MATLAB News & Notes - November 2003
Meet The Developers
Developers Bill McKeeman, who works on the MATLAB JIT technology in the Language of Technical Computing group, and Darel Linebarger, a manager in the DSP and Communications Design Automation group, joined the company after successful teaching careers.
|
|
|
|
|
|
Bill On
my 12th birthday,
I told my mother that I wanted
to be a university professor. I had
no idea what I was asking for. But I
never bothered to rethink my goal, and
through 20 years of further adventures,
including pilot school and teaching
physics at the Naval Academy, I finally
became an assistant professor in the
brand new Computer Science
Department at Stanford.
Darel As
an undergrad
co-op student at Southern Methodist
University in Dallas, Texas, I learned an
important lesson: school was a lot more
fun than work! I had done well in my
classes, so when I found out that I could
be paid to go to graduate school in
engineering, I said “Sign me up!” BILL McKEEMAN On my 12th birthday, I told my mother that I wanted to be a university professor. I had no idea what I was asking for. But I never bothered to rethink my goal, and through 20 years of further adventures, including pilot school and teaching physics at the Naval Academy, I finally became an assistant professor in the brand new Computer Science Department at Stanford.
Ten years later, I was department chair at the University of California, Santa Cruz. Along the way I did a lot of teaching, which I liked, and dealt with journal referees, grant officers, and campus politicians—none of whom I liked. My love affair with compiler writing was the constant in my life, and still is.
After a visit to Xerox Research, I came east to head the new Master of Software Engineering degree program for Dr. An Wang, the founder of Wang Laboratories. It was a kind of academic heaven: no tenure, lots of teaching, and well funded, but Dr.Wang ran into hard times, and Wang Institute was closed.
Following a sabbatical at Harvard, it became clear that it was time to move on. During a job interview, a dean did me a real service. He asked me how much money I was going to raise the next year. Anything less than $2 million was the wrong answer. I had discovered that a suit, tie, and silver hair got me offers to manage organizations (granting agencies, departments, even whole schools), or entry into grantsmanship, but not much chance to teach.
Instead, I took a position as programmer in the compiler group at Digital Equipment Corporation. Toward the end of a fruitful decade at the keyboard, I was offered a chance to teach part time at Dartmouth. I accepted, then retired from Digital and came to The MathWorks. The MathWorks honored my commitment to teach, and suddenly I had the best of both worlds: interesting work in Natick and good students in Hanover. That three hours of asphalt separating them is a nuisance I can deal with.
Forty years ago, I specialized in fast compilers. This made real sense in the days of the PDP-1 and IBM mainframes. As cycle times went from milliseconds to nanoseconds, all compilers became fast enough for developers. The market for my tricks had dried up. But in the brave new just-in-time world, compile time is user time, and The MathWorks needed just what I knew how to build. You never know.
DAREL LINEBARGER As an undergrad co-op student at Southern Methodist University in Dallas, Texas, I learned an important lesson: school was a lot more fun than work! I had done well in my classes, so when I found out that I could be paid to go to graduate school in engineering, I said “Sign me up!” I was happy to spend a few more years in school, and thought that afterwards I would consider academia as a career.
I entered graduate school at Rice University in 1981, where I obtained an M.S. and Ph.D. in Electrical Engineering. At Rice, I especially enjoyed writing software to handle simulations for the problems that I was studying. During that time, I learned C and was exposed to the Cleve Moler Fortran version of MATLAB. I also remember some engineers from Texas Instruments talking to us about this new device they were building: the TMS32010, which was the first DSP processor. All of these factors were to play a large role in my career.
After Rice, I did a short stint at Martin Marietta, but when I heard that the University of Texas at Dallas was starting an engineering school, I jumped at the opportunity. The possibility of going home to Dallas and back to academia was very appealing, and I was fortunate enough to receive an offer.
In 1987, I became one of the first six faculty members in the Department of Electrical Engineering at UTD. In those early years, my colleagues and I established B.S., M.S., and Ph.D. programs. We got our first ABET accreditation and built up respected undergraduate and graduate programs. We also published enough and raised enough money to be granted tenure.
Although I liked my job at UTD, I did not enjoy writing proposals or the administrative workload of building a department. Then, through some consulting opportunities, I recalled how much I enjoyed writing software for mathematical modeling. One consulting job was in the Wireless Communications group at Texas Instruments (TI), where I learned a lot about fixed-point DSP implementations for wireless applications. While this experience helped me teach courses directed at the DSP world, I also found that I was enjoying my work at TI more and more.
In 1999, I decided to look for employment in industry and seized an opportunity to join The MathWorks to work on tools for fixed-point DSP. It was the perfect match. Although I miss some aspects of academia, I now do more work that I enjoy and less that I don’t enjoy. I lead a great team of engineers, and working on the MathWorks DSP tools, especially fixed point, is a blast!
| For more information, visit |
|
