Omni Chat:
Eileen Gunn and
David Gelernter

Eileen Gunn:

Welcome to Omni's E-Media. Our guest tonight is Dr. David Gelernter, professor of computer science at Yale University. Dr. Gelernter's newest book, Machine Beauty, discusses the need for elegance in computer design. In it, he takes the debate beyond functionality, into the realm of aesthetics: some designs are simply more beautiful than others.

In his book, Dr. Gelernter talks about elegance as a combination of simplicity and power, and gives a number of examples of elegance, drawn from engineering and industrial design in the Hoover dam, for instance, and the classic 1937 telephone design of Henry Dreyfuss. Dr. Gelernter argues two points — one, that elegance, "machine beauty," is the driving force behind technology and science, and two, that our society fears or resents this machine beauty: it bothers us.

The epitome of elegant computer design, it's widely agreed, is the Apple Macintosh desktop, which widely agreed, is the Apple Macintosh desktop, which built on the work of Xerox PARC's interface design team to produce a powerful, fun, easy-to-use system.

David, am I reading you correctly here? Would you expand on this point or correct me if I'm misinterpreting you?

   
David Gelernter:

I think that's fair. Beauty is tremendously important in itself, for its own sake but it also plays a tremendously important pragmatic role in technology, and there's no technology where its role is more important than software and computing.

   
Eileen Gunn:

Do you think there's survival value to an appreciation of beauty?

   
David Gelernter:

Yes, I have to assume that any characteristic that is a big deal in a human being, that is important in understanding what a human being is, is probably there for a reason. My guess is that, as many people have said, that the beauty sense is helpful to us in a lot of ways, probably more than we realize, but it's also important that we not be trapped into a position of the pragmatic Soviet commissar approach, where we value beauty only where it buys us something practical. It does buy us practical things, but it's an emotional issue, more than a practical one.

   
Eileen Gunn:

What is the pragmatic role of beauty in technology?

   
David Gelernter:

The pragmatical role of beauty in technology — in general, there are a million possibilities in any technological decision and beauty is the best guide we have in picking or guessing the right direction when we're faced with a tremendously large number of choices. It's also the case in all of technology, but mostly in software, that the biggest problem that exists is managing complexity, not getting overwhelmed by it...Software lends itself to things that are more complex than anything else, the most complex machines on Earth are the software machines. We know all the problems with complexity, nothing doesn't have bugs, but beauty is the best value in software because it's the best tool in managing complexity. If I have to build a tremendously complex machine, if I make the machine elegant, the lines clean, simple & beautiful, the chance is greater that I will be able to understand it and master it than if I slop together some jalopy of a program.

   
Eileen Gunn:

Not unlike the function of beauty paradigms in choosing a mate?

   
David Gelernter:

Well, you know, 85% of the world's greatest literature exists to convince us that beauty is not a good heuristic in placing one human better than another. The face that launched a thousand ships, is, I guess in a sense, the most famous face in literature, the enthusiasm for the Trojans and Paris for Helen of Troy was a bad idea, or at least that's the verdict of literature. So in general, this is one of the great tragic themes in literature - the beauty that attracts us in human beings can be dangerous and misleading. But that doesn't mean it isn't tremendously important to us, because it is. It doesn't mean that there isn't a man who doesn't want to find the most beautiful woman at all times. Is that the way we want to operate? No, it's a stupid way to operate, but that's the way we are.

   
Eileen Gunn:

Well, I understand that certain recent research indicates that waist to hip ratios are good indicators of fertility.

   
David Gelernter:

Don't believe everything you read. ;)

   
Eileen Gunn:

LOL! ...But why would people distrust technological beauty?

   
David Gelernter:

It rubs them wrong — it seems out of place. It doesn't fit our preconceptions, doesn't fit our worldview. We tend to think, particularly Americans, Americans historically are incredibly brilliant technologists, but they particularly tend to think of beauty and aesthetics as frivolous topics, not really serious, and engineering and technology as serious, important, worth money, challenging, hard, difficult, so forth. So although the best technologists and scientists and the best mathematicians talk about beauty all the time, I think we tend to edit them out, we don't want to believe them because it seems too improbable and it takes us a long time to grasp that it's true, and some never grasp that it's true.

   
Eileen Gunn:

Do you think this is unique to American society?

   
David Gelernter:

Absolutely not. But, there certainly is more of a tradition of admiring the beauty of great engineering, great structural engineering and beautiful machines in general. That way to look at the world is particularly better established in Europe than it is in the United States. You still see it today in car design, for example, where the leading Italian design studios, other leading studios, continue to turn out the most beautiful cars in the world. And the best European cars are certainly more beautiful than American cars, even though ugh they are not necessarily better engineered. The difference between a Porsche and a Renault is part of a tradition of admiring the bridges and the other structures and machines of the twentieth century, and of a society that used to put art much closer to the heart of a society than Americans did. That's changed — art in Europe is moribund, and New York is thriving, but traditional beauty & engineering aesthetics have been of less interest to Americans than Europeans.

   
Eileen Gunn:

Do you think that attitude is directly responsible for the Macintosh's failure (that may be too strong a term) in the marketplace?

   
David Gelernter:

I'd be astonished if it weren't a factor. I can't believe it's the entire explanation. I'm positive it's not the entire explanation. In the 80's, when you had the Mac going against the PCs and clones, generally running DOS, it was no contest on beauty. Everybody could see, and nobody disputed that the Mac was a vastly more beautiful machine. There were solid business reasons that gave Microsoft an advantage in the 80's, particularly IBM's status as an important business and the way the applications developed, but when you add up all the mundane reasons why Microsoft should have beat the Mac in the 80s, there's still something surprising in the fact that Apple didn't do better going against as hideous a system as DOS. It's particularly surprising when you factor in the studies showing that the Mac was the most cost-effective machine in dollars and cents, even if you paid more for it than a PC. It was easy hardware & software, and those cost-deficiency arguments are the kind of arguments that are important to American business, that they take seriously for good reason. So why didn't Apple sell more Macs? I can't explain all of the difference unless I can believe that Apple knifed itself in the back by playing to American's uneasiness of elegance and beauty. Apple itself promoted the Mac as "cute," Apple called it "cute," critics called it "cute," people who liked it called it "cute," people who hated it called it "cute." If you're a businessman, you don't want to invest money in "cute." Cute is the opposite end of solid, profitable. Yes, I think it has to do with the consumers emotion unclarity with elegance. I think that's important, it's still important.

   
Eileen Gunn:

Personally, I think the axis on which Apple's fortune turned was the middle manager's need to have a safe choice, a choice that could be defended if anything went wrong. I don't think it was so much "cute" as "chancy."

   
David Gelernter:

That's exactly the point. If I'm going to lay out money for a bunch of computers, I want to be able to defend my decision in legitimate-sounding dollars & cents business terms. And I certainly don't want to be caught telling people I spent money for this machine because it was beautiful. I'd get fired. The word "cute" itself is ubiquitous. It's *strikingly* ubiquitous in the Mac discussion of the '80's. You run into that word so often that you're forced to come to grips with it. I'm speaking here not as a technologist, but more as a critic. This word plays so many roles that you must come to grips with it. At the same time, I must say that it's impossible to quantify this kind of argument. I don't claim that the intelligence of the Mac, or the cuteness of the Mac wasn't the only problem with it, but I can't believe it wasn't a factor.

   
Eileen Gunn:

I was on the front lines at the time, and certainly Apple's promotional strategies didn't help. But it was the same problem that DEC succumbed to in the 1970s. It wasn't the safe choice, and big business likes a safe choice.

   
David Gelernter:

Right. I agree. Exactly my point.

   
Eileen Gunn:

Do you think the Window's desktop design was successful =because= it is clunkier and less elegant than the Mac?

   
David Gelernter: No. I think it was successful because it came from Microsoft, and Microsoft had the kind of solid business credibility that Apple didn't. Microsoft was in a position to make elegance respectable and Apple wasn't. As I said in the book, it strikes me, broadly speaking, as a John Wayne endorsing pocketbooks for men kind of deal. Pocketbooks look like handy gadgets, but needless to say men aren't going to carry them. But on the other hand, some people would have more credibility than other people in making that case. If Andy Warhol had made that case, men would not have been impressed. John Wayne had more credibility. Microsoft had more credibility too. Still does.

   
Eileen Gunn:

Well, I was in charge of Microsoft's advertising at the time, and I worked hard to make them sound like IBM.

   
David Gelernter:

You succeeded!

   
Eileen Gunn:

Do you think if the Mac look and feel had been licensed to Microsoft, it would have been successful? (I realize Steve Jobs would rather have chewed his own foot off than done so, though.)

   
David Gelernter:

Well, Microsoft scored a tremendous hit with Mac's look & feel just by appropriating it with Win 3.0, and to be fair, Microsoft was really ahead with DOS before that. So, would Apple as a company done better, if it had made a deal with Microsoft? I don't know, I'm not enough of a businessman to say. It wouldn't surprise me, but hindsight is too easy. It's also clear that there's too much Microsoft bashing in the world today. It's an easy position because there's some basis, but it's clear that Microsoft is a better business than Apple. Technology is not the only issue, of course.

   
Eileen Gunn:

With respect to the question of elegance — do you think it is necessarily a drug on the market, though?

   
David Gelernter:

No. In the long run, it always wins, one way or the other. If consumers can justify elegance to themselves in pragmatic terms, they'd much rather have it than ugliness. After all, Windows 3.0 took off, it didn't have to, there were loads of DOS applications, there were DOS shells that made DOS less of a pain to deal with, but as soon as you could buy elegance legitimately, instead of on the black market for cuteness, as soon as it was a legitimate buy, people bought it. I'm convinced that beauty is not only the driving factor in the evolution of technology, but is in the end, the most important issue in the market for technology. It just takes us as consumers quite a while to work our way around to that issue, but it's where we want to be in the end.

   
Eileen Gunn:

Would you talk a bit about your development of Lifestreams, your candidate for a new information-control interface? How does elegance figure into it?

   
David Gelernter:

Originated in my unhappiness, bordering on disgust, with every operating system on the market, the Mac desktop was revolutionary in the 1970's, and was beautiful in the early 1980's, but in the late 1980's, it was getting old, and today it's pathetically obsolete, whether you buy it from Mac or in the form of Windows. After all, it comes out from an obsolete, long ago, technology era that doesn't match today's computing environment at all. Matches it so badly that it's an intolerable pain to deal with. So that for example, the system was designed when the Internet was not the internet, email was unimportant, very few people used it at all, computing cycles were scarce & expensive, memory was expensive, and just as important, or more important, all computer users were new users. So in the 1980's, people didn't have many files, many directories, because they hadn't been online for very long. But today, when compute cycles and memories are cheap, and the problem isn't how to conserve those resources, but how to squander them reasonably, and the internet is bigger than ever. So many people use their computers as text managers exclusively. The operating system designed long ago for radically different computers doesn't work anymore. For that matter, the whole underlying thesis of an operating system is obsolete. There is absolutely no reason that I should ever have to think about where I have a file, what machine I'm on, what my files are named, what directory I stuck something in. What I want is to be able to walk up to a computer anywhere, and tune in my electronic life. I don't care if it's a Mac or PC just as I don't care if, when I tune in CNN on TV, I don't care if it's a Toshiba TV or a Hitachi TV. In short, for all these reasons I've sort of hinted at, I found myself so disgusted with what was available, I figured there had to be something better. Although the research I had been doing on software in the 80's was fairly esoteric stuff having to do with programming & distributed systems and artificial intelligence, I had to turn my attention to everyday computing needs because the situation was, in software terms, so incredibly awful.

   
Eileen Gunn:

As a matter of fact, your description here reminds me very much of a book I wrote for DEC in 1977 on the future of distributed systems. It was science fiction in 1977 — how close is it now?

   
David Gelernter:

It's up and running. The technical underpinnings of what you might call delocalized data, data that just floats in the network, nowhere in particular — those underpinnings have existed since the mid-80's, and the Lifestream system itself, which of course, isn't the only conceivable system of this sort, there could be, and I'm sure, will be, loads of them, but the Lifestream system in particular was in prototype form as a doctoral thesis a few years ago and is now the basis of a vigorous little startup company.

   
Eileen Gunn:

Realistically, how are you delocalizing data now? The system runs now on UNIX, right?

   
David Gelernter: The system runs on UNIX and on various Microsoft platforms and on the Mac and on a lot of systems. I don't want to get into boring implementation details, The existing commercial software hasn't achieved as much delocalization as we want it to, and as the research prototype did, but the prototype was based on the shared virtual memory systems that had been developed on the distributed systems during the 1980's. The point being, designed for systems where every network node shared access to a common bunch of data objects, even though there was no physically-shared memory.

   
Eileen Gunn:

So the system keeps track of where all the data is, and the user doesn't have to?

   
David Gelernter:

That's what the research system does. The commercial system at the moment is more like a conventional client-server deal, but it's moving in the direction of exactly what you say, complete delocalization. This is the characterization of Linda systems today and that's the technology we're building on.

   
Eileen Gunn: Is there a demo on the Lifestreams website? (It was down when I tried to access it.)

   
David Gelernter: Yeah, I believe there is. The URL is http://www.mirr orworlds.com.

   
Eileen Gunn: I think it's time to open the floor to our audience. We already have a number of excellent questions.

   
Hipbones:

Hi David. Thanks for at least two stunning books, and a marvelous rabbinic interlude in one of them. How would AI go about recognizing beauty? My own hunch is that an aesthetic sense is *the great sorting principle*, that it has to do with pattern recognition, and specifically the recognition of isomorphisms parallelisms in deep structure. So an AI that recognized deep isomorphisms across wide topic distances would be the ideal web navigator as an AI that recognizes deep isomorphisms across wide topic distances is a creative mind. It would also be playing Hesse's Bead Game, no?

   
David Gelernter: Hipbones, I think basically, that's exactly right. I wrote a book about this issue of what you call recognizing isomorphisms in widely different domains, a tremendously important issue in how the human mind works. I think you're exactly right in that if we were able to get software to be able to fake the human beauty sense, the fakery would be based exactly on the ability to make links between seemingly very different items. I think that the key to doing that is to, if we want to do it, and there's no question but that it could be useful, but if we want to do it. I think the key lies in fake emotions in the computer, and that's a fascinating, difficult problem, but there has been some progress made on it. of course, a computer can simulate human emotions up to a point — it can never have them — it's all fakery, it's a trick, but it's potentially tremendously useful.    
Eileen Gunn:

I've known a few people who seem to be simulating human emotions, myself. Next question?

   
Ishtar:

As a consumer I often feel I have a choice of (a) functionality (b) beauty or esthetics. Do I have to settle for this rotten choice? Or is there a way a consumer can voice their opinion in marketing of computers and technology products?

   
David Gelernter:

Ishtar, I hope there's a way. The biggest problem of the commercial software business today is the complacency of the consumers. For some reason, people are willing to pay money for garbage software. I think their standards are far lower in software than they are almost anywhere else. Many consumers are intimidated by software or whatever, they're complacent. So the choices available in commercial software today are often bad. They aren't always, there are some good programs for sale, but too often, the choices are bad choices, reflecting the choices consumers have made to plunk down. dollars for garbage. When consumers get more demanding, the industry will be forced to respond. But so long as people are willing to shell out for junk, why bother to... if they'll drink Kool-Aid, why bother to serve them champagne?

   
Parkbench:

Do you believe we should program AI in an empathetic style for human interaction?

   
David Gelernter:

Closer to: being able to simulate human thought is too tempting not to try. We've been trying and we're going to continue to try because that's our nature. Building fake humans has always been a preoccupation of humans and human art. I do think it's potentially useful, it's not just a scientific exercise. There are also potential pitfalls. We already know that people fall in love with dogs and creatures of certain sorts. If we give them the possibility to fall in love. with machines, the lonelier ones will do that, and it will be a tragedy when that happens. It'll probably inevitable.

Question — balforce says: in your book Mirror Worlds you said you believe technology will always be benign. Do you still think that is so?

Balforce, I would never have said that in a million years! It goes without saying that any invention of any sort will be used by good men in a good way, and by evil men in an evil way. Technology by itself is neither good nor bad. The idea that technology is necessarily benign is certainly wrong. It certainly is the case in Mirror Worlds that although the book ends with ambivalence (sort of a theme of the book), on the whole I thought that the evolution of technology had done vastly more good than harm. And yes, I still think that.

   
Melody:

Do you think the Science Fiction ideal of AI technology before practical application has led to any preconceived notions of what the functionality or visual presentation should be?

   
David Gelernter:

Melody, I don't know. I do know that science fiction is a factor in technology research. Personally, I can't stand it, and I never read it, but I know a lot of smart people who read it. In general, SF has had an influence on technology evolution.

   
Demosthenes:

Intel is now talking about the new interface paradigm as "visual computing". Do you think Merced and similar powerful chips will finally atrophy the keyboard and allow us to communicate in voice, or other means, with our computer ? Or is that still further away ?

   
David Gelernter:

Demosthenes, there's been impressive, you might say spectacular, progress in recent years on voice recognition, but it will never replace the keyboard, not that I can see. It will be useful in a lot of circumstances, but speaking now as a writer, I want to write my books not dictate them. I know a lot of writers who feel the same. Writing in one form or another is here to stay and the keyboard, in practical terms, has proven itself. I think it will be hard to dislodge it. And I think writing will never be dislodged, augmented, but not dislodged.

   
Forest4trees:

What will our children use for software and how will it differ from what we use today?

   
David Gelernter:

Forest4trees, I hope it will be a lot better. Children's software today, there's some great titles, I have two young boys and the computer is a great toy, but there are a billion things I wish I could get children's software to do. It's stuck on the same topics and they emerge again and again. I think, basically, it's something like software in general, they're complacent on the whole customer base, too easily impressed, too undemanding. In general, I think children's software today that's strictly for fun does well — there's a lot of good games. Children's software that's supposed to be educational strikes me as a flop, and in general, it strikes me that the uncritical embrace of computers in schools is a disaster. I think if every computer in every elementary school in the country was unplugged and tossed in the trash tomorrow, we'd probably be better off. They're being used today as just another way to play games and waste time (not everywhere). I tend to think that will change. I think software in general, there's a lot of garbage, but it's in it's infancy. You have to give us some time, but we'll get it right. I hope. Or at least better than we have so far.

   
Tropicalcode

Question — tropicalcode says: Do you think people want more emotion from their technology? In a vocal or visual sense?

   
David Gelernter:

Tropicalcode, not unless they're awfully desperate. There are a lot of desperate people who will take emotion when and where they can get it. But by and large, emotional computers are not high on most people's lists of essential items.

   
Eileen Gunn:

I'm afraid we've run out of time. Thank you, Dr. Gelertner, for taking the time to be here with us tonight. And thank you, everyone, for coming and participating in the discussion.

   
David Gelernter:

Thank you very much for inviting me.

   
Eileen Gunn:

Good night, everyone.

   
More about David Gelernter:

How I Survived the Unabomber, by David Gelernter, on Time magazine's website.

   
Web sites mentioned in this interview:

Mirrorworlds Technologies

   
This interview was conducted on Omni Online, March 16, 1998
Copyright 1998, General Media Corporation and Eileen Gunn

Home.