Dennis Ritchie's Interview for Samizdat

Tuesday, June 01 2004 @ 09:25 PM EDT

Contributed by: PJ

Dennis Ritchie was mentioned in the AdTI press release about Ken Brown's forthcoming book, "Samizdat", the book attacking Linus for not "inventing" Linux. As you will recall, the press release said the author's book was based on "extensive interviews" with Richard Stallman, Dennis Ritchie and Andrew Tanenbaum:

"In one of the few extensive studies on the source of open source code, Kenneth Brown, president of AdTI, traces the free software movement over three decades -- from its romantic but questionable beginnings, through its evolution to a commercial effort that draws on unpaid contributions from thousands of programmers. Brown's account is based on extensive interviews with more than two dozen leading technologists including Richard Stallman, Dennis Ritchie, and Andrew Tanenbaum."

Today, Dr. Ritchie told me that was overstating it, at least with respect to himself:

"I think that the teaser for it, mentioning extensive interviews with me among others, is overblown in my case. Brown sent an initial (email) probe asking for an interview, in response to which we invited AdTI to send some sample questions (which I answered). This happened just before Brown's visit to Tanenbaum. The only other interaction was a brief phone call from a staffer who asked only about a couple of fact things: how many lines of code in some early kernel, what date was it released."

So in his case, the "extensive interviews" consisted of one email. Here it is in its entirety, with his prologue to me, published with his permission. You will see Brown repeatedly trying to elicit negative responses from Ritchie, who replies at one point: "the specifications for Unix were always quite open". Mr. Brown, therefore, put out a press release saying something very different from what he was told by Dennis Ritchie, from my reading. Our thanks to Dr. Ritchie for putting this information on the public record.


Brown sent an initial (email) probe asking for an interview, in response to which we invited AdTI to send some sample questions (which I answered). This happened just before Brown's visit to Tanenbaum. The only other interaction was a brief phone call from a staffer who asked only about a couple of fact things: how many lines of code in some early kernel, what date was it released.

The main communication was this, from me to Brown:


1) Tell me what the environment was like after ATT/Western Electric decided that they didn't want the Lyon's book around. How did you react to the Lyon's Book and its subsequent recall?

We in the research group reacted with great pleasure to Lions's book; it was very well done. Indeed the early Unix Support Group (that became USL etc.) were pleased as well, and in fact invited Lions for a couple of stays with them to help annotate more documentation. Lions also visited us later in the research group and did some annotation on early Plan 9.

What was decided from the licensing point of view (after 6th edition) was that teaching whole classes from the source might be worrisome, and I think he was asked not to continue doing this. The original book was never officially published except within UNSW for his course; however it was reprinted by AT&T for internal purposes and it was also was made available to AT&T/WEco Unix licensees. Lions was disappointed that things changed so that he couldn't teach courses from a newer edition, but I don't recall any animus.

As you may know, the 6th edition source and Lions's commentary about it was formally reprinted and published by Peer-to-Peer, which seems to have morphed into Annabooks, in 1997. Of course this was after much water had passed under the bridge, in particular after the earlier SCO had bought the rights from Novell, so it was SCO who approved this publication.

2) People say that the earliest version of Linux 1.0 had a very similar resemblance, line for line to Unix? What did you recall?

I have no idea whether this is true, since I've never looked, but I doubt it.

3) Have you ever discussed the Minix/Linux migration or any other topic with Professor Tannenbaum in Finland? What are your thoughts about his decision to create Minix based on Unix, regardless of the efforts by ATT to restrict its use?

Since you've visited him, you know that Andrew Tanenbaum was and is at VU in Amsterdam. About the similarity and "restriction": the specifications for Unix were always quite open. Tanenbaum, in Minix, wanted to use the specification but with his own outlook for didactic, research and release purposes. I don't think AT&T ever bothered him. He (like Lions) spent time with us more than once during the writing of his later books.

4) There is something missing in the legal history of Unix to Minix to Linux to me. Help me understand a few things:

1) ATT was obviously angry that Unix code was going everywhere at once. Did they look at Linux? Did they look at Minix? Why didn't they feel either products were copyright violations?

I don't think that they were seriously looked at, and I don't think that AT&T was angry. Both Minix earlier and Linux a bit later took pains to avoid copyright issues and use documented and, by that time, probably internationally standardized interfaces. In the case of Minix especially, the accent was also clearly in the research/academic area as opposed to commercial.

By the time they came along there was plenty of material like the SVID, Sys V Interface Definition, Maurice Bach's book and so on. AT&T and USL generally encouraged such public efforts. I can't recall the dates, but, for example, the company contributed to the IEEE POSIX standards effort.

2) In your opinion, why did ATT lose its case against Berkeley?

This is the one case where USL did get angry enough to go to court. You will have to read the court's decision about it to see why USL lost (or at least didn't get what they wanted). In the event, the crucial decision by the court was to deny an immediate injunction against BSDi and UCB. Probably you've seen it, but the decision is at and it's probably best to let it speak for itself.

3) In my opinion, you wrote Unix (UNICS) from scratch. In my opinion, Linus Torvalds did NOT write Linux from scratch. What is you opinion? How much did he write? I talked to a Finnish programmer that insists that Linus had the Unix code (the Lyon's Book) and Minix code. Without those two, who could not have even come close to writing Linux. I hate to ask such a bare-knuckle question, but I really feel that this part of history is very gray.

Say what you feel is ok to say.

We did indeed write Unix from scratch (though with intellectual influence from aspects of Multics and other systems). I don't know what Torvalds started with or what he had read. It seems plausible from his writings that he was distancing himself more from Minix than Unix as such. See, for example,

4) Could I get a copy of the original version of Unix that was released? My team is comparing Linux 1.0, Minix 1.0 and your first versions. If you can help with this, let me know.

We have only parts of the earliest releases. has a good collection of what is available, including a version of the kernel from ca. 1973, and also partially complete renditions of 5th through 7th editions. The earliest available material (from 1973) is at