Q: What does "computer-based education" mean?

A: With regard to medicine, I think of computer-based education as falling into three categories: the basics; using computers to teach; and teaching students how to use medical applications of computers. The first category really means: Are computers available and do people know how to use them? In general, I've had more problems accomplishing this with faculty than with students.

Compared with other medical schools, however, I think we're in very good shape and way ahead of most. The second category really means software that takes advantage of the unique aspects of a computer to improve the educational experience. There are two problems in advancing this goal. The first is measuring "improvement." If you take a population of high achievers (i.e. AECOM students) and make one change in a large and complex curriculum, it is very hard to measure any change -- positive or negative. You can measure changes in student satisfaction, but there are those who don't consider student satisfaction an important part of the educational experience.

The second problem is that good educational software requires a lot of work and has a very limited market. To work well, a computer-aided instruction program requires the same development effort as a good word processing program. Yet instead of selling millions of copies, the programmer can expect to sell only thousands at most. The fact that there is any software at all is a result of dedicated faculty at various institutions who invested a lot of effort for the sake of education.

Even so, these individual efforts mean that when it comes time to take software from some other institution and make it available at our institution, we are likely to run into problems and in a few years when the next wave of operating system upgrades comes along, the software is likely to stop working. If an upgrade becomes available, it will be months or years behind schedule; the faculty member who originally developed it has other work to do.

Two things are changing this picture: publishing companies are beginning to get savvy about electronic publishing and the world wide web has made electronic publishing similar to paper publishing. Faculty are already very familiar with the paper publishing process so it requires much less effort to make the jump to electronic publishing. Also, the "operating system upgrade" problem goes away. Faculty can concentrate on updating the knowledge they are presenting rather than updating arcane programming language. Compared with other schools, I think we are about in the middle of the pack but gaining ground fast.

The last category refers to teaching students how to use drug databases, electronic decision support tools, and computerized patient records. Certainly we teach them how to use bibliographic databases like Medline (TM) but in the other areas we are behind. In some ways this is due to the state of affairs within our clinical affiliates. If access is difficult for practicing physicians at our clinical sites, it is even more difficult for our students. Teaching students about something they won't get a chance to use is problematic. We could be doing a better job. Students do get exposure to computerized patient records, but every site they go to has a different system and it is difficult for them to pick up what the common themes are just by exposure. Students do get access to a drug database (as part of a school site license) but they are unlikely to be able to use this at a patient's bedside or in the nursing station.

Hospitals are starting to experiment with providing internet access to other medical resources but they are moving cautiously because of patient confidentiality and security concerns.

Q:What impact has the student housing network had on computer-based education here?

A: I can point to several things demonstrating that it has already had a positive impact.

A year ago, it was difficult to convince faculty to place material on the school's web server. When I succeeded in convincing them, we would have complaints from students because there would suddenly be overcrowding at the on campus computers -- especially the night before an exam.

This year (1999) the over crowding has disappeared. At the same time, faculty have become more open to the idea of placing material on the web. Compared to this time last year we have doubled the number of courses that are providing color images via the web. The only alternative way to provide these images is with microfiche.

Microfiche are arguably more portable, but they lose out in the area of cost and turn around time. If a faculty member finds a good image half-way through the course, we can provide that on the web within a few days. With microfiche, you have to wait till next year. The student housing network will also have a more subtle effect. I believe it will indirectly increase student computer literacy.

The fact is that computers and software are immensely complex things. There's no getting away from that.

A computer science saying comes to mind: "If you create a fool-proof system, only a fool will use it."

I think computers will become easier to use as software programmers and human interface designers learn more about creating intuitive programs but there will always be complexity on the cutting edge.

For basic things, it is possible to provide advanced training, but for the more complex tasks, no amount of training will make some sequence of operations stick until the student is actually ready to use them.

One sequence in particular that is important to learn is figuring out how to do what you want to do on the computer. If you are sitting in a computer room and can ask the person next to you how to do something, you will promptly forget it and have to ask again the next time. If it's 2:00 am and you have to figure it out yourself, you may make mistakes and it will certainly take longer, but not only will you remember how to do this one task, you'll have a better idea how to navigate the often obtuse computer manuals that come with software these days. Unfortunately, this line of reasoning doesn't appeal to my wife.

Q: It sounds like you advocate a "sink-or-swim" philosophy?

A: Not really. I think of it more as an aspect of human nature. No matter how many times you tell someone to make back-up copies of important computer files, no one (including myself) ever really learns this until they lose a couple of weeks worth of work because of a bad floppy disk.

Q: So if students have a computer problems, they should just think of them as learning experiences?

A: Yes and no.

Yes, it's a learning experience but sometimes it isn't the student who needs to learn the lesson. Take the student housing network, for example. Were there problems? Absolutely. Were they the student's fault? Absolutely not. Were they avoidable problems? Again, the answer is absolutely not. This was a huge project which in some ways was more difficult to carry off than the original network wiring of the campus research buildings.

When the campus was originally wired, it was done because the school had a long-range vision about the importance of the Internet. But it was also done in a relatively leisurely manner. There were a handful of faculty who wanted the connection "immediately" but there was no specific deadline.

For the student housing, not only did all the work need to be done within a few short months, but at the end of all the wiring work we could expect several hundred new users within a short period of time. Perhaps some decisions should have been made differently but once made, I think it would have been a mistake to try to change things in mid-stream.

I have a good imagination and I think there were a lot of problems we did avoid that would have been worse than the ones we went though. I also predict next year will go much more smoothly.

Q: Can you talk a bit about the new requirement that all incoming AECOM students purchase a computer? Is this the right move at the right time?

A: Yes, I think it is important now.

At the moment, it is needed to convince faculty to move in this direction.

By next year, I believe students who don't have a computer will be at a real disadvantage because of the extensive faculty reliance on them. Eventually, the requirement will become superfluous -- it will be like requiring students to own a telephone. At that point, I'll be out of a job.

Q: So are computers the solution or the problem?

A: It's a cliche, but computers are just tools.

I have students and faculty who come to me and are very enthusiastic about whether we can do X or Y with our campus network. It always sounds simple but I'm very aware of the hidden complexity.

It brings to mind another computer science saying: "To err is human; to really foul things up requires a computer."

Nor is the complexity all a matter of programming. Sometimes it has to do with the sociologic interactions between students and teachers. I heard a complaint from one faculty member who was disturbed by the fact that a student sent him an E-mail question at 2:00 am.

Other faculty members worry about appearing foolish by trying new technology in front of students who may be more savvy than they. Students, too, have problems adjusting to change. It may be hard to believe, but I think compared with past years, we have cut down on the deluge of paper information we send to students. But there is a growing deluge of electronic information being generated. It requires different skills to sort through electronic garbage. You can't use a high-lighter to mark the important passages.

From the class mailing lists alone, students are getting an average of four messages a day (more for years 1 and 2 and fewer for years 3 and 4). Some students are grumbling that this contains a lot of junk mail. The result is students are skipping everything and occasionally missing important announcements.

Students need to learn when to hit the delete key, when to scan a message, and when to read it in detail.

The problem will only get worse the longer they use E-mail and the further they progress in their careers. I get about 60 messages a day, and that includes weekends. I dread Monday mornings.

Just for the record, about one-third of students' E-mail is coming from faculty and the rest from fellow students.

At the same time, it isn't all bad. We are just starting to take advantage of things that are unique to computers. I already mentioned that the web provides a way to provide images quickly which can't be done any other way. This has other implications. For example, in case conferences with progressive disclosure (i.e. you have to think about Part I before you get to see Part II), we can provide new images (e.g. the results of a chest x-ray) at a specific time and an automatic way. It isn't dependent on someone physically handing out new pictures.

We've just completed our first on-line questionnaire on nutrition awareness in the Class of 2002. The results will be part of a lecture being given in February, but it is a short step from on line collection to processing the data and giving immediate feedback. My hope is that this will lead to offering on-line practice exams for students. When access improves at our affiliate hospitals, I foresee clerkships at dispersed sites having real-time case conferences. This would provide a more uniform clerkship experience across sites without all the tedious mucking about in taxis and buses. In some ways these all sound like simple things -- not the kinds of things most people think of when they think of "intelligent computers."

But it's just these kinds of simple ubiquitous uses that have the most impact. When someone asks what impact computers have had on society, I point to the automatic teller machine (ATM), a computer most people use without ever thinking of it as a computer.

Q: When and where did your fascination with computers begin?

A: I first became interested in computers in high school.

I think there are lots of kids who are fascinated by raw power and can imagine moving their arm to move a lever that moves a hydraulic piston that moves a backhoe that crushes a building or a rock or something. That kind of fascination probably has something to do with adolescence.

Anyway, I was fascinated by the idea of writing a mathematical equation that generated a sequence of logic that resulted in a bunch of calculations that printed a whole bunch of results that I could conceivably replicate only if I had a million years of free time. And of course the results had something to do with playing 3-D tic-tac-toe or something equally useful. Computers are also intriguing to me because if you understand how they work, they embody order and logic. It makes a nice escape from the real world. By college, I was interested in comparing the order and logic of computers to things in the real world. And how they could be used to better leverage human thought. This naturally leads to trying to understand what human thought is so you can program it into a computer. That's what my interest in psychology, neurology, and education grew out of.

Q: Is your goal to make computers intelligent or make computers better teachers?

A: Both of course. A computer (or a teacher) can't really teach until it (or she) understands the student and understands what the student's misunderstanding is.

In a limited way, we can make computers do that, but it's a lot of work. I'm interested in tackling one aspect of this: using semantic networks to represent knowledge. If a computer has a flexible way of representing knowledge, it can begin to generate questions that test the student's method of representing knowledge. When it finds a place where the student's ideas don't match the computer's, it can explore this in more depth. Maybe the student will learn something. Or maybe the computer will.

Faculty won't be out of a job for a while. It will be even longer before computers can act as role models.

Q: Do you believe that eventually computers will be "intelligent?"

A: There is an essentially "religious" debate about whether computers will ever be truly intelligent.

The arguments on both sides are fascinating, but like many philosophical arguments I think it ultimately comes down to what you believe. I believe eventually it will happen but eventually is a long time in the future.

Computer scientists joke that "artificial intelligence" is everything we cannot yet get a computer to do. Once they get a computer to do something, it no longer becomes part of what defines artificial intelligence. Classical examples include chess, problem analysis, and goal planning. There are now examples of these kinds of problems where computers surpass most, if not all, humans yet no one thinks of them as intelligent.

Alan Turing devised a test in which a person enters a room with a computer terminal. The person types into the terminal and gets responses back. At the end of an hour if he or she can't tell whether or not the thing they were communicating with was a human, then whatever it was it passed the test for human intelligence.

The Boston Computer Museum now holds an annual contest in which judges converse with "entities" about a specific subject. The entities might be computer programs or humans. To win, a computer program has to be rated as "human" by more than half the judges. So far, no computer has won, but there was one year when the topic was Shakespeare that a Shakespearian scholar was rated a "computer" by more than half the judges.

Q: So am I conversing by E-mail with a human or a computer?

A: You be the judge. I won't be insulted either way.