Irreal

At the 2022 Linux Conference Australia, Brian Kernighan gave a delightful talk on the early days of Unix at Bell Labs. It’s not so much that Kernighan tells us stories that we haven’t heard before but that he puts everything in context and explains how the unlikely miracle of Unix came about.

At the end of the talk, Kernighan speculates on whether Unix or a similar groundbreaking project could be undertaken today. There are certainly smart guys with good taste like Ritchie and Thompson around now and the open source infrastructure, whose predecessor had a lot to do with Unix’s success, is already in place. What’s missing, he says, is something like Bell Labs.

As Kernighan says, the Labs were basically a tax on telephone service so they had stable funding. More importantly, management took the long view and encouraged the researchers to work on whatever interested them. Industry can’t do that because they are too focused on the short term. AT&T was a monopoly with essentially a guaranteed income so they were relieved from worrying about quarterly results. Academia can’t do it either because they don’t have the funding.

It’s a nice video and well worth watching. Kernighan is an excellent speaker and always enjoyable and informative. The video is 48 minutes, 46 seconds so plan accordingly.

Everyone with even a tenuous connection to our field knows who Ken Thompson and Dennis Ritchie are. They probably also know who Brian Kernighan is but unless they’re fully immersed in Bell Labs lore they probably don’t know about Doug McIlroy except maybe for hearing his name as “one of the gang”.

That’s too bad because as anyone who was there will tell you, McIlroy, in addition to be the manager of the Computing Techniques Research Department—the birthplace of Unix—was also one its leading lights. It was McIloy, for instance, who conceived of the idea of pipes and insisted on their implementation.

Kernighan and McIlroy have a delightful video in which they discuss some of the Unix lore including the invention of pipes. They discuss a lot of interesting technical details, including the history of many of the Unix utilities such as Diff.

To me, though, the most interesting aspect of the video was the discussion of how Bell Labs worked. Although on paper they were hierarchically organized, it didn’t actually run that way. There were no junior engineers, senior engineers, or distinguished engineers there were only members of the technical staff. As for compensation, that was entirely based on contributions not rank. Ken Thompson, for example, for paid more than McIlroy, his boss. That meant that those interested in remaining individual contributors could do so and still progress in their careers.

Because, unlike academia, there was no notion of tenure, the researchers felt free to suspend whatever they were doing to help a colleague solve some problem. This synergy was responsible for a lot of the Lab’s discoveries.

If you have any interest in Unix lore this video is definitely worth your time. It’s just over 52 and a half minutes so plan accordingly.