A Conversation With Steve Jobs

News

NeXT CEO Discusses Firm's Strategy, Exhorts Unix Community To Band Together

UNIX Today!

November 11, 1991

Steve Jobs, the president/CEO of NeXT Computer and co-founder of Apple Computer, was the keynote speaker at Unix Expo. A few weeks before his speech, Jobs met with UNIX Today! senior editor Paul Krill and UNIX Today! news editor Evan Schuman. Following are some highlights of their conversation:

UNIX Today!: NeXT is one of the few general-interest workstation players heading into 1991's third quarter without a RISC product.

Jobs: That's true.

t this point, most users would probably not consider a RISC-free strategy to be a risk-free strategy. (Jobs laughs) Given the way the market is shaping up today, how effectively can NeXT mount a successful commercial push without a RISC product?

Jobs: Good question. First of all, not to be defensive at all, a lot of this RISC-CISC stuff is marketing. The 68040 has a lot of instructions that execute in a single cycle. A lot of these RISC processors have a whole lot of instructions to be RISC. So I think you have to separate the engineering reality from the marketing hype. As an example, the 68040 runs faster than most of Sun's SPARCs, everything but the SPARCstation 2s, and yet, it's a single-chip solution that costs a lot less and lets us deliver computers for 20 percent less money than its equivalent Sun counterpart. And that's important, so we don't want to get totally wrapped up in this marketing stuff.

Now, having said that, we observed that we're sort of on the second generation of RISC processors right now, generally running anywhere from the 25-to 40-MIPS range, and you'll see the R4000 machines come out around the high end around 40 MIPS. We see Sun at 28 trying to get up to 40, a lot of work going on in the 30 and 40 range right now that'll be out in the next six to nine months.

Of course, HP is claiming numbers in the 50s and 60s.

Jobs: Our general belief is if HP was quoting engineering MIPS vs. marketing MIPS, those numbers would be in the 30 and 40 range. But we are also working on RISC very strongly, and you will see us participate in the third generation of RISC. We think that the second-generation processors will get eclipsed by the third-generation RISC processors in 1992.

UT: What kind of time frame?

Jobs: Early 1992. I don't want to comment any further than that. But we feel the best thing about what we're doing isn't just going to be how blazingly fast it is, but that Next Step is going to run the same on it as it does on the others-just a whole lot faster, and apps will look and feel and work the same. Our company jewel is Next Step, and everything we do will run Next Step in a totally compatible fashion.

UT: There had been a popular belief that you were leaning toward the 88K and an equally popular theory that the Apple/IBM talks might have put some doubt into your mind. Can you offer us any reality to counter those speculative thoughts?

Jobs: Well, we haven't really commented on what RISC processor we're using. One thing I will say, though, is that Apple requested Motorola to build a version of the RS/6000 which is bus-compatible with the upcoming 88110-that's a very interesting request.

UT: Why do you consider that so interesting?

Jobs: Just think about it. Why would one request such a thing? And it appears that Apple's use of the RS/6000 is at least 2 1/2 years away-that's a long time for Apple to go without a RISC processor. So who knows what's going to happen?

UT: So you don't see any of those talks changing your opinion either way on the 88K?

Jobs: Well, without going into what RISC processor we're using, if one changed their product strategy every time a new alliance was formed, one would be a very busy person and not get much done. So, our goal is to build great products, and I think the stuff we're working on in the lab is outstanding. Just outstanding. So, when it's ready, we'll introduce it, and I think it'll be really good.

UT: What is your assessment of some of the Mips offering? Mips computers and chips?

Jobs: I think they're good second-generation RISC processors. They don't have any multiprocessing capability, and they don't have any graphics instructions built in. We believe very strongly that graphics are very important, and we also believe that multiprocessing is pretty important, so I think that those deficiencies are certainly easily correctable given a few years of work. But they certainly exist in the R4000.

UT: And you think they would not, do you feel that they don't exist in the 88K?

Jobs: You should go talk to Motorola.

UT: You mentioned multiprocessing as being important. I don't believe the RS/6000 chip has that yet, where 88K does. Does that sort of imply that you're still leaning toward the 88K?

Jobs: Who knows?

UT: You do.

Jobs: Yeah.

UT: Has the decision been made on which processor you're using?

Jobs: We have some RISC products that will see the light of day in 1992, and as I said, I think they're going to be dynamite. That's all I can say.

UT: And those are currently working within your labs?

Jobs: I won't talk about it. Just won't talk about it. I'm sorry, I'd love to but I won't.

UT: I had a reseller that I was speaking with recently suggest that NeXT might go SPARC. Has that been under any kind of consideration?

Jobs: You know, what's interesting about this conversation is that you all seem to think it matters what processor we use. Why? Who cares? Who cares what processor we use, as long as it's decent? It's funny. In the Unix marketplace, people seem to care about two things: which processor you use and which version of Unix you use. Who cares? These aren't where the battle's being fought anymore. The battle's being fought in how long does it take to design and build my app and what sort of applications do you have if you're in the scientific and engineering market. There's certain apps that are critical to you if you're in the commercial market. It's more productivity apps than who's got what. People could not care less what processor's in the box. That's yesterday's battle, and people who are still fighting it haven't figured out that the battle has shifted. The battle shifted to what's your system software like and what momentum do you have with whatever application developers are important to your marketplace. That's where the battle is being fought. So whether we use a SPARC, or an 8810, or Mips chip, who cares? You know, we don't sit around here and discuss which disk controller chip we're using, who makes our keyboard. Who cares? And I think that some people are stuck in these older battlegrounds-and it doesn't matter anymore.

UT: Is NeXT profitable?

Jobs: We are waiting until the time when we're very profitable to make a big announcement about that, so give me another few quarters and we will maybe amaze you with how profitable we are.

UT: Are you marginally profitable?

Jobs: You could say that. But we're not saying anything until we're impressively profitable. So, give us another few quarters.

UT: You've been appointed to the President's Export Council recently. What kind of insight do you expect to offer the panel?

Jobs: Well, most importantly, at 36, I'm actually kind of an old man in this business, so having done it for about 16 or 17 years, maybe I can offer some insight into what it's like for a high-tech company.

UT: In the last five years, there have been a lot of industry consolidations, with HP buying Apollo, AT&T buying NCR, Sun and HP working together, IBM and Apple working together. Throughout all of this, NeXT has remained a solo player. Why have you remained without consortium?

Jobs: Well, good question. The answer to the question is that actually hasn't been the case. As you know, we licensed Mach from CMU (Carnegie-Mellon University), and one of the things we did was-in addition to licensing Mach-we hired Dr. Avie Tevanian, who, along with Rick Rachid, is really the father of Mach, and we commercialized it. Because Mach was at that point a research entity, and we put a lot of work into making it a real product. We then gave that back to CMU because we like to give things back to CMU-our contract calls for us giving some of them back-and we give things back even beyond that because we want there to be one Mach. Our work is exactly what OSF picked up, and we talked extensively with OSF and even helped them to some extent with that, so we feel we've made a significant contribution to OSF's use/adoption of Mach and the acceleration of their having a product. And, even though we haven't joined OSF formally, we're certainly spiritually connected with them and so that's one part of the answer. The second part of the answer, though, is our industry. It's easier to make alliances than it is to make products, and we're seeing a lot of alliances but we're not seeing so much innovation come from these alliances. Our goal at NeXT here is really simple: We want to make great products. So to the extent that joining an alliance will help us accomplish our goal, we're willing to join an alliance. But we haven't really seen that it will help us. We've seen that it's just a lot of meetings to sit in-and what's come from some of . . . I'm not saying that alliances are bad, but if you look and say, "What's come from some of these alliances, you know, what came from the Sun/AT&T alliance?"

Well, OSF, you know. Was Unix really furthered by that alliance or did it really help to fragment it? My guess is that it probably hurt more than it helped. We've seen HP buy Apollo, we've seen all sorts of things happen, but have these alliances really produced the products that were promised when they were announced or do most of them just fade into the distance with a neutral effect at best, and with a fragmentary effect at worst? So, it's not that we have anything against alliances.

UT: You said that you have an almost spiritual bond with OSF because of their Mach work. At one point, when they were moving over to Mach, some OSFers believed that you were going to join them. Why did you end up not actually going ahead and joining OSF?

Jobs: What we try to do, we believe very strongly that these Unix battles are foolish, that they're not serving our needs as the Unix community and, more importantly, the problems with the Unix marketplace are not that there are three versions of Unix, all 5 percent apart, that need to battle it to the death for one version. The problem is that Unix is only part of what is considered a modern operating system. You know there are very robust graphics necessary, and we believe now that there's a whole object-oriented framework necessary. So while enormous amounts of money and people's time and market positioning and capital has been wasted on these Unix wars, it hasn't helped Unix to win against its true competitors, which are Microsoft and others. So joining a camp formally we didn't think really served what we thought was the right goal of the Unix community, which was to just settle these differences once and for all, either by saying there will be two versions of Unix, each 5 percent different, or agreeing on something in an afternoon. But forming two camps and waging war I think is an obsolete exercise in futility. So that's why we didn't join one camp or the other.

UT: Do you think the Open Software Foundation adds anything to the industry with the RFT efforts?

Jobs: I don't know. It seems strange to me that we should have to have these extraordinarily formal and long-term, or time-consuming mechanisms to resolve things that shouldn't be that difficult.

UT: What is your retail selling strategy in the wake of the disappointing results of the Businessland venture?

Jobs: We obviously made a real mistake with Businessland, but what we've done is we've begun to sign up who we feel are some of the best computer dealers in the United States and we now have about, I think it's about 40-45 dealers that are reselling our products-they're all trained, they're all really good, and we're not picking them based on affiliation with a chain. We're picking them based on going to users' groups and other local customers and asking, `Who's the best dealer in this town?' and going and finding that dealer. And we also have about almost 50 VARs signed up now and we're seeing some real success there as well. Because VARs add value, most of them by writing custom software. So NeXT, all of the advantages of NeXTStep accrue to the VAR as well. And they can bring value-added to their market much faster than they can on traditional platforms.

UT: What was the problem with the Businessland deal?

Jobs: The problem that we're seeing now is that in corporate America, five or six years ago they needed help with PCs, as an example. And there were some dealers that were very good and all too eager to help them and some, just some, of the dealers built their whole business around serving large companies. The problem was that large companies started to hire really smart people, in some cases, right out of school into their IS departments. And their IS departments got dramatically better during the second half of the '80s, and these people knew just as much about the products. They knew more about the company being on the inside, they had much lower turnover, they were paid more and so it wasn't very long before they got better than the dealers doing the support. And someday, somebody inside one of these companies, a purchasing person or someone, asked, why am I paying for support twice? I'm paying my internal group, why am I paying this dealer an extra premium? And so the criteria for buying PC products became price, price and price. And the dealers that had made their strategy servicing large companies basically were turned into fulfillment organizations, not demand-creation organizations, but simply moving products based on the lowest possible price. And that squeezed them to the point where they didn't know how to create demand anymore. The people that knew how to create demand were long gone because they weren't necessary to fulfill demand at the lowest possible price. And so that's not at all what NeXT needs, and it was just happening as we were entered into our relationship with Businessland-we didn't understand it at that time. But it accelerated rapidly during the last few years, and I don't think that channel is really capable of creating demand for new products. I think they're just capable of delivering commodities.

UT: You're selling through the computer companies, like Computer Attic, they're also a PC . . .

Jobs: Yeah, but they're at the high end of the PC business, and their whole business is creating demand and supporting companies that don't have an internal IS department, like medium-sized companies. You can go to a medium-size company with 500 to 1,000 people in it and they don't have an IS department-so they need one, they need the equivalent, and Computer Attic and good dealers just like them are ready to provide that and have survived for the last year, five years, by focusing on companies that need what they have vs. companies that don't. And most of our distribution to large companies is now direct; we have a direct sales force.

UT: About two years ago, you spent a lot of time in an IBM headquarters building in New York City meeting with reporters to talk about Next Step on the RS/6000. Whatever happened with that?

Jobs: One of the things I've learned in dealing with IBM is that there's a lot of people there, and a lot of people have a lot of opinions, so they're not a monolithic entity. We hear from some people that Next Step is alive and well, and we hear from other people that Next Step is never going to see the light of day from IBM, and I honestly don't know. We would love for IBM to move forward with it. In my opinion, IBM has a really serious problem, which is that it can't continue to sell things when someone else down the block will sell you the same thing for 30 percent less money. So the way out of this problem is to differentiate oneself: to add value and, in our marketplace, the competitive advantage of hardware doesn't seem to last longer than about a year to 18 months. It seems like a lot of people know how to make hardware pretty fast, so the only way I understand, the only way I think there is to achieve sustained competitive advantage is through software differentiation, software value-added, and it seems to take people a lot longer to catch up with software. And as an example, look at Microsoft with the Macintosh. The Macintosh came out in 1984. Microsoft was working with us in 1982 on it, on some application software for it, so we disclosed it to them. And it took them till 1991 to come out with a competitive product, Windows. That's by a generous account, seven years since the introduction of the Mac and probably closer to a decade from whence they first saw the product under nondisclosure, but seven years from the date it was publicly available. And with all the resources Microsoft brought to bear, that's an awfully long time. So closing the gap in software competitive advantage can really take quite a while, and so we thought initially-and so did many, so did some senior people at IBM-that Next Step was their answer, that Next Step could give them a sustainable competitive advantage. And I believe that that's still true today.

In our view, we handed them a diamond that could get them out of their difficult situation, but, as the people change jobs at IBM and the responsibilities for Next Step were handed from one person to the next, somewhere along the way this diamond fell in the mud, and it's sitting on someone's desk today who thinks it's a dirt clod and doesn't realize there's a diamond inside. And will they? Will they wash off the dirt clod to find the diamond or not? I don't know.

UT: When you're saying people changing jobs, are you referring to (former Advanced Workstation Division President and now president of mainframe unit Nicholas) Donofrio?

Jobs: Well, Nick's a really good guy, but he was only one of four people who's had responsibility for Next Step over the years.

UT: And who else has shifted that would have made a difference?

Jobs: I'd rather you ask them. We still have a relationship with them, and I take the long-term point of view that, over the long term, NeXT and IBM will have a very fruitful relationship, even though today it may seem that the relationship is not producing any products.

UT: When I spoke with them, they said they had developed Next Step Version 1 for IBM and then you came out with Next Step Version 2, and they shelved Version 1 until they could finish their own implementation of Version 2, and that's where they stand at this point. Does that make sense to you?

Jobs: No.

UT: Why not?

Jobs: Well, you know, I hope that's right. I hope they finish Version 2 and come out with it soon because I think a lot of customers would like to see that.

UT: But, the reason that it doesn't make sense to you would be what?

Jobs: Well, it should be very rapid for them to come out with Version 2 if they want to.

UT: So it's taking them longer than you think it should be taking?

Jobs: Yes. Next Step is very processor-independent, and so it's a fairly routine and expeditious task to port it to various processors, the RS/6000.

UT: Next Step itself has pretty much been identified as being very easy to program to, and yet there have been criticisms that there are not many applications for the NeXT platform. Why is that?

Jobs: Well, I don't think that's true. Next Step is a complete object-oriented system. It's not just a thin object frosting on top of a traditional cake. It's a true object-oriented system, and the advantage is not just in the object-oriented framework. This object-oriented basket, if you will, is filled with five years' worth of objects we've developed. One of the greatest benefits of object-oriented programming is that you can take objects that other people have written and simply subclass them and change them a small bit and rapidly reuse that code. So you need a collection of objects that come with the system, otherwise the whole thing doesn't do you much good. So Next Step is both of these. It's both the object-oriented framework and five years' worth of objects that come with the system. And so what's happened is that people ranging from WordPerfect to internal development projects that many Fortune 500 companies have discovered that they can build their mission-critical custom apps three to 10 times faster using Next Step than anything else they've ever seen, especially Suns. And that is the largest reason that people are buying NeXT computers today. There are others, but the largest single reason that people are buying NeXT computers today, usually over Sun, is that they believe that they can develop their mission-critical custom apps three to 10 times faster and, indeed, there's tons of proof that that's true. So, that part we really believe we have a 3- to 5-year lead in, minimum, over any competitor. So we think that that's a very, very important and sustainable competitive advantage that we have, Next Step. Now, we do have a lot of application software. We have the best spreadsheet, in Improv, which is totally compatible with 1-2-3. We have the best version of WordPerfect, the only real WYSIWIG version of WordPerfect, and it's compatible with all the other versions. We have the best version of Adobe Illustrator that blows the Mac version away and on and on and on . . . We have about 200 Next Step apps shipping. This compares with about 50 Open Look apps from Sun. They'll tell you that they have 2,300 apps, but only 50 of them are Open Look. We run all Unix apps, you can take any Unix app and recompile it and run it on our system just like you can on a Sun or any other Unix box, but in terms of Next Step apps which are usable by mere mortals, and if you compare them to Open Look apps or OS/2 apps, we have more apps than Open Look by about a factor of four, we have more apps than OS/2 by about a factor of two.

UT: What about Motif?

Jobs: In the productivity area, we blow them away. In the scientific and engineering area, they will have more apps than us. Because our target market-we are not going after the traditional Unix scientific and engineering work-that's not what we're doing. What we're doing is taking Unix into the commercial market. We call it the professional workstation, for lack of a better term, as opposed to the scientific and engineering workstation. We're bringing Unix to the commercial market where what they care about is designing, creating and deploying mission-critical custom apps alongside those in a multitasking environment. They want to be able to use a suite of productivity apps that are compatible in the data formats with all the ones they use in the PCs. And that's where we have an offering that Motif has nothing to compete with and Sun has very little to compete with. But if you take us into the traditional scientific and engineering marketplace, we are a pretty powerful and very cost-effective Unix box. And to address X, we do run X on our product very well. There are two: one X product, soon to be two, that you can buy from third parties that run X-Window right alongside their Next Step windows, and they're quite good. So we don't have anything against X for what it was designed for, but X has not made the crossover into the commercial marketplace.

UT: So you're saying on productivity applications, you have more applications than Motif or Open Look or OS/2?

Jobs: Oh, absolutely.

UT: But as far as HP/UX or the RS/6000, or the DECstation, all of these have more . . .

Jobs: {NeXT has more} productivity {applications} by a factor of several times. And also please remember that NeXT is selling more desktop Unix workstations than DEC is today, in unit volume. And we are caught up with HP, we believe, or close to catching up with HP depending on what numbers you believe. {This is referring to} the number of desktop Unix workstations. How many desktop Unix workstations do you believe HP is selling per quarter? We get different numbers from different people.

UT: Apparently, they have a backlog of thousands of orders at this point.

Jobs: Well, we're shipping about 10,000 units a quarter, which the best market research numbers we've got-since HP doesn't release those numbers-is that they're shipping about the same and that DEC is shipping about 8,000 a quarter. So we think we've passed DEC and we've caught up with HP-but maybe our HP numbers are wrong-and that Sun, of course, is shipping in the 40,000s of units per quarter. I think they just released their numbers, what was it, 46,000 for last quarter, something like 46,000. So, and maybe more important for the productivity software vendors like the Lotuses and WordPerfects and others, we're selling all of our units into the target-rich commercial environment who want to run productivity apps, as where most of the Suns, and the HPs and the DECs are being sold into the traditional scientific and engineering market, which traditionally doesn't buy productivity apps.

Jobs: We think our customers are really liking our products. We've had some very significant orders recently: the L.A. County Sheriff's ordered several hundred units from us, and we just closed an order for 500 systems yesterday. A lot of orders, and what's really fun is that we're finding that Sun is our main competitor. That's because they're the only one of the workstation manufacturers that really has made it, sort of made somewhat of a crossover in the commercial market, although they're not that strong. But they have some presence there, and what we're finding is every time we get a chance to suit up against Sun, we're winning. Like 95-plus percent of the time. Our problem is that not everybody thinking about buying a Sun for commercial applications is calling us on the phone. We wish they would. If I could give you one message to put in bold headlines, I would request that your readers, all of your readers that are considering Suns for commercial apps, call us up because everybody who is, 95 percent of them plus, are deciding to go with NeXT. And they're deciding to go with NeXT because the reason they're buying Suns in the first place, if they're a commercial customer rather than just a PC or Mac user, is because they have a mission-critical custom app that just doesn't make sense to try to program on a Mac or Sun or a PC. And so they're stepping up to a workstation but in doing so, and the main reason is to develop their mission-critical app, and we can do that three to 10 times faster for the primary reason they're going to workstations in the first place-we're three to 10 times better. In addition to that, the minute they decide to go to a Sun, they give up all their productivity apps which they'd like to keep, being a commercial customer, and on a NeXT they don't have to give those up. As a matter of fact, in some cases, they can get even better versions that are available on a PC. So, our biggest challenge is how do we get a chance to be considered, how do we get a chance to suit up, because it seems like every chance we get that opportunity, we have something the customer wants.

UT: HP and Sun have a joined forces to build an object request broker for the Object Management Group. What is your assessment of those activities?

Jobs: Well, I think anything that furthers object-oriented programming is a good deal, so I don't want to knock anything. I think it's very important to get an object-oriented system on the market and to get customer feedback and developer feedback in particular rather than sort of academically designing one that may or may not be implementable, that may or may not have the performance that one would need in a real live system. So my advice to Sun and HP would be to get an object-oriented system on the market as fast as possible. The brokering is a very small part of that-very, very, small, 5 percent or less of it-so I would suggest that some effort be given to the other 95 percent of it and get a system on the market as fast as possible to start getting feedback, because it's an extremely sophisticated thing. A computer based on object-oriented system software is a very difficult and sophisticated thing to engineer and actually make workable, and I would advise them to get one out as quickly as possible, rather than just worrying about some of the smaller issues.

UT: Would you adopt the technology that the OMG endorses?

Jobs: We'll take a look at it, and we may even contribute some technology to that as well. We've been asked to help with some of that technology. We have some solutions for, we think, some of the problems that they're looking for. We're not averse to using any good ideas. Picasso had a great saying. He said: "Good artists copy, great artists steal." So without taking out the legal aspect of that, my belief is that we're all in this industry together, and we should learn to cooperate a little bit more.

UT: One of the strongest value-adds of NeXT, at least one of the strongest perceived advantages, has been multimedia. A lot of the advantages that you've demonstrated with E-mail, with video and sound are very impressive for demonstrations, but how practical do you think that is going to be as a selling tool for the average commercial site that you're targeting?

Jobs: Well, we're selling our computers now for two reasons, and I don't think I ever got to the second. One is, people buy them because they want to implement mission-critical custom apps three to 10 times faster than they can on Suns primarily. And the second reason is interpersonal computing. Now what is that? If you go back to the mission of personal computing, it was to improve individual productivity and at best, creativity, and most people seem to think that that's mission accomplished. There were 50 million shipped in the 1980s-it pretty much worked. Now, a lot of people feel, and we're some of them, that competitive advantage of the '90s is going to need more. We're going to need to improve the productivity of entire organizations. And so what we're seeing is personal computing growing into intrapersonal computing, whose mission is to improve group productivity and collaboration. Now, what's really important is that we've been hearing about the year of the network for years and it's never happened, and it never will happen because networking is not a market-it's a technology. It's not an end, it's a means to an end. Same thing with multimedia; there's no such thing as a multimedia market-it is a technology that's a means to an end, and where multimedia will be used the most is in interpersonal computing, improving the productivity of groups and their collaboration. So we're seeing a lot of that right now, we're seeing a lot of our apps incorporated, we've been seeing a lot of the custom apps incorporated and the state of multimedia on a lot of computers is pretty bad. On a lot of them, you can't even put up a good color photograph, or two. You know, take SGI's new machines-they have 8-bit color frame buffers in them, you can't even put up a bunch of beautiful color photographs without having one of them look good. The rest of them look blah because you can't do it in 8 bits. You know what I'm talking about. So most computers haven't even gotten past the stage where you could put up a bunch of color photographs, much less other types of media, or moving, dynamic media. And I think that that's an advantage that we'll continue to keep.

UT: But a lot of the RISC players, your point is very valid that a lot of their current multimedia offerings are extraordinarily weak.

Jobs: And that's probably because the scientific and engineering market doesn't demand it.

UT: But right now, there is a lot of multimedia development with the other players. HP has got a lot of those things in the works, as have Sun, IBM and DEC. What reason do you have to believe that in two years, when multimedia is a far stronger requirement, that they're not going to have revved up their products and seize control of the market that you've helped create?

Jobs: We think multimedia is absolutely moving from the domain of hardware to software. Doing stuff with all this is really becoming much more software-intensive, and it ties in beautifully with the object-oriented environment that we have. I think the future of multimedia is very software-intensive, and it also fits very, very, wonderfully into the object-oriented environment that we have. So the multimedia battle will rapidly turn into a software battle, and that's one where I think we can make some real contributions and stay ahead.

UT: When NeXT started out, you had an academic-only marketing strategy, and then you went commercial-what was the purpose of the original strategy, and was it successful?

Jobs: Well, about 30 percent of our sales are to higher education today, 30 percent of Q2 was to higher ed, and we have now emerged as the No. 2-selling computer on a bunch of campuses in this country, including Stanford and MIT. We outsell everything but Macintosh, and we're gaining on them. Matter of fact, on some of these campuses we are now selling as many NeXTs as Macintosh 2s.

UT: What about Berkeley?

Jobs: Ah, Berkeley. I don't know where we're at at Berkeley. But, we're being sold on about 300, we're being used in about 300 campuses in the country, and we're being resold on over 100. I think it's about 110 now, so we're really happy with how that's going, and we think we've got some dynamite products for these people. As an example, we bundled Mathematica, and Mathematica's a pretty hot thing in higher ed, both for labs and for students and faculty. And it turns out that Mathematica on NeXT is by far and away their best version. It runs about 10 times faster on a NeXT than a Mac 2 FX, if you can believe that, about 10 times. So we're really doing well because we have, a) the best version of Mathematica and b), we bundle it. And I think the version on Apple costs about $695/$795. The version on Sun's even more, so it's a tremendous . . .

UT: And the version on Next Step is what?

Jobs: It's free-it's bundled with every system we ship to higher education. Comes right on the Winchester, so, actually no, it comes on a floppy. Two floppies, that's right. So it's a great deal, and it's the hottest version of Mathematica, so that helps us a lot. And we are enjoying some real success there.

UT: The historic advantage on going academia is of course that three or four years down the road those graduates will be wanting what they worked with in school.

Jobs: Of course.

UT: Do you sound like, have you started to see that effect yet?

Jobs: Yes, yes. It's also a very good way to . . . it's also a very good thing in terms of recruiting people, both for ourselves or our customers.

UT: Do you see yourself as an outsider to the Unix community?

Jobs: Well, I'm not a programmer, I'm not technical so as a non-technical person . . . No, actually I think we're in the Unix community-hook, line and sinker.

UT: In what way would you not consider yourself technical?

Jobs: I'm not a computer scientist. I mean I am very good conceptually, but you would not be well-advised to hire me as a programmer to write your next important program.

UT: With some of your more recent revenue figures, about half of your revenue was from Europe and Asia. Has that been the case through much of NeXT's history?

Jobs: No. Well, as you know, we actually went to Asia before we went to Europe with our relationship with Canon, and that's working out very well. And the most fascinating thing is that we've seen some real success over there without shipping a Kanji product. We shipped our Kanji product for the first time in June, so we're expecting sales to increase even more as a percent of our business over in Asia, as well as absolutely.

UT: Well, why didn't you ship Kanji before June?

Jobs: Oh, we were getting it done, but we came out really sweet. We're just getting the reviews out now, and we're getting reviewed as the best Kanji system in the world. And there's two primary reasons for that. One is that when you have an alphabet with 7,000 letters, which is about the minimum number of Kanji characters you want to ship on your computer, a bit map font to put on the screen takes a lot more room than a romance language counterpart, and it takes so much room, as a matter of fact, that all of our competitors ship only one size screen font. We use display PostScript, and so we went and licensed the beautiful Kanji fonts that Adobe, the outline fonts that Adobe has spent millions of dollars creating for laser printers. We licensed them and we run them on the screen because we use display PostScript. And we have the most beautiful Kanji fonts you can imagine, and they're fully scalable to any size you want-the marketplace loves this. They've never had it before. The second thing is that, as you know, with 7,000 letters in your alphabet, you don't want to have a keyboard with 7,000 keys on it, so how do you specify which Kanji character you want? Well, you enter it phonetically and then convert to a Kanji character. You have a dictionary look-up. It works pretty well; the way everybody else implements it is they have a window on the bottom of the screen where you type in your phonetics and then convert to Kanji. When you get the character you want, you hit a button and it puts it in the cell of the spreadsheet or wherever you want it. That's not in-line entry. What everybody wants is in-line entry.

Well, it turns out that's really tough. Lotus took about two years to build in-line entry into 1-2-3. What we did was because we're object-oriented, we have a text object, and almost everything that uses text in our system uses that text object. So we worked with Canon, because they have probably the best conversion technology. We borrowed that technology from them and put it inside the text object, so everything that uses the text object gets Kanji conversion automatically in-line for free.

So in Interface Builder, a software developer can just drag over a text field and start typing in phonetics and convert to Kanji. We took Lotus Improv right out of the box, object code only, binary only, and ran it on our Kanji system. And because we have dynamic run-time binding, it just linked with the Kanji text object at run time, and you could go up to any cell in the spreadsheet and start typing phonetics and convert to Kanji-for free. We've shown this to software developers in Japan, and it blew their minds.