An Interview with Peter Denning
by Charlotte Thomas
Editor, Graduating Engineer
November 1996
In November 1996, Charlotte Thomas interviewed Peter Denning in preparation for an article in Graduating Engineer, a magazine distributed to seniors expecting to graduate from engineering school within a year. Excerpts from the interview were published in the spring issue. Here is the complete interview.
CT: Given the changes that technology has made in our culture, how has your definition of the new engineer changed since 1994?
Denning: The discussion of the New Engineer is about who we will need to be to succeed in the world and workplace of the 21st century. As part of this discussion, we have emphasized human communication and relationship skills. But this should not obscure the importance of practical engineering competence. People will expect enginers to solve problems in a professional, human way.
CT: Hasn't this always been so?
Denning: The engineer gets plenty of theoretical training in school. We need theory but we also need the ability to solve practical problems. As an example, think about the ubiquitous computer. Along with telephones and faxes, computer are being connected into a highly complex network. Many people rely on that network. To keep it working and to improve it constantly, we need technicians and computer engineers. They help set up systems, configure them, and troubleshoot them. In the future, when people want to make their computers to work, they will look for engineers with practical skills. Right now it's easy to go through a computer engineering program just concentrating on theory and programming. You're not guaranteed much practical experience to go with the theory.
But I'm not talking about substituting practical skills for theoretical knowledge. That would be going much too far. Today you can be certified as Novell network engineer. This is very practical and professionally valuable. But you are certified only in Novell systems. You may not be of much help to someone with a Unix system networking problem. What's missing for the certified Novell engineer is a larger, systems point of view. Without a systems view, it is difficult to take practical knowledge from one domain into a new one.
Engineers need the ability to think at the systems level and solve practical problem for clients. Students must include the practical aspects in their training. Students in information technology tend to get less engineering practice in their training than students of the traditional engineering disciplines. Some students get the engineering practice from lab work, some get it from coop programs, and some get it from senior design projects. But it's quite variable depending on the school and program.
I think an important element of engineering practice is learning to deal with real customers or clients. So, in my mind, the most educational senior design projects are those with an outside customer, usually from industry. This helps them to learn how to fit engineering theory and practice into the concerns of customers --- such as deadlines, upward compatibility with existing systems, The more valuable senior design project is one with outside an customer for a project. Students shouldn't settle for a faculty member being a customer. They should request a real customer from the outside.
"I know the perfect database system we could use to solve your problem," says the student. The customer replies, "We have a legacy database. We have to use it. We can't change it. It's not an option." The student might say, "Yes, but this database system really is perfect for your situation." And the manager insists, "No, we have to stick with the legacy system. That's that." Students can and will encounter this sort of reality in projects with an outside customer.
The other thing students encounter in projects with outside customers is an emphasis on systems integration. Real customers want to use as many existing components as possible to solve the problem. They can't afford to start from scratch.
Sad to say, many computer science and engineering departments don't require senior design projects. The projects students get are mostly programming projects. I would like to see them all do it. I advise students to seek out senior design projects even if they are not required.
CT: engineers used to deal with numbers in order to solve Problems. Now they are asked to deal with people. Why the Change?
Denning: We live in an age of cartoons. Scott Adams, creator of Dilbert, invented class warfare between engineers and managers. Through Dilbert, he celebrates the nerd. But this is barren. Engineers are human beings. Without skills in relationships and communication, they will flounder, eventually winding up in the kinds of dysfunctional relationships so artfully depicted by Adams. I think this is something educators ought to worry about. I think students ought to worry about it even if their professors aren't. Otherwise they'll end up in Dilbert's workplace and hate it, not realizing they have the power to change it.
Many student engineers already understand that most of what they are going to experience in the workplace will occur in relationships with other people. Even the mundane matter of solving problems is a set of relationships among the client, the engineer, and technical people who can help. The central practice of engineering --- the project --- is a team that lives or dies on the quality of its relationships. Adams celebrates the individual nerd, who ends up at war with management, is endlessly cynical, and believes his boss to be "clueless" --- even while he himself is clueless about his own inability to sustain satisfying relationships. In the modern information age, it's easy to get wrapped up in programming, in computer-aided design, in e-mail, in net-surfing, and a slew of other individual technical activities --- and forget to take care of your relationships.
I've seen students who get so wrapped up in their own stories that they come to believe that everybody else is messed up except themselves. In the extreme cases, these students tell professors they are out of touch with the real world and have no right to tell them what's right or what's wrong. They will definitely have problems in the workplace. The other students should not look upon them as role models. I emphasize that this is an extreme. Most students I know are more aware of the importance of relationships than are many of their teachers. Most professors are not models of people who understand how to make teams work or how to manage teams. The academy celebrates individual performance. It celebrates the star. Teams are not rewarded. This is a different style from the real engineering The team is a fixture of the workplace. Even Scott Adams recognizes this when he depicts broken attempts at review meetings where everybody wonders how a project gets done at all.
Another dimension to human skills is your professional identity. Your professional identity is the set of stories that others tell about you when you're not around. It determines what doors will be open for you, and what doors will be closed. If you are seen as an untrustworthy person, how many jobs do you think you'll get? If you have the reputation of doing good work, but never on time, do you think you will get jobs at firms that value keeping time promises to customers? What if you are known for being argumentative and putting people down but you're technically very good? You'll be selected for one-person projects, but never for teams. These are simple examples of how opportunities can pass you by if you have a negative professional identity.
It's easy to confuse your professional identity with the identity you have with friends. The truth is, we have many identities. Your personal identity is the way you see yourself in the privacy of your own mind. Your family identity is the way your family sees you. Your work-associates identity is what your colleages at work will tell you about yourself if you ask them. Your professional identity is the toughest of all to observe, because you're not around in the hallway when others are discussing you, or in a meeting deciding who should be on the team. The conversations you don't get to participate in affect you. If you widely trusted, have high credibility, high integrity, get your work done on time, and strive for high quality, you'll get selected for the projects that matter to you. You'll get promotions.
Students don't think about these matters and learn how to observe and affect their identities. We faculty don't talk to our students about such matters. We should.
In addition to the two human dimensions I've mentioned --- relationships and professional identity --- there is a third dimension to real workplaces that we don't teach our students: power. A lot of people react negatively to power when others have it, and yet they want more power for themselves --- witness the yearning for "empowerment". What is power? Power is the capacity to influence action. People who have power make things happen. Powerless people can't make much happen at all.
There are different kinds of power. Financial power means I can influence what happens by paying for it or by backing it with capital. Coercive power means I can influence what happens by force or threat of force, such as with the police or the military. Production power is the capacity to manufacture large numbers of items. The power of office is the set of authorities that come with an office, such as the President of a company. Symbolic or personal capital comes to certain people who have great stature and special respect --- for example, the President of South Africa has such great moral authority that he is able to keep factions from warring with one another. Social power is accumulated by organizations through their social networks, giving them the ability to influence events by exercising their contacts and speaking for many people. (People pay high prices to go to Harvard, not because the education is better, but because of the immense social power of the Harvard Alumni network.)
I'm suggesting that engineers don't appreciate these different kinds of power. They think that power is most coercive and will corrupt. Consequently, many of them don't develop much power and wind up unhappy. Look at Andy Grove, chairman of Intel. He is an engineer who has come to understand social and production power. He leads a huge organization and keeps them at the forefront of their market. In his book, Only the Paranoid Survive, Grove attributes his understanding of power to a set of practices he developed over the years for worrying about the marketplace and the competition.
Those who don't understand power often resent those who do. They say they are victimized by the power elite. They hide their envy by criticizing and carping. How often have you reacted to a complainer with the thought that this person has deprived himself of power by imagining himself to be the victim?
CT: But wait, hasn't downsizing justifiably created this paranoia in engineers?
Denning: Well, it's certainly popular to say that corporate executives live off fat salaries earned by terminating the jobs of many people. The stories circulating make the problem sound a lot bigger than it is. Unemployment is now lower than at any time in the past two decades. The 1996 Bureau of Labor statistics show that the number of jobs lost to downsizing is around 4%, and that most of those people get other jobs fairly quickly. The hiring market for engineers is very hot, especially for engineers who understand information technology.
What then is the concern? I think it is job security. Most engineers cannot count on working for a company for the next 40 years. They were brought up by parents who experienced job security but find themselves in a workplace where mobility is the norm.
I have a student who worked for a defense contractor. He found himself being moved from one narrow task to another. He began to sense that there would come a time at which the DoD would no longer be interested in his specialty. At age 48 he enrolled in our Master of Science program. Within the year he was given a pink slip and went into a long search for a job that would use his skills. He eventually changed to new areas thanks to his Masters Degree and is now satisfied with his career. He was luckier than others, the ones who keep their nose to the grindstone and don't keep themselves from becoming obsolete.
CT: Aren't engineers aware that they have to keep up to date?
Denning: Sure, but many engineers think that keeping up to date means staying abreast of technical developments. There's much more to it than that. For one thing, your social network is very important. Your well-informed friends can help you spot little ripples in your field well before they've become mass currents. Your friends can get you contacts in times of need. Their many eyes and ears augment your own. Those who celebrate nerdhood often don't take care of their relationships and are taken by surprise when their obsolescence overtakes them. Then they become the victims of circumstances they might have avoided by taking care of their social networks as well as taking courses and seminars.
Stay proactive. Stay on top. Don't settle for just reading. Make sure you are acquiring skills that are valuable. Nurture your relationships.
CT: What is the knowledge society and why do engineers have to deal with it?
Denning: This term is used to describe a change over the past half century from labor-intensive to conversation-intensive work, from hands to brain. Today about 80% of the jobs are in white-color, service work; only 20% are in manufacturing and other forms of labor. Most work today depends on knowledge and intellect, not muscle. The person standing before a workstation may not be moving much, but nonetheless can be engaging in significant transactions that affect many people's lives.
Look at what we mean by vacation in the knowledge society. The assembly- line worker goes away and someone takes his place; when he returns he resumes his place and there is no backlog. When the knowledge worker goes on vacation, no one can fill her place; when she returns, she has to catch up on many things that accumulated in her absence. Because of your special knowledge and familiarity with the project, you're a key person who can't be replaced easily.
CT: Doesn't this have to do with being a generalist of specialist?
Denning: That's a false dichotomy. If you don't have engineering competence at a deep skill level and knowledge you're dead. The challenge is to become a very good specialist at things that you have to be, without losing sight of the larger picture. You shouldn't try to choose between generalist and specialist. You must be both to be valuable.
Specialists have skills that are often quite valuable; their biggest challenge is to learn new skills as the technologies change, so that they remain specialists in current technology. The generalist has a bigger challenge: avoiding being branded as a dilettante who can't solve real problems.
CT: Is Sense 21 still in existence? What kind of impact has it made? Since 1994, what have you added to your ideas about new sense?
Denning: Sense 21 is shorthand for "a new engineering common sense for the 21st century." Before we can talk about a new common sense, we must discuss the current common sense about engineering. The term "common sense" means a set beliefs, values, premises, suppositions, myths, superstitions, rules, commandments, conventional wisdom, standard practices, habits, and the like that are shared by everyone, often tacitly.
CT: What do your students say when you ask them what their sense of engineering is?
Denning: One of the first things they say is that engineering is problem solving, mostly in the context of systems. They think of systems being built to satisfy requirements and specifications given by their customers or clients. They think an engineer is a professional at these matters with a code of ethics.
But they move quickly from this conventional wisdom to their concerns about career and jobs. They believe there is a lot of uncertainty in the marketplace --- uncertainty about which of many alternatives available to them while building a system is the best and uncertainty about their own positions and employability. They don't understand the phenomena behind the uncertainty --- social relationships, professional identity, power, and design --- and therefore they don't know how to take effective action.
The engineer of the 21st century (which by the way is just a couple of years off) must shift attention from requirements and specifications to design. The requirements and specifications are not going to go away, but in their rigidity they do not deal with flux, uncertainty, and shifting expectations. Engineers will need designs that adapt and evolve easily. This kind of design is not simply a system with good feedbacks in it. The skill of design has a big dimension o flistening --- listening to the concerns of the client and finding processes and systems that help the client take care of those concerns. The listening engineer will often find simpler solutions for a client's problems than an engineer focused on the wonders of the modern technology.
You know the story of Mitsubishi and Sony, who had competing approaches to video taping some years ago (VHS and Beta, respectively)? Even today, most engineers say that Beta is technically a better system; its error-correcting codes defeat many distortions arising from copying or noisy receptions. Why is it that the VHS system became the market standard? Mitsubishi understood marketing, social networking, and business relationships better; its design included its market strategy. They won.
Something similar happened with Microsoft and DOS. DOS is universally acknowledged as one of the most primitive and difficult operating systems. So how come DOS became the standard for personal computers? Many engineers can't understand what happened. What happened was that Bill Gates made a deal with IBM to include DOS in every IBM PC --- and he retained rights to sell DOS to other computer makers. It didn't matter to the customers of PC that their computers had DOS and not VMS, OS360, or Unix. It mattered only that their PCs worked. People became dependent on DOS, and nearly every PC operating system since then has had to be compatible with DOS conventions. Bill Gates understood that savvy deals must be part of his overall design. He was able to take care of the concerns of a lot of people with an operating system that was no match technically with most other operating systems. Bill gates exemplifies traits that the new engineer must master. Andy Grove of Intel is another engineer of this kind. He saw his company's purpose as listening to the customer and building systems that help them with their concerns --- not just systems that meet technical specifications.
Sense 21 is about learning this. Students quickly understand it at an intellectual level. The hard problem is to teach them to think and practice this way. Students tell me that the Sense 21 course opens up their minds. But I have found they quickly slide back into the more traditional habits of engineering unless they have a structure to help them maintain their ability to think this way. So I formed the Sense 21 club as a way for them to keep in touch with interested, like-minded colleagues. About half the graduates of my course, which started in 1993, have remained active in that club. It's quite remarkable, given that there are so many demands on their time pulling them away.
I have warned the Sense 21 members to maintain their technical skills because otherwise they will appear to the traditional engineer as a soft generalist, and will be quietly but firmly ostracized from the social network. In the new engineering common sense, students can be both generalist and specialist.
Eventually, I'd like to see the precepts of Sense 21 become standard practices supporting the engineering curriculum. I think this will eventually happen because the demands of the marketplace are too strong for hundred-year-old engineering traditions to resist.