acm - an acm publication

Articles

A conversation with Jef Raskin

Ubiquity, Volume 2003 Issue July, July 1 - July 31, 2003 | BY Ubiquity staff 

|

Full citation in the ACM Digital Library


Jef Raskin created the Macintosh computer project (naming the computer for his favorite variety of apple) and is the author of "The Humane Interface" (Addison-Wesley, 2000). He is also, among a great many other things, musician, mathematician, professor, and ... well, read the interview.

UBIQUITY: What are you doing these days that you�re most excited about?

RASKIN: "The Humane Environment", which we call "THE" from its initial letters.

UBIQUITY: Could you elaborate?

RASKIN: What if we could use computers without all the little annoyances that constantly irk us? What if you applied common sense to how computers and PDAs and cellular phones worked, and let the simple things be simple again without losing all the richness and variety of tasks you can perform with them?

UBIQUITY: That sounds fine, of course, but can it be done? And if so, why hasn't it been done?

RASKIN: I think that THE shows that it can be done. An interesting part of THE is available online for free downloading ( www.jefraskin.com) and development is proceeding. The implemented part shows that text creation and editing, calculation and programming can be made much easier. Why hasn't it been done? Because the industry is stuck in the GUI (graphic user interface) paradigm. For many people, the current interface style embodies computers and other devices, and they can't imagine anything else.

UBIQUITY: We will get back to THE, but can you tell us first something about your educational background?

RASKIN: I received a BA in math and a BS in physics from the State University of New York at Stony Brook; my minors were philosophy and music. I studied for a PhD in mathematical logic at the Pennsylvania State University. However, I became persona non grata with my advisor by using a computer.

UBIQUITY: Really!

RASKIN: I was discussing models of learning with a professor and I said, "Here's a possible model of learning," and he said, "No, that's not a possible model of learning." So, I wrote a computer program that exhibited the behavior that we agreed upon was learning, but which used my proposed mechanism. One of the nice things about computers is that, unlike minds, you can look inside the black box. But he wouldn't even look at the computer output. I felt like a Galileo with the church fathers refusing to peer through the telescope. And so I switched from philosophy to computer science.

UBIQUITY: And then?

RASKIN: I completed my thesis but Penn State did not get its expected accreditation for a PhD in computer science, which left me high and dry. So I took a Master's degree in computer science, which is all they could offer at the time, and I left.

UBIQUITY: Did you ever bother getting a PhD?

RASKIN: I made one more try. I went to University of California at San Diego and I became a music graduate student.

UBIQUITY: Doing what?

RASKIN: Mostly composition, performing, and conducting. I've been alternating between a musical career -- I play keyboard instruments and early winds -- and a computer career for many years. I completed my studies there and became a professor in the Visual Arts Department.

UBIQUITY: After studying music? How did you become an art professor?

RASKIN: Well, I was doing a lot of art at the time. My work was exhibited in the Museum of Modern Art in New York, the LA County Museum, the Brooklyn Museum, and a number of galleries. I was one of the artists representing the United States at the Edinburgh Festival of the Arts. This created an embarrassing situation: a music graduate student who was doing more art shows than most of the faculty of the Art Department, so I was hired as a professor. I taught there for five years. The Regents of the University of California said that you're supposed to be able to teach in one department while getting a PhD in another. However, in my case they decided that music and art were too closely allied, and so once again I completed a course of study and I didn't get a PhD. It was very frustrating.

UBIQUITY: It sounds like a catch-22.

RASKIN: Sure does. I should have known; the very first day my mother took me to sign up for kindergarten, we were directed to an office that was at the top of the stairs. We went up, turned a corner and took another flight up, then another corner and as we started up the last flight, we noticed that it ended at a blank wall. As I said, I should have known. Finally I figured: I'm already a professor, so I did not return to graduate school. That's my formal educational background. Mostly on my own I learned how to design and build electronics and how to work with a machine shop; useful skills for an inventor. More recently, in order to better understand how to design human-machine interfaces, I spent a number of years studying cognitive psychology.

UBIQUITY: What do you think of it, looking back on it now?

RASKIN: I have a great deal of fun and learn a lot.

UBIQUITY: What kinds of art courses did you teach?

RASKIN: I taught the usual introduction to Western art. I taught photography and computer animation. I had been doing computer graphics for many years, and computer art was really hot and in the news. I was often on TV or doing something newsworthy by using computers. This was before personal computers made such activities commonplace. Also, I carried around a computer in the back of my truck.

UBIQUITY: What kind of computer was it?

RASKIN: It was a Data General Nova. Most people didn't run around with computers in the mid- to late-'70s. I also taught courses such as computer programming for people in the arts and humanities and music theory for engineers and scientists in the Music department. I was the school's most interdisciplinary professor. I even taught a course in the Phys Ed Department, since I was a bicycle racer in those days.

UBIQUITY: That's quite a range of activities to be so proficient in. Were you looked upon as a marvel back in those days?

RASKIN: People have often expressed surprised at the range of things I do.

UBIQUITY: Do you still interweave all these different things, or are you now more of a specialist than you used to be?

RASKIN: I have cut down on the number of instruments that I play to piano and pipe organ and the recorder family, and strive for a decent level of technique. I have limited the kinds of remote-control airplanes I build and fly to those that are very small and electric-powered (though I did build a large one to do remote sensing and surveillance for the U.S. Department of Agriculture a year ago -- my one stint as a contractor to the government).

I am no longer an active amateur radio operator as I used to be but I have not been able to abandon my love for aerodynamics. I enjoy writing and I have all kinds of various oddities scattered around my office. Nothing of monetary value, but fun or of interest, for example, a paper air-powered engine and a little black square that floats over four gold cubes ... and which puzzles almost everyone. It's great fun to see people who know physics stumped by it. I've always wanted something that would float in the air forever, not using any power, and now I have it.

UBIQUITY: There you go. Answer a trivia question for us. Why do you spell your first name Jef rather than Jeff?

RASKIN: Jef is my official legal name, it's not a nickname.

UBIQUITY: Have you ever seen anybody else have that as an official name?

RASKIN: Yes, I'm not the only one.

UBIQUITY: But you're probably the only one known as the creator of the Macintosh computer. Tell the story. What does it mean to have created the Macintosh computer? What did you create?

RASKIN: Well, when I joined Apple, I had the concept in my head of a computer that would be purchased at reasonable cost, would be graphic from the get-go, and would be far easier to use then existing computers because it would have a much better interface. So I drew up the specs for it and started hiring the people, designed a lot of the software, designed a lot of the user interface and managed to get a project started inside Apple called "the Macintosh Project." I called it "Macintosh" because the McIntosh is my favorite kind of apple to eat. And I figured that if I was going to have an apple I might as well have a tasty one.

UBIQUITY: Who did you report to then?

RASKIN: At different times I reported to different people. For a few years I reported to Steve Jobs. Later I reported to Tom Whitney.

UBIQUITY: Some innovations have many parents, but this really doesn't. "You're the man," as they say.

RASKIN: There were many threads leading to the Mac, but within Apple I was its first and, for a while, its only champion. I first wrote about the Macintosh in 1978, and the Mac come out in 1984, by which time Jobs had pushed me out.

UBIQUITY: Why did he do that?

RASKIN: That's a very long story, having to do with personality and other stuff.

UBIQUITY: His or yours, or both?

RASKIN: Probably both. I think that he likes people to jump when he says jump. I felt that he was untrustworthy, and that he does not take kindly to being found wanting. I was unimpressed by his thoughts on how computers should work (he said for a long time that the Mac was the "dumbest idea" at Apple and tried repeatedly to kill the project). He doesn't seem to like people who see him without a halo.

UBIQUITY: How many years did you work at Apple and work on the Mac project?

RASKIN: I first started working with Steve Jobs and Steve Wozniak in 1976, when they were still in the garage. I was hired to write the manual. I became an employee instead of a consultant in January of 1978 and left in 1982. I was the 31st employee at Apple Computer.

UBIQUITY: Do you have any thoughts that you would like to share on the Macintosh after you left the project?

RASKIN: A book full.

UBIQUITY: Have you written a book on your experiences there?

RASKIN: Well, it hasn't seen the light of day, and won't until I finish the more technical books I am currently writing. It's called "The Mac and Me."

UBIQUITY: Does it have a story? A moral?

RASKIN: It the story of how I came to have enough information in my head to design such a thing (and of the many people who guided me), and how I came to learn to put human needs ahead of technical concerns. It's about the things that gave a strong humanitarian and altruistic direction to my life and a lack of interest in making money for its own sake. My parents were a great inspiration; in the 1950s they risked their livelihoods to defend racial equality -- this was before Martin Luther King's prominence and the famous civil rights movement. They achieved some of their goals for our town, but we were boycotted by the white community, lost our family business, and had to move a step down from the middle class. But there was no iota of regret in our household, the moral victory was ours and I have the additional satisfaction that the equality we joined the fight to achieve, if not complete, is far greater today than it was then.

UBIQUITY: So, the Mac was, from the beginning, a truly idealistic project?

RASKIN: Absolutely. But I also knew that it had to be a commercial success in order to be influential. Amazingly, I can still buy a current Macintosh model that has the attributes I dreamed of -- and more -- a full quarter century ago. But it and its progeny (e.g. Windows) have gotten too difficult to use, so my work at making computer technology accessible is not done.

UBIQUITY: Did the people that you hired, and the people that you worked for, all share that vision?

RASKIN: Most of the people that I personally hired did. People I worked for certainly did not, except insofar as it was a profitable idea.

UBIQUITY: Where would the tensions have arisen between idealism and commercial viability?

RASKIN: There's not necessarily a tension there because I realized that if the machine were to sell by the millions it would be very influential. Some of the tension was that I wanted Apple to license the interface and Jobs didn't. In my opinion, after I left and before the product was released, the interface got somewhat more difficult to use because some of the people who were designing it afterwards had less of an understanding of human factors. They just liked to play, and didn't care about the user as much as they cared about the joy of showing off technology. We will never know if the changes made it more of a commercial success or less. I could have been right, or they could have been right, we'll never know.

UBIQUITY: Give us some examples.

RASKIN: There's the overuse of the mouse. At one point some of the people on the project became enamored of the quest to do everything with the mouse. This is an interesting exercise, but has led to systems that are slower than necessary and cause unnecessary repetitive stress injuries -- which (as recent research has shown) is caused far more by mouse usage rather than keyboarding.

Another example is the absurd application of icons. An icon is a symbol equally incomprehensible in all human languages. Whatever language you know, you have to learn the meaning of an icon anew. There's a reason why humans invented phonetic languages where just a few symbols can be combined to produce any word. I've met many interface designers who assume that you should use icons in an interface. I ask, "Is this the best interface that can be designed, or is it spawned out of habit and convention?" Most designers just stumble along in the present paradigm, ignoring the decay underfoot.

Just look at a typical directory menu; a flood of identical icons distinguished only by (gasp) words! Ever notice that those little tips come up to explain icons are words? Note that we don't have little icons pop up to explain words.

UBIQUITY: What surprises me is your current interests in theory, for example, your paper on Turing and G�del and so forth.

RASKIN: I don't see why that should be surprising. I have always liked mathematics and the theory of computation. It is widely believed that what computers can and cannot do is limited by what Turing machines can do. Turing machines are a mathematical invention of Alan Turing, and have proved invaluable in understanding the nature and limitations of algorithms. However, many philosophers have assumed that the limitations of Turing machines apply to computers. This is incorrect (I recently pointed out the error to the editors of the Stanford Encyclopedia of Philosophy, and they changed the entry on Turing machines accordingly. See http://plato.stanford.edu/entries/turing-machine/ ).

UBIQUITY: How do you know that computers can do more than can Turing machines?

RASKIN: Only one example is needed to make my point: the digits of some real numbers, such as pi, can be generated by a Turing machine. I call such real numbers "tame". There are at most a countable number of Turing machines, and thus at most a countable set of tame numbers. But the reals are uncountable, so there are uncountably many -- the "wild" real numbers -- whose digits cannot be generated by a Turing machine. However a real computer has access to randomness (e.g. the time between clicks registered by a Geiger counter) and can generate the digits of a wild number.

UBIQUITY: That may be interesting to computer scientists, but how about the rest of us?

RASKIN: The significance of this discussion is that if computers were limited by the results of Turing and G�del, then there are certain tasks that humans apparently can do that we could prove computers cannot do. This, in turn, would show that human minds are in essence different than any mechanical computer. Many prominent thinkers, such as Roger Penrose, John Searle and Martin Gardener have made this claim. By showing that real computers are not limited by the abilities of Turing machines, this particular bar to machine sentience is lifted.

I am not the only one to realize that physical computers are more powerful than Turing machines. Peter Wegner and a few others have come to a similar conclusion independently, but I have enjoyed investigating the philosophical implications of this insight, and correcting the record wherever I find it wanting.

UBIQUITY: OK, but what surprised me is that compared to the Macintosh project that you conceived and executed so brilliantly, it's on a different level. It seems strictly theoretical.

RASKIN: There is much beauty in mathematics, information theory, and the formal development of computers. Theory guides practice, and practice shows where more theory is needed.

Besides, math is fun and I enjoy sharing it. Two years ago my son and three other students at their high school in San Francisco had completed the upper advanced placement (AP) calculus exam, and that's as far as they could go at that school. So, I volunteered and taught a course on the foundations of abstract mathematics to them -- a sophomore or junior level (post-calculus) course for math majors. It was a joy to see them light up as they learned the material (especially when it comes to understanding infinite sets), and it gave them a leg up in college.

But what is most relevant here is that mathematics can be applied to some aspects of interface design, and if I had not found math and physics to be comfortable realms, I might not have gotten as far as I have in understanding and developing quantitative tools for interface design analysis.

UBIQUITY: Can you give us an example?

RASKIN: An eminently applicable and ground-breaking book, "The Psychology of Human-Computer Interaction" (Card, Moran, and Newell. Lawrence Erlbaum Associates, 1983) brought a measure of quantification to analyzing interfaces, in accord with experimental evidence and information theoretic concepts. I rely heavily on their work and on that of others that amplified it, and was able myself to introduce an objective measure of interface efficiency that is of great help in creating efficient (and thus time-saving) interfaces. Using their and my methods, you can predict a lot about the performance of an interface and its human users without having to build and test it.

The theoretical approach is useful in rejecting certain classes of poorly designed interface widgets and methods, but there are many other factors of importance, and testing is still required to make sure that an interface works well.

UBIQUITY: Do you have something in mind that would be as world changing as the Macintosh?

RASKIN: Yes, the big project I'm working on now is THE -- The Humane Environment.

UBIQUITY: Tell us about it.

RASKIN: We already discussed people's annoyance with icons and the dangers of consorting with mice. Well, there are many other things about computers that I'm absolutely sure annoy you. There's the cut-and-paste where you forget to paste and lose what you've cut. If you have to set up a network it's likely to drive you crazy. In your word processor there are commands you never use, but they clutter your menus anyway. You can accidentally hit keys that make the program do something unexpected and it takes you forever to figure out how to undo it, because it's hidden on the third level of some obscure menu. If there's a new and great application you hear of, your interest may be tempered by the fact that you are also reluctant to get it, saying, "Oh, no, I have to learn yet another computer package." Am I guessing correctly that you've personally had all these reactions?

UBIQUITY: Oh, yes, absolutely.

RASKIN: These feelings are universal. You know that your brand-new portable in three or four years will be useless even though it will still have the capability of doing everything you need. But the computer world will have changed enough that it won't be able to use it. Everybody hates a lot of things about computers, so I've been designing a system that is unlike anything currently available, which is far easier to learn and use, and which has the promise of presenting greater stability for the future so that a person does not have to relearn as much or as often. It gets rid of the whole idea of applications, so that you purchase the abilities you need and don't have to buy or learn any others.

UBIQUITY: Is this something you would download from the Net?

RASKIN: Of course. THE applies much of what has been learned about human cognition in the 20-plus years since I created the Macintosh project. Few people are applying this information because everybody's so caught up in the present GUI [graphic user interface] paradigm. Almost every designer and manager in the industry believes that it's impossible to change the way that we use computers, except for marginal improvements or prettier boxes. That's said, because users should not have to deal with an operating system, installing and learning applications, and so forth.

UBIQUITY: So you would take GUI and throw it in the ocean and replace it with THE software. What would be a simple way of conceptualizing what THE really means, on a less than perfectly abstract level?

RASKIN: It's based very strongly on our knowledge of those things that are inherent to the way our minds and bodies work.

UBIQUITY: Well, wouldn't some people who developed GUI say, "Well, GUI is exactly that. We have an icon and that's universal." How would you attack that?

RASKIN: I would ask if you recognize all the icons your computer presents, and when you said "no," that's your answer. Remember that the modern GUI work was started way back in the early PARC days, and the mouse was still earlier. The GUI was based on our understanding at the time. It was also partly based on the guesswork of a number of very bright, but still fallible, people. It is no condemnation of them that they (and I) didn't get it all right. I'm sure that my present work leaves room for improvement as well. But it sure is better than what we have.

To understand how humans and their artifacts interact, I dropped out of the computer world for about six years. You'll find if you look at my vitae that was a period when I had very few publications and did not make presentations at conferences. I was studying cognitive psychology. I was reading papers, understanding the experiments, separating out the miniscule grains of wheat from the incredible amount of chaff that the field produces. That's what led to my book, "The Humane Interface," and to the realization that a lot of interface approaches we took were simply wrong, including some things that I had designed.

UBIQUITY: How would you relate your ideas to those of Gelernter?

RASKIN: He has a strong idea about organizing files, but the basic interface style is the same old GUI. It's one useful element, but it is not the needed revolution. The Cat product that I designed for Canon Inc. in the 80s kept things in a single chronological file, though you could regroup by moving items about, so in some ways I anticipated Gelernter's main thrust.

UBIQUITY: So, when you throw GUI into the ocean, what will you replace it with?

RASKIN: It won't go into the ocean. It'll be with us indefinitely, even if I'm successful. With all the computer languages that have come along since, we still have Fortran. We have cars, but horses are not extinct.

The best I can expect is that it will gradually become the system of choice for a minority of computer users that will, with time, become the majority. I am confident that we can do better than GUIs because the basic problem with them (and with the Linux and Unix interfaces) is that they ask a human being to do things that we know experimentally humans cannot do well. The question I asked myself is, given everything we know about how the human mind works, could we design a computer and computer software so that we can work with the least confusion and greatest efficiency?

UBIQUITY: Can you explain to people what makes THE different from GUI? For example, you might say, the graphic user interface is graphical. Whereas, my THE interface is more . . . what? Fill in the blank.

RASKIN: Well, that's a bit of a problem, and why we have part of it running via a free Net download. You should have seen the problems I had trying to explain how the Macintosh would work to people who'd never seen anything like it.

UBIQUITY: That's an interesting point.

RASKIN: That's why I just call it "The Humane Environment" because it's meant to be humane. It treats human beings well. By contrast, the present GUI interface was meant to make conventional computer structures more accessible. I think that's the fundamental difference: the GUI starts from a model of how the machine works, THE starts from a model of how we work. For example, the reason why PARC invented the desktop was to make the idea of an operating system and other technical underpinnings of computers easily understood by representing them graphically. The better idea would have been to eliminate the irrelevant technical details. If you wish to write a memo, what portion do you create while on the desktop? Zilch, zero, nada, nothing. The operating system is the program you have to hassle with before you get to hassle with the application. So instead of trying to gussy up everything and put these various wolves in sheep's clothing, I got rid of the wolves.

UBIQUITY: You say there is no operating system?

RASKIN: Not as far as the user is concerned.

UBIQUITY: Are you happy with what you perceive to be the state of education more generally right now at different levels?

RASKIN: That's a big switch of topic. It depends on what schools you're talking about. In my opinion, the local schools in my town are pretty dreadful. In the paper I read that the valedictorian last year has aspirations to become a cosmetologist. But there are also good schools. Public education in California, with some exceptions here and there, is substandard, under funded, and does not enjoy real public support. The entire climate of our country seems anti-intellectual; education that is deeper than that needed merely to hold a job is not treasured except in some minority subcultures and in the rare and exceptional family.

UBIQUITY: Now, you are currently, speaking of education, a professor at UCSD, right?

RASKIN: No, I'm not. I left being a professor there many years ago. When I resigned I got into a hot air balloon in the middle of Revelle Plaza and flew over the Chancellor's residence playing my sopranino recorder so that he would hear the sound. He came out and I yelled down that I was resigning and floated off. I was an art professor at the time and it seemed arty to leave that way. I have been invited to teach at the University of Chicago in the fall, and I am looking forward to that.

UBIQUITY: You are, since 1989, described as an independent consultant and a writer? What do you enjoy about being a consultant?

RASKIN: The variety of projects and the opportunity to be with my family far more than if I had a regular day job.

UBIQUITY: Give me some sense of the variety of projects that you work on.

RASKIN: Most of the projects involve helping a company make its product work better for its users, so it's mostly interface stuff. But, usually, when I get to a company I discover that the real problems are structural with respect to the company, and I end up having to work on that a lot.

UBIQUITY: So they are essentially all human interface problems, right?

RASKIN: Of one sort or another. Now and then, to keep ennui at bay I do other kinds of work. I have helped Toyama Musical Instrument Company in Tokyo design its line of recorders (the musical instrument, not sound recorders). I think I mentioned designing and building a surveillance aircraft for the U.S. Department of Agriculture as another project.

UBIQUITY: But, I think you were suggesting that when you got into a particular project it occurred to you frequently that the real problem was not product interface, but organizational in nature.

RASKIN: There is a cultural problem in many organizations that makes it hard to achieve quality.

UBIQUITY: Do you find that fixable?

RASKIN: Sometimes. Usually not.

UBIQUITY: How do you fix it, when you can fix it, and why not when you can't?

RASKIN: The majority of the projects I work on -- in fact the majority of the projects that any consultant works on, as I've learned -- never come to fruition. There is a general lack of understanding of how a good interface is a powerful marketing tool. I remember at one company I worked on there was this huge multi-million dollar project to improve a product. I was a consultant to this project but also worked with a group of four programmers off in a side room. The big group had huge meetings and tremendous resources. But the small group completed the software for a few hundred thousand dollars while the other group was still floundering. We had something that the customers loved but it wasn't this juggernaut that the corporation had decided on. Because I was working on both projects I went to the main project's leader and said, "OK, here's the solution. You already have it in hand, all you have to do is distribute it." He agreed it was better, and it was done, but -- and he was very candid with me -- he said, "If I admit that this little group did with a few hundred thousand what I told the company was going to cost $30 million, I'll be out of a job and so will all these people. So we're going to forget about it."

UBIQUITY: Is that a typical scenario?

RASKIN: That's one of the more extreme examples, but that kind of thing has happened again and again. More people say, "Oh, we can't make that much of a change. That will scare away our present customers." And I say, "If you keep to your old ways to avoid scaring your present customers then some competitor is going to take them." People don't understand, for the most part, the idea of competing with yourself. If you can do something better, put it out alongside what you presently have and let natural selection take care of it. You win in either case. If your old product remains popular, you win. If your new product takes off, you win. This was the argument I used at Apple. I was told, "We can't possibly produce the Macintosh. It doesn't have Apple II compatibility." They said, "Do you realize we have sold 10,000 Apples IIs, and we're going to lose this wonderful customer base?" That's why I wrote the article "Computers by the Millions." that's now available on my Web site.

UBIQUITY: Is there anything else you would like to add to this interview?

RASKIN: If people with Macintoshes surf to www.jefraskin.com they can download some of THE and use it right now. It's free. We are currently porting it to other platforms. My book, "The Humane Interface", which we mentioned earlier, explains why THE works well in detail.

I must admit that I do enjoy being an iconoclast. My article debunking Rogerian nursing theory has helped eliminate that nonsense from some nursing schools (the article is on my Web site).

One last example. Almost everybody is taught that airplanes fly because the curve on the top of the wing makes the air go faster than the air going by the flat bottom and thus (by Bernoulli's principle) the wing has less pressure on top and the plane is sucked up as it moves forward. As I noticed when I was a kid, this explanation can't be right because planes can and do fly upside down, when the curve is on the bottom. My Web site has an essay (look for the Coanda effect article) which explains what is really going on.

COMMENTS

POST A COMMENT
Leave this field empty