Wawancara ini ditemukan lagi gara2 lagi beberes jurnal2 dan majalah2 lama. Konon Oom Made Wiryana mau berbaik hati menyimpankan dokumen2 berharga itu di calon perpustakaan akademis non-kampus yang akan didirikannya. Para pendukung Open Source boleh keki baca pendapat Ken tentang Linux. Mudah2an pendapat beliau udah berubah dalam 5 tahun ini.
Computer:What about the development history of Unix?
Thompson: The early versions were essentially me experimenting with some Multics concepts on a PDP-7 after that project disbanded, which is about as small a team as you can imagine. I then picked up a couple of users, Doug McIlroy and Dennis Ritchie, who were interested in languages. Their criticism, which was very expert and very harsh, led to a couple of rewrites in PDP-7 assembly.
At one point, I took BCPL from Martin Richards at MIT and converted it into what I thought was a fairly straight translation, but it turned out to be a different language so I called it B, and then Dennis took it and added types and called it C.
We bought a PDP-11one of the very firstand I rewrote Unix in PDP-11 assembly and got it running. That was exported to several internal Bell telephone applications, to gather trouble reports and monitor various things like rerouted cables. Those applications, independent of what we were doing, started political pressure to get support for the operating system; they demanded service. So Bell Labs started the Unix Support Group, whose purpose was to serve as the interface to us, to take our modifications and interface them with the applications in the field, which demanded a more stable environment. They didn’t like surprises. This grew over time into the commercial version from AT&T and the more autonomous version from USL.
Independently, we went on and tried to rewrite Unix in this higher level language that was evolving simultaneously. It’s hard to say who was pushing whomwhether Unix was pushing C or C was pushing Unix. These rewrites failed twice in the space of six months, I believe, because of problems with the language. There would be a major change in the language and we’d rewrite Unix.
The third rewriteI took the OS proper, the kernel, and Dennis took the block I/O, the diskwas successful; it turned into version 5 in the labs and version 6 that got out to universities. Then there was a version 7 that was mostly a repartitioning of the system in preparation for Steve Johnson and Dennis Ritchie making the first port to an Interdata 832. Unknown to us, there was a similar port going on in Australia.
Around version 6, ARPA [Advanced Research Projects Agency] adopted it as the standard operating system for the Arpanet community. Berkeley was contracted to maintain and distribute the system. Their major contributions were to adapt the University of Illinois TCP/IP stack and to add virtual memory to Bell Lab’s port to the VAX.
There’s a nice history of Unix written by Dennis that’s available on his home page [ed.”The Evolution of the Unix Time-Sharing System,” http://cm.belllabs.com/cm/cs/who/dmr/hist.html].
Computer: What accounted for the success of Unix, ultimately?
Thompson: I mostly view it as serendipitous. It was a massive change in the way people used computers, from mainframes to minis; we crossed a monetary threshold where computers became cheaper. People used them in smaller groups, and it was the beginning of the demise of the monster comp center, where the bureaucracy hidden behind the guise of a multimillion-dollar machine would dictate the way computing ran. People rejected the idea of accepting the OS from the manufacturer and these machines would never talk to anything but the manufacturer’s machine.
I view the fact that we were caught up in thatwhere we were glommed onto as the only solution to maintaining open computingas the main driving force for the revolution in the way computers were used at the time.
There were other, smaller things. Unix was a very small, understandable OS, so people could change it at their will. It would run itself you could type “go” and in a few minutes it would recompile itself. You had total control over the whole system. So it was very beneficial to a lot of people, especially at universities, because it was very hard to teach computing from an IBM end-user point of view. Unix was small, and you could go through it line by line and understand exactly how it worked. That was the origin of the so-called Unix culture.
Computer: In a sense, Linux is following in this tradition. Any thoughts on this phenomenon?
Thompson: I view Linux as something that’s not Microsofta backlash against Microsoft, no more and no less. I don’t think it will be very successful in the long run. I’ve looked at the source, and there are pieces that are good and pieces that are not. A whole bunch of random people have contributed to this source, and the quality varies drastically.
My experience and some of my friends’ experience is that Linux is quite unreliable. Microsoft is really unreliable but Linux is worse. In a non-PC environment, it just won’t hold up. If you’re using it on a single box, that’s one thing. But if you want to use Linux in firewalls, gateways, embedded systems, and so on, it has a long way to go.