(Kind Of)
Link to Time Line
Being at a Jesuit university, we wonder about things that people at other schools do not even think about. For example, who is the patron saint of computer science? As far as I can find out, there is none currently assigned. I respectfully suggest Mickey Mouse as a candidate for that position and I would like to explain my reasoning.
Mickey is a natural. He has three fingers and a thumb on each hand. With our four fingers and thumb, we count in base ten but the computer is binary. Mickey naturally counts in base eight. Did you really believe that this is a coincidence?
As further evidence, I remind you that computer science is often referred to as a Mickey Mouse field. Now you know what people mean when they say that.
You might say that Mickey is not a real person and that this would
disqualify him to be a saint but I was raised with St. George, the dragon
slayer, in my lives of the saints. Saints were much more interesting before
Vatican II demoted some of them.
Mickey Mouse was “born” in 1928. When he was born, you might say that there were no computers in the world. A more precise statement would be that there were many computers but no automatic computers. “Computer” was a government job title for people who spent the day with adding machines computing tables. The work was tedious but the need was real. There were always a backlog of work that needed to be done.
By 1946 when Mickey was 18 and considering college, there were no computer science programs as there were only a handful of hand-made automatic computers in the world. For example, at Harvard, they had built the Mark I out of telephone equipment and in England, they had used the Colossus during the war to decipher German codes. In Germany, Zuse had developed a computer to do calculations but it didn't survive the war.
During the war, the human computers could not keep up with the developing technology in artillery shells. They were needed to calculate ballistic tables to predict where a shell shot with a certain charge at a certain elevation, with a certain wind would land. A group of engineers at the Moore School of Engineering of the University of Pennsylvania in Philadelphia proposed using an automatic computer to do this and in 1946 they were finishing up the work on the ENIAC.
The ENIAC was a plug programmable computer with a main memory of twenty
ten-digit decimal numbers. It performed 5,000 operations per second, had 19,000
tubes, weighed thirty tons, filled a 6,000 square foot room, and used 200 KW of
power. In comparison, The 386 chip on a personal computer has the equivalent of
275,000 tubes and can still be found in computers that sit on a desk and are
plugged into the wall.
In 1949 when Mickey was 21 and thinking of what to do with his life, computers were on the verge of major changes. The transistor had been invented the year before but it would be 10 years before it would be useful. The next year (1950), EDVAC would be completed. It was the first stored program computer and had a 1,000 word main memory. The year after that (1951), the UNIVAC I would go on sale, the first commercial computer. It had the equivalent of 24,000 tubes, 1,000 word main memory, and magnetic tape secondary. Programming was in assembly language. They would make and sell forty-six of them.
Ten years later in 1959, when Mickey was 31, a lot was happening. Transistors were starting to be used extensively. This caused a revolution in computers. With vacuum tubes, the tubes often burned out and extensive down time was a standard daily event. With transistors, down time was much less. Higher level languages were starting to be used: FORTRAN, COBOL, ALGOL, and LISP. Operating systems were small but that was soon to change.
By five years later in 1964, operating systems were large and helpful. Time-sharing was common. The IBM 360/370 series came in with virtual memory. It had a tremendous main memory for the time. You may find this hard to believe but this state-of-the-art mainframe actually came with 16 megs of main memory! Of course, that is below minimum for a personal computer today.
BASIC was developed to give computing power to people who wanted the resources of a computer without having to become skilled programmers. Kemeny, at Dartmouth, one of BASIC’s inventors, stated that it was so simple that even a student could easily program with it. What he did not want to say is that they made it even simpler than that. It was so simple that even a faculty member could easily program with it. Mickey was 36 at this time.
By the time that Mickey was 40 in 1968, the PDP-8, the first really
successful minicomputer, was introduced. This made computer power much more
widely available. Many colleges bought computers at this time as they could
finally afford them. The language Pascal was developed that year. This in some
sense was the start of the discipline of computer science. It was the first
language designed to teach programming conceptually. Before that, we taught
programming and tried to add theory. After that, we taught concepts using
languages as tools—we taught computer science.
In the next decade, as Mickey edged towards 50, computer science made some very major changes. In 1971, the first microprocessor, the Intel 4004, was created. As the decade developed, 8-bit microprocessors were developed and, using them, the Apple Computer. Computing power became cheap enough that people could buy them for home use--the personal computer. When Mickey hit 50 in 1978, the first 16-bit microprocessor was developed. It was the foundation of the IBM PC line of computers. The entire history of personal computers is less than a quarter-century. That is just beyond the lifetime of most students in college today. There are very few aspects of our society that are so pervasive that have such a recent development.
Fifteen years ago, it was clear that the expectation of a computer in every home was ridiculous. Word processing was great, especially on a Mac, but unless you did writing at home—or had kids who used the computer for school—there really wasn't much that a computer had to offer. What could a computer really do for you that could not be done as well—and cheaper—without it? That was the reality then. Reality has a short “shelf life” in computer science.
In the years since then, the rapid pace of advancement has not slackened.
The personal computer that you buy today is a thirty-two-bit machine just as
the big machines used to be. The typical home computer today is so powerful
that fifteen years ago, if you were looking for a similar powered computer, you
would have to buy at least a super-minicomputer. And today’s computer will sit
on a desktop or slip into a backpack. But your home computer is not alone. You
will most likely link it into a network. Networks are becoming pervasive and
ultra useful as well as inexpensive to access from anywhere. Computers have
become gateways rather than repositories just as libraries have.
Mickey is 72 now and retired from computing. The torch has passed on and it will be passing to you. What will you see in your lifetimes? In Mickey's life, he saw an entire field start from literally nothing to be a dominant factor in our culture. He saw an accelerating growth that has not yet peaked. Fifty years from now when you look back at computers in your life, you will see similar changes. It is hard to predict the future in a field where you cannot keep up with the present but we will try.
What is possible now? In a local area network where when you need more resources, your operating system will go out into the net and negotiate with other computers for the resources. A database that gets used to the way you use it and starts prefetching data before you specifically ask for it. A database that reconfigures itself dynamically as it grows to maintain efficiency so that it is always fast to use? Intelligent agents that wander about the net in off-hours, collecting information that it has learned that you like. This would include professional journals that may no longer be actually printed.
In the software development process, coding is 15-20% of the work. That percentage will decrease. More automatic tools such as program generators will become common. The emphasis will continue to shift to analysis, design and verification as the dominant elements of programming.
Twenty-five years ago, typical programs were 50,000-100,000 lines of code. That is not much code today. A CD-ROM is a minimum requirement for loading much software and we need gigabytes of disk.
The space shuttle needs fourteen million lines of on-ground code to get it off the ground, through a mission, and land safely back on Earth. Between the Challenger disaster and the Discovery liftoff, NASA made 3,800 requirements changes. When Discovery lifted off, the programming contained 1,800 known bugs. And this is an old program in maintenance phase. The work that will be needed for the new space station or for a manned flight to Mars will easily eclipse the space shuttle program in size. And it has to work correctly the first time.
Yes, the field will change dramatically during your lifetimes. But who are the people who will change it? We expect that some of you here will be involved in that. Every software designer is a researcher as we never really do the same job twice. Each new program is a new challenge that pushes back the frontiers of the field. This is the world that you will be creating during your working lifetime.
Dr. Dennis S. Martin
Based on a talk given to UPE, the Computing Sciences Honor Society a few years
back and updated a bit.
© 2000