Harry L. Shipman
Updated April 2000
This document has an interesting history. Some of the text originally appeared in - of all things - a University of Delaware committee report, written ten years ago in 1990. I was chair of the committee. I'm including stuff from 10 years ago for two reasons:
I've interspersed some "Y2K updates," in a different type font, which update the predictions at the present time.
The overall goal of this computing initiative, originally written in 1990, is both simple and ambitious:
Five years from now, every student at the University of Delaware should be competent in computing. A student should be familiar with the way that computers are used in his or her major field, as handlers of library information, and in the working worlds to which this major is coupled.
Faculty committees at Delaware have been calling for widespread student computing competence since 1983 if not before then. It has not come to pass. If our students are to become leaders in the 21st century, they must know what this technology can do. Computers and the closely related area of electronic communications have already become central to many enterprises. For example, leaders of multinational corporations need to understand this technology's potentials and limitations if they are to use it to create and manage far-flung global companies. A person's ability to use the computer as an information management and information search tool is critical here. As another example, creators of artistic works ranging from much of modern music to technical cinematic breakthroughs like "Roger Rabbit" can only use technology to extend their creative genius if they know what the technology can do. On a more munbdane level, our recent graphic arts graduates tell us that in the area of commercial art, the Macintosh has virtually replaced pens, paints, and palettes. Journalists and technical writers say the same thing. At this point in time, a student with computer competence has a competitive advantage, but in a very short time computer competence will mean basic survival. All of our students, not just our engineers and computer scientists, need to be competent in computing in order to thrive in the 21st century.
Y2K Update: In 1990, neither the other authors of this report nor I predicted the explosion in electronic commerce which has been led by companies like amazon.com. The internet was barely on the technological horizon in 1990. However, the point we made is even more true now than it was ten years ago: Every student who graduates from a university needs to be competent in computing.
This discussion leads to a second question:
(Continuing Y2K update) Every profession is going to be affected by computing, as some of the details later in this document show. Teaching, journalism, finance, music, art, . . . all are affected by computing. And in all of these businesses, you are ultimately the person who is in charge of your career. You have a choice -- to be an ostrich or a tiger.
The ostriches ignore the changes that take place around them, just like real ostriches who think they can hide from the rest of the world by burying their heads in the sand. I have known quite a few ostriches. The lucky ones have kept their jobs but are very unhappy people. They hoped that the wave of change would subside, but it hasn't. They are counting the days until they retire, even if there are a large number of days like 10,000. Their co-workers and students suffer along with them. The unlucky ostriches lost their jobs and, because of their refusal to deal with anything new, have difficulty finding new ones.
The tigers accept the changes that take place around them. Some tigers not spend hours and hours learning more and more about the technology and worrying about whether it is best to buy a computer with a Jaz drive or buy a CD-burner. Most don't. They enjoy their jobs and learn enough about the new technologies so that they are aware of how their working lives will be affected and improved by them. When they see something that works, they adopt it. They realize that computing will keep changing, and take the initiative so that their working and personal lives are enriched as a result.
Which will you be - an ostrich or a tiger? A tiger, I hope. The rest of this page contains some forecasts of where computing is going.
Software will continue to mature, with more features and greater complexity. The greatest constraint in future uses of computing will be software. Since custom-developed software is quite expensive, it often only possible to use a computer for some application if someone else has developed the appropriate software. The price of more features and greater user-friendliness is that the software voraciously gobbles computer resources. For example, WordPerfect 5.1 basically requires a hard disk and 640K of memory to run decently. The same software with somewhat fewer features, sold four years ago as WordPerfect 4.0, could work with 256K of memory and a two floppy disk machine. Graphical user interfaces place even more demands on machine capabilities.
Y2K update: The prediction is right on target. The graphical user interface, where the user clicks on icons rather than using textual commands, has become universally accepted in computing. New software does require increasing amounts of computer resources. These resources are becoming available, since the cost of computing resources has followed a pattern called "Moore's Law" for at least 30 years. Moore's Law postulates that the cost of virtually anything associated with computing drops by a factor of 2 in a period of 18 months - for example, someone can buy a machine now which has 128 megabytes of memory for $1700, and in 18 months you can spend $1700 and buy a machine which has 256 megabytes of memory. If all you want to do is connect to the internet, a few years ago it took a pretty capable PC to run internet browsers like Netscape, and pretty soon machines which hook up to your TV will become widely available for $500 or less. Software will continue to change, becoming more and more capable.
Increased computer capabilities results in our doing different things with computers, not just doing the same thing faster. Roughly speaking, an 0.1 to 1 MIP machine, the PC of 1990, permits one to control words. (A MIP is a million instructions per second, one imperfect measure of computational speed.) A 100 MIP machine, which was a supercomputer in 1990, permits one to control pictures and sound. A 1000 MIP machine may permit voice recognition and machine vision.
Y2K update: Again I see no reason to update the prediction, even though 10 years have passed. The use of pictures in text generated by ordinary PC's has now become pretty routine. Some web pages have little animations on them; and some more sophisticated ones permit the user to have access to full-motion video, though you usually have to download some pretty large files to get the video. Voice recognition is available in some cellular phones. Computers which understand human speech are on the market but their use still seems to be limited, and I'm not really sure why. All of these applications require lots of memory.
Ten years from now, we will be doing very different things with computers. Blockbuster Video could well be out of business - you simply download your movies over the net. Change will occur very quickly.
The increased interest in very large databases and increased use of pictorial material will result in increasing storage needs - perhaps as much as a factor of ten in the next five years. A centrally run "disk farm" can play a variety of roles. The U.S. Census, for example, can be stored centrally in a university, and subsets of the census database can be downloaded to individual users. When users work with graphical information, the old idea of backing up on floppy disks, barely practical when one works with text, becomes completely out of the question. Backup copies of individual user data may well be most cheaply kept on a centrally run disk farm.
Y2K update: While we were right about the need for more memory, we were absolutely wrong in the way it was to be provided. We did not forsee the development of high-capacity, cheap storage devices that people could easily attach to PC's. A few years ago, the capacity of floppy disks increased enormously when Zip drives appeared, drives that could store 100 MB of information on something the size of the standard floppy. A year or so after the Zip drive was introduced, the same company (Iomega) came out with the Jaz drive, which holds 1-2 GB (recall 1 GB = 1000 MB). In the past year CD burners have become cheap enough for individual users, and with those hundreds of GB can be stored on one CD. Nobody needs a central facility at a university or company to store or back up data. Most users will have access to the memory that they need on their desktops.
Thus, while the cost of hardware decreases, we do not end up spending less on computing. Because software becomes more bloated and because of the changing nature of what we do, requirements for even minimal levels of computing power increase at almost the same rate as the unit cost of hardware decreases. Thus the net cost can remain the same.
Y2K update: This prediction was really pretty accurate. Let me be rather specific. The first PC which I bought, in the 1980s, was a state of the art machine and it cost me $2000. The next PC which I will buy, in the summer of 2000, will be a similar upper-level but not leading edge machine and will still cost me $2000. The two machines are enormously different in memory and speed. But the computer that's needed to process words has become much more powerful.
Ten years from now, I think that a basic PC will cost roughly the same as it does now - between $700 and $2000. But what is beginning to change, and will change even more, is that a range of cheaper machines will appear. These machines will not do everything that a PC can do, but they can do simpler things. Palm Pilots and personal organizers are much cheaper than PC's - whether the software developers can figure out sensible uses for those things remains to be seen. (My friends are split on electronic day-timers, and I know one person who used one for a while and gave it up.)
And while a user can try to stop change and refuse to upgrade, this user will find that she or he is unable to open files which are sent by team members which use a more capable word processor. The word "bloated" to describe more advanced software was a bit controversial among our group, but I think there is some truth to the statement that version 8 of my friendly, non-Microsoft word processor (Word Perfect) doesn't do an awful lot more than the version which I started with, which was version 4.0. Ten years later, it requires 100 times as much memory. The next ten years will be the same as the last ten - software will need more memory and more speed. Thanks to Moore's Law, the hardware capabilities will be there.
Access to national networks will become an increasingly normal part of the way that the research community does its business. Editors of conference proceedings in a number of fields are beginning to ask for the electronic submission of manuscripts. A number of national agencies are asking that proposals be sent over the network. Researchers will use data stored in national databases and will exchange programs with each other.
Y2K update: This prediction seemed radical 10 years ago and if anything it was too timid. The scientific research community did exactly what we said it would do. I will edit a conference proceedings this summer and I will collect the manuscripts electronically. When I submit proposals to the National Science Foundation, I do it electronically - the paper option disappeared a few years ago.
But we were only thinking of the research community. Neither we nor anyone else predicted that everyone else would be using the internet too. No one predicted that amazon.com would be one of the most explosively growing companies in the late 1990s, or that aol.com would be on the Fortune 500 list of companies with the greatest sales in the U.S.
Someone I know who works for a travel agency predicted that in 5 years all airplane tickets would be sold over the internet, and that travel agencies like the University Travel office in the Trabant Center would be obsolete. The future will tell if she's right.
Training users so they can take advantage of these technological changes will be critical. The University must give students enough background so that they can thrive in this technological environment, or at least exist in it with a minimum of discomfort. Faculty and staff need to keep up to date.
Y2K update: Some units in the University have understood the need for computer training. The School of Education, which provided hardly any computer training in 1990, now does one of the best jobs of any unit in the university in training future elementary teachers in the use of technology. A number of other programs like accounting do a similar job. However, there are still some places in the University where computing is not important enough to majors. Faculty may have kept up to date in some senses - you can't survive in this university without e-mail - but some curricula still don't include enough of it.
However, you as a student have an opportunity. The University has many training opportunities and workshops which you can attend. I hardly ever see students at the workshops run by IT User Services. Once you leave the university, you can't simply take a free course in how to animate web pages - someone will have to pay a good deal for it, or you'll have to teach yourself.
Comments on this page are most welcome.
April 3, 2000
Harrry Shipman