Topic: Network Computers
Content Types: Article, Interview
CLIENT NEEDS: Wired needed research and publication of an authoritative, interesting article on the history of the personal computer as it relates to the emergence of network computers.
DELIVERED: A wide-ranging article about the first personal computer–the Micral–and an interview with creator and network computing pioneer Andre Thi Truong–the text of which can be read below.
A Talk with the Father of Computing
Andre Thi Truong is not often confused for the father of the personal computer—at least in the US, heartland of digital tech. But, as it turns out, the 61-year-old French Vietnamese entrepreneur is exactly that.
In 1973, two years before the debut of the famed Altair, his two-year-old company, R2E, created the Micral microcomputer based on an Intel 8008 processor—the genetic ancestor from which all PC generations have followed.
And last year, Truong’s France-based Advanced PC Technologies spawned another paradigm-breaking variation by creating the first network PC—the NPC 97. Soon after, the “group of five”—Microsoft, Intel, Compaq, Dell, and Hewlett-Packard—conjured up specs for network PCs that they felt would drive an immediately emerging industry shift—one marked by more centralized administration and data locale.
Stepping far ahead of the crowd is not Truong’s only hallmark. In 1988, he was awarded the French government’s Technical Education Gold Medal for bringing PC technology into French schools. Wired spoke with Truong about the yesterdays and tomorrows of computing.
Roderick Simpson: Describe the emergence of the Micral.
Andre Thi Truong: The Micral in the beginning of 1973 was very rudimentary. It had only 2 Kbytes of RAM, could only be programmed in binary, and that year we sold 500 of them only in France. In 1974 we wrote the first assembler, so I could demonstrate the first complete microcomputer using an Intel 8080 at the National Computer Conference (since replaced by Comdex) in Chicago.
RS: Does anyone have a beef with you claiming to have created the first personal computer?
ATT: Nobody has really contested that the Micral was the first microcomputer. In 1986, The Boston Computer Museum held a contest to elucidate the pioneers of microcomputers. Its examiners, including Steve Wozniak, gave the title of the first microcomputer to the Kenbak, which was designed in 1971 with discrete logic and not with a microprocessor. The Micral was nevertheless recognized as the first microprocessor-based personal computer, and the first microcomputer largely and commercially distributed, and it’s now displayed at the museum.
RS: How did you begin thinking about the network computer and your NPC?
ATT: The idea of the NPC has several origins. As the architect of an information system in France in 1979, I completed the world’s first downsizing by replacing mainframes with a system of PC terminals in a client/server relationship. I think that was the first large realization of downsizing in a client/server environment based only on PCs.
This architecture has been adopted universally, but with the increased power of the PC, clients have become more and more fat and at the same time, PC servers have become more and more powerful. And the client/server architecture has concurrently become increasingly difficult to administrate.
What this has done is raise the total cost of ownership of networked desktop machines from several thousands of dollars to what is now some tens of thousands of dollars. Three years ago, the Gartner Group signaled in their surveys that the cost of administration had become the biggest cost in a client/server system. It made sense to me to go back to a more centralized client/server architecture. This was the first idea when we began to think about the NPC. The other idea was to replace our old Minitel [France’s proprietary network] that had become obsolete.
These are the reasons that led us to Windows NT as an operating system, and that led us to boot NT from CD-ROM on the NPC 97, which is destined first for the intranet market.
I have to admit that Larry Ellison’s idea was a very good one, yet the concept of the Oracle NC was nevertheless based on an obsolete microprocessor (ARM). And I have to say, it is unthinkable to even think, as did Oracle beginning in 1996, about creating a new operating system. The installed base of Microsoft’s operating systems is so huge that nobody can realistically think they’re going to write a new operating system themselves. We have based all of our development on the bet that Windows NT will be the only dominant operating system.
RS: So you’d say that Larry Ellison was a bit premature when he claimed a few years ago that the PC was dead?
ATT: I’m positive the concept of the general-purpose PC is not a dying one. But if you look at the PC industry right now – it is a mature industry. By this I mean an industry that has reached a plateau and is definitely ripe for change. We certainly are today at a turning point, and this industry needs to now find new niches in order to grow again.
The momentum of the PC, however, will only accelerate thanks to the NPC fulfilling the low-end promise of its market. Taking this further, one can imagine that Windows CE incorporated in mobile devices will wire the entire world thanks to Iridium or Teledesic. On the other end of the spectrum, the arrival of the Pentium II is going to make the PC very competitive against traditional workstations like those made by Sun or Silicon Graphics.
For me, all these machines have to cohabitate to solve different problems with the same architecture, like cars: You have small cars, big cars, trucks, but they all have the same engine architecture and they all use gasoline. Same as the PC with different motors like the Pentium and quite the same gasoline—Windows X, Y, etc.
RS: How do you see the prospects for Sun and Oracle’s NC models?
ATT: The Wintel model has already beaten the Oracle model, which is naively opposed to Microsoft. And the Sun model that has been created with native Java running on a Sun NC with a Sun chip, is obviously proprietary as well. Sun will be the only company to support this product and I don’t think that Sun, as the lone hardware company, can support the software. In the past we have seen the similar failures of IBM with OS/2 and Apple with the Mac OS.
Furthermore, I have never considered the concept of multi-platform as a huge advantage. It was the case with Unix, which was profitable for a short period of time for limited manufacturers—not for the end user. I really think the Wintel domination will be so thorough that being multi-platform won’t be necessary.
RS: If you only had one choice out of the following, who would you pick to forge the next millennium of computing and why? Larry Ellison, Bill Gates, Scott McNealy, Andy Grove, Steve Jobs, a monkey named Bobo.
ATT: I think I’ll choose your monkey named Bobo because I know what the five others have done and are going to each do in their respective areas. And, with all technologies that they have created, perhaps Bobo could imagine a utilization of these technologies which none of them have.
RS: And if Bobo had prior commitments?
ATT: In this case, I’d choose both Bill Gates and Andy Grove, simply because, tied together, they are indomitable.
RS: If the terminals we use become dumber and dumber—say, to where Bobo can use them—and the network gets smarter, what effect will this have on users of both machines?
ATT: I do not think that the NPC transforms human beings into extensions of a dumb terminal. But people have thought the same thing about every new invention. Take television. There are of course a lot of idiots watching televisions, but televisions have also transported knowledge to the most distant corners of the world.
The terminals that we use, instead of becoming increasingly stupid, are going to become increasingly intelligent, because these machines are not dumb terminals if we use them in a client/server architecture. Users are going to become more intelligent because they will access more rapidly and easily the bulk of multimedia data that is increasingly complex in the voluminous Internet. This move is profound, because in one generation we’re moving from the ASCII terminal to an interactive TV terminal.
Near term, PC technology will increasingly penetrate the TV industry, and we’re working on the next NPC with this in mind. As the first step, the NPC 98 will replace all the boxes that have approached the TV from a computing standpoint—the set-top box, WebTV, and so on. I’ll have a prototype at next years’ Comdex.