dsp education became clear to
me. It is amazing, I thought, when you clean it out, how many bits and
pieces of your life you find. I usually dump the contents every few
months, but remnants persist. It sort of collects artifacts from your
life, and becomes a time capsule for its owner. This last cleaning, in
May, produced silver tinsel from Christmas, a few Christmas tree
needles, parts to my children's games, and of course, a missing puzzle
piece. One of my students told me he always finds at least one Lego in
his, even though it has been a long time since there were children in
his house.
  
We often ask engineering students to suck up all sorts of matter,
telling them this stuff will be useful someday.  A good engineer
collects all this knowledge in bits and pieces, not quite sure how it
all fits together, and eventually finds a use for most of it in the
most unusual places. If you examined the mind of some of my students,
I am sure it would look like my shop-vac's drum - a bunch of
disconnected junk just laying around - occasionally connected in
strange and unusual ways that you would have never imagined - most
often not connected in the "proper" or "expected way." No matter how
we try to impose structure on our curricula, it is never absorbed the
way we expect it to be. And it all just seems to come together a few
months before graduation.
  
Then there is the scalability angle. Small shop-vac's just don't do
the job. Fragments of engineering knowledge often don't do the trick
either - a good engineer needs perspective, breadth AND maturity.  One
programming class doesn't work - we really need to evolve a philosophy
of computer-based education. Programming, or other mainstream topics,
taught in the abstract to engineers today, just don't seem to work
unless students see the bigger picture.  A more successful educational
model seems to be the "sink or swim" approach - immerse students in
thick of a large problem and let them acquire knowledge on an
as-needed basis. Psychologists call this "hands-on" learning and argue
this is the way it has always been.  Certainly, when it comes to
computing, we all feel like we are sinking in the middle of a big
lake, and we have all we can do to tread water.
  
Finally, my favorite connection between the two is the "been to Hades
and back" angle. An engineering education is like a good tool - it
will be there when you need it most, and do the job. We often talk
about an electrical engineer AS someone who brings his OR HER bag of
tricks to the job and knows how to pull out the right one at the right
time. The truth is, academics can't predict the future any better than
people in industry. Education often lags the real world in content
(how many schools still teach Fortran), but, as a senior professor in
our department quipped, an engineer with strong fundamentals seems to
always survive the boom-and-bust cycles in engineering.  This is
primarily because he or she can always fall back on this strong
foundation known as a B.S. in electrical engineering.
  
On the serious side, this month's issue is devoted to digital signal
processing (DSP) education. DSP, like many modern engineering
disciplines, is one of those areas that was once exclusively the realm
of graduate school (where all the interesting stuff is taught?), but
now has migrated down to junior level undergraduate courses.  In
addition, it is a heavily computer-based and software-based curriculum -
which translates into a constant struggle how to balance chalkboard
instruction with hands-on computer-based learning. This is a major
issue in modern-day engineering curricula.
  
In this issue, we present three articles dealing with approaches to
DSP education, primarily focusing on undergraduate education.  The
first article, by Dolores Etter of the University of Colorado and
Geoffrey Orsak of George Mason University, deals with a number of new
approaches to education: distance-learning involving teams of students
at multiple universities, the use of a highly-evolved computational
tool known as Matlab, and the role of simulation tools in
computer-based problem-solving. This paper provides a glimpse into a
future that might bring "virtual universities" or "classrooms without
walls."
  
The second paper, by Fred Taylor and Jon Mellott of the University of
Florida, provides a hardware perspective on DSP education. They
discuss the use of a popular low cost DSP hardware package developed
by Texas Instruments, and show how students can be exposed to
real-time DSP simultaneously with their theoretical development. This
philosophy reflects a growing trend amongst universities to respond to
pressure from government and industry to emphasize more hands-on
learning and practical problems in undergraduate education. Taylor and
Mallott teach a course based on an industry-leading processor, and
attempt to produce students with a practical working knowledge of DSP
at the expense of some of the theoretical foundations (practice first,
theory next might summarize this approach).
  
The third paper, by William Ebel and Nick Younan of Mississippi State
University, provides a high-level discussion of several competing
numerical simulation tools, and attempts to compare and contrast their
effectiveness in education. This is no small chore, because as we all
know, the learning curve on such tools is steep, and once an
investment has been made, it is hard to go back. The effort required
to computerize a curriculum is overwhelming, and once you settle on a
particular tool, it is hard to change. At a recent meeting in our
department, we listed all of the computer tools students must learn in
their undergraduate education today - it made me glad I am no longer a
college student.
  
I am confident that  this collection of papers will stimulate thinking
on the role of computer-based curricula in undergraduate education.
This is perhaps the greatest challenge we face today in academia - how
to balance our high-level mission of developing rigorous analytical
and problem-solving skills in students with the need to improve the
marketability of our graduates. No jobs - no students: this is a fact
of life in engineering education. Students no longer walk into jobs
that train them for several years - in the competitive markets today,
students must be immediately productive in the workplace. It is
forcing us to reevaluate many aspects of higher-level education.
  
One more question left to answer: why isn't Jack Deller writing this
column? Not sure, but I think he is tired of getting email messages at
2 AM about useless household appliances and their relevance to his
column.
  
-The Terminal Man (Joe Picone)