Olsen bursts the NC hype balloon

Q. Okay. Ken, when was Modular founded?

A. I think it was about the very end of '92.

Q. Okay. Now that was - and when did Larry (Ellison) and Scott (McNealy) start pounding the drums about network computers (within) clients and all that sort of thing?

A. That was a year ago October.

Q. Okay. So it sounds like you were ahead of the curve. It's interesting that these guys -

A. That's not quite a comparison. You see, we are doing something quite different. Most of the - I'm not sure when these people got involved but most of the time the early talk was about a much less expensive home computer and we had a completely different approach, much less obvious approach which was to offer a quite traditional PC that runs all the operating, normal operating systems including OS2 and NT, not NT then but NT now, in addition to the Windows and DOS, and to run all the applications but just as they are without anything special, and the thing that we offered that was unique was (diskless) and -it was floppyless and it was CD-less so that it only ran company software and it was designed for company use.

Now, if price -if the cost is important, the price was not the main emphasis because the cost of operating is a big cost and, when it is used in business, I argue, you should buy the best (cathode ray tube normally is) 17H1 and the best keyboard and then the best desk because you are going to work there all the time but then, define what his problem is, what his job is, and offer that software and have company software stored in a server that everything he needs is the same software everybody else uses, so that it all adds up afterward and there is nobody that has the opportunity to introduce games or funny things or, above all, viruses, and the unique idea here is not so very clever, will never make me famous but it is, computers are a tool.

They are there to do a job. They are not there to show off how much technology we have and what wonderful things we can do and so our idea is - will never make the paper like these other things will do.

We don't have meetings for standards because all the standards are the same we have for PC except that the company picks the software, tests it thoroughly, and nobody can introduce funny things.

Nobody can do, like a few people did; a company was running DOS and they bring Windows in illegally from home, and then call service to make it work ?? anybody else.

Q. Right. Now, the way that the sort of NC vision has evolved though is going after the corporate marketplace, is storing the applications on the server. I think the main difference between what you are doing and what they are doing is that they want new stuff built from the ground up for Java whereas you are pushing existing, proven solutions.

A. Yes. We do Java by emulation and, as far as we are concerned, Java is not the problem right now. We - the problem is doing ordinary things and, in fact, limiting the job to what ordinary things had to get done.

Q. What's your take on the network computer, the Java based network computers? What kind of role do you think they are ultimately going to play in that corporate America?

A. There's two questions; what part should they play and what part will they play? Will they play is so much a matter of people's attitude. They don't want to be left behind in technology.

Hey, it's the latest thing. We have got to do it.

I would be embarrassed not to be using the latest thing and, you know, information technology is the only thing in business which is not decided on rationally or by plan, or by study. It's always done by emotion or I don't want to get caught behind, which is a strange thing, you know.

Your people would not tolerate having a more powerful computer at home than at work and so people just do things for no rational reasons, or buy -technology is going so fast, we better buy ahead.

They don't define what they want done with it and most of the things that (people) say they do don't happen and most of the software is written by people who never did the job they are supposed to be helping, and most people never stop to think what really goes on in business.

So, I think it will quite popular just becuase no one still - or very few, rather - very few people are analyzing why they want something. They just go from the headlines and just go from there.

Now, of course, there's another movement which I think is a movement. It's a little bit like the rich people now are showing off by how economically they can eat, and there is a, it seems to me, a number of companies who are so proud of the fact they use 286's. They are doing very well financially. They are doing a great job and they use all 286's, and that's a point of pride. So there's that growing movement.

We are not a part of that one either. The 286, we wouldn't recommend to anybody. The 386, hey, that's not bad but you see, this market is not a rational market and maybe it's becoming more rational, and part of it is.

Now, we have been selling our diskless machines for three years, four years - three - four -three or four years, and not in huge quantities.

We haven't made a big deal out of it but NASA uses them because, in their launch facilities because they don't want somebody introducing new software off the network the day before the launch, and we have a number of people who are using OS/2 becuase they use OS/2 because they want the ruggedness and the stability of OS/2 in the financial world, and they love this idea of nobody introducing new software when all their financial transacitons are dependent on it.

Q. Right.

A. So, the - oh, and the other thing, of course, we are interested in is people like this also on high quality servers that are built like computers used to be built.

They are put in a locked computer room. Nobody trips over the wires. The janitor doesn't move the machine when he sweeps at night and everything is on racks. It's open, accessible and we like the idea that it is expandable. They can - anytime they call up, we can add more servers or add more disks.

So, that's the other part (of the busines that) goes with it and actually what we did first, and these clients were just a component that came along with the whole idea of secure, safe, low maintenance computing in the office.

Q. Right. It seems like you answered what you think will happen with network computers but what do you think should happen?

A. What should happen?

Q. Yeah.

A. We believe that what we are doing is so obvious. You know, you don't have big committee meetings. You don't write standards. You just say - what we do, just build a simple computer. Don't let - and have discipline in the organization so that people don't use random software.

Don't introduce any software at all except the small (new) - (people that have to) test new software for the company or have to do special technical jobs, they need to introduce software but, in general, you use a desk (to) work and you want - and the computer is a tool and then too often it's a toy and most people's job is fairly well defined.

Even engineers' jobs are usually fairly well defined and so something like this, like what they do would work well but I think it also should tolerate a choice of operating systems, partly becuase people have, for example, OS/2 in their organization already, and some people want OS/2 simply because it is security and they should have an opportunity to use it. Other people wouldn't be caught dead with OS/2 because they read it was obsolete and they want applications which aren't available on OS/2.

Now, they may be wrong there but they should have that freedom. So, we think it should be the freedom and we don't see any need for standards. I mean - well, no. I'm not - there should be standards for the whole PC business more and more, and less wild things.

Now, one other interesting thing; the fellow who manages our - (internally) eighty people or so, eighty or a hundred, and the fellow who manages our network says he would need five more people to support our desktops if we didn't have diskless computing. So, that's the big reason for doing it.

Q. Right. Now, the Microsoft's Net PC specifications, even though nobody has built systems based on it now, seems a lot more in line with your model. Is that true? What's your opinion of the Net PC?

A. I glanced through this. I glanced through this so I would be ready to answer the question and now I - Robert is the expert on these things.

Now, they also make a big deal about management. The way we do it, it doesn't take much management. They are all the same and what you need, what you are doing is in your file, and the software is right - is also on file and so you can sit down anywhere and do it. You can take it home and do it.

So, ours is just much more obvious and simple, it seems. Now, they may get down to this as they get further along.

Q. Right. Right. Just kind of a general question about Java since you are supporting Java on these devices; how important do you think Java will be in the long run? What kind of impact do you think it's going to have on the marketplace, this new style of programming?

A. I thought it was all foolishness for a long time because Microsoft owned everything with the way it was doing it but, apparently, I'm wrong and that's the limit of my knowledge.

Q. Right. Right. Ken, I know you were still with Digital or I believe you were still with Digital when the Alpha processors was announced. I think you probably ran the press conference. Do you have any thoughts on the recent law suit against Intel for infringement?

A. I know nothing about it except what I read in the Boston Globe.

Q. Yeah. One of the things that you mentioned got me thinking, Ken. You mentioned how everybody seems to be going in Microsoft's direction. It seems like a lot of large companies have kind of fallen in lock step with Microsoft's direction and I am just wondering what you think the impact of that is on the computer market, whether that is actually stifling innovation. I think you could argue that Digital is an example of that.

Rather than creating bold, new architectures, it seems like Digital is right now trying to support whatever it is that Microsoft is doing on the NT side, on the messaging side, on the Windows '95 side. What kind of overall impact does this have on the industry?

A. It's hard to tell but I hear more and more people disillustioned with the instability of NT and either going or desiring to go to alternate systems even though they are not as powerful just because they love the old days when things always worked.

The idea of crashing regularly is just dillusioning people. So, there appears to be a market out there growing for other alternatives that do things maybe simpler and more disciplined, more controlled, and always work.

Q. But is there less innovation?

A. What's that?

Q. Is there less innovation now? I don't see IBM making as many bold moves. They seem to be selling whatever the market is.

A. It might be just he opposite, that people are tired of innovation and they don't need innovation. What they really want to do is get their job done and have the system work.

Q. Interesting.

A. Innovation, we have gotten to define as being wild, way out, wonderful things and it's not all clear what we need now. Oh, for the day of 300 bauds, mail and it worked. Today that sounds beautiful.

Q. Right.

A. Some of these young fellows never heard of 300 bauds.

Q. Yeah. I mean, you can characterized what you are doing at Modular as both innovative and completely non innovative. You are running existing operating systems. It's kind of an old style of computing but it's just - you could argue that it is just as innovative as these Java based in clients and whatnot and, given that that is kind of your mindset, what you are doing now, I'm wondering what - how Digital might be different today if you had stayed in charge there, how your vision might have been implemented within a corporation of that size?

A. I thought about that many times but I never say anything.

Q. Oh, really? That's the question I am the most interested, though, Ken.

A. Yeah. I know. My history has always been to solve problems. (We rarely) had the fastest computer. We did take the best care of customers and we - it was problem solving (we) solved (which) frustrates many people because we rarely had the fastest machines, but that's not what customers want and that's exactly what we are continuing here, to solve problems, make them reliable, often charge more for hardware because it's built like a computer is supposed to be built.

That means rugged, locked up in an air conditioned room, no access to anybody, readily serviceable and expandable and we think that computers - well, when servers became popular, they were just PC's turned up on end and put on the floor, and people tripped on the cables and the janitor swept dust into the fan, and then people said that PC's weren't doing anything useful anyway, so what difference does it make and, of course, that wasn't true.

They sometimes really do do useful things and they often are very critical and, when they are critical, the server is the key item and the server should be the best quality we can conceive of, and sometimes redundant, full tolerant quality, completely redundant some distance away because the information is that critical and, in that case, the client part of client server computing is just almost a small segment of the overall picture and it has to be one that does not introduce problems.

It has to be one that's reliable doing what it does do.

Q. Right. I'm wondering how much of the stuff that you hear today that hyped so much just sounds old to you. You know, the network is the computer. For example, I think every press conference I have ever seen you at, you had an Internet - - an Ethernet cable - excuse me - and you talked about hte importance of connecting things together. I think the VAX mate was kind of - kind of modern when you think about it now. It was built - it had the networking built right in. It was designed as kind of a quasi network computer. How much of this stuff that people are saying is new is actually old in your estimation?

A. (Old) ?? you say. (Whirlwind) Computer at MIT, started in 1945, was quite different than the other computers becuase it was 16-bit and very fast, very highly intensity designed circuits with a very small number of them, and very simple architecture as compared to the 64, 100 bit (word link) computers then being built, which there were only two or three of them and the (Whirlwind) was, by any definition except size, a PC.

It had (cathode ray) ouput. It had a keyboard. It had a printer.

It didn't have mouse. It had a joystick and a - not unlike the little red button in the middle of our IBM keyboard, and it had a light panel, and it had sound. It had music. It came out accidentally.

Someone - the ?? element interfered with the intercom and someone figured out how to hook up the intercom to the ?? element and played Christmas carols for a Christmas party but, after that, every computer had loud speaker.

Now, we took that personal computer - we called it interactive then - and made Digital to introduce that. Now, before that we - the air defense system was built. The Russians (exploded) atom bomb in '49. Instantly there was tremendous pressure to make an air defense.

So, built on (Whirlwind), the air defense system was built, which contained local area network in each center which had, oh, two or three dozen stations.

The (refresh rate on the tubes) was two seconds, which a psychologist said was the worst possible. It was the best we could do but it was a local air network and there were several wide area networks; two sweeps of radars across northern Canada, visible observers in John Hancock and Empire State building and towers on the ocean and airplanes with radar, all in a wide area network.

So, when we - after we formed Digital, this was our background and we gave a computer to MIT and we ?? one from part of Lincoln - ?? and the students there took these interactive computers, worked twenty-four hours a day, never - they stopped studying, stopped washing, stopped eating, stopped going out with girls, something then they didn't understand.

Now, we understand but out of that came games, and there came word processing. They call expensive typewriter and came timesharing.

They discovered that it took an operator a while to respond after the computer answered them and in that time somebody else could be working, so timesharing came out of it.

So, with the timesharing and that background we had and developed at MIT we, in effect, didn't make mini-computers. We made personal computers.

So Digital made personal computers and then the servers to serve them, serve the then (dumb) terminals. Now, if you - and then, of course, that was networking. We didn't sell anything to the governemnt.

We said we won't sell to the government. We weren't pacifists. We just thought it was too much trouble.

I was called to Washington once saying, hey, you have got to sell to us. All our labs and all the academic labs have DEC equipment and DEC networks and you have got to sell to us because we want to perform this network called (ARPAnet, which was later on called Internet and it was becuase we had everything networked, come from that original background we had at MIT.

Now quite early, twenty-five, thirty years ago, we had the whole world of Digital tie together with networking.

I could write notes, to the dismay of everybody, I could dictate for two hours at nine o'clock in the morning and it would be all over the world.

You would break a meeting anywhere in Europe, in Mexico or anywhere in the states, they would spread out and go to the secretary's desk, get their mail, answer their mail, go back in the meeting, get all the things Ethernet, including all kinds of (interest routes), some embarrassing, some very good, and I spent so much time tryng to sell networking.

The New York Times; they couldn't register at all why you would want to network.

Even my friends at Ford, who I knew very well; why would you want to network, and so I spent most of my time selling fast computers and networking, often without success.

So, yes, I guess in some ways it must have been successful because there's a lot of them out there but the - yeah, so it's all, in one sense, old stuff and another sense wonderfully new and exciting, still so much more to do, get more done today, more fun today than ever and then, on the other side, the same old stuff.

Q. How much time are you spending, let's say, per weekend at Modular and what other types of things are keeping you busy?

A. Oh, doing everything. At Digital I spent most of the time in finance and organization; here, still quite a bit, most of the time in technical things, often the physics and the packaging and electrical engineering part of it.

Q. What else are you doing with your time, Ken?

A. There's not much else left.