Maybe its pervasiveness has long obscured its origins. But Unix, the operating system that in one derivative or another powers nearly all smartphones sold worldwide, was born 50 years ago from the failure of an ambitious project that involved titans like Bell Labs, GE, and MIT. Largely the brainchild of a few programmers at Bell Labs, the unlikely story of Unix begins with a meeting on the top floor of an otherwise unremarkable annex at the sprawling Bell Labs complex in Murray Hill, New Jersey. [...] It wasn’t until late 1971 that the computer science department got a truly modern computer. The Unix team had developed several tools designed to automatically format text files for printing over the past year or so. They had done so to simplify the production of documentation for their pet project, but their tools had escaped and were being used by several researchers elsewhere on the top floor. At the same time, the legal department was prepared to spend a fortune on a mainframe program called “AstroText.” Catching wind of this, the Unix crew realized that they could, with only a little effort, upgrade the tools they had written for their own use into something that the legal department could use to prepare patent applications.
The computer science department pitched lab management on the purchase of a DEC PDP-11 for document production purposes, and Max Mathews offered to pay for the machine out of the acoustics department budget. Finally, management gave in and purchased a computer for the Unix team to play with. Eventually, word leaked out about this operating system, and businesses and institutions with PDP-11s began contacting Bell Labs about their new operating system. The Labs made it available for free—requesting only the cost of postage and media from anyone who wanted a copy.
You could even say Multics was widely used into the very late 1990s, with some of the most important and largest sites not shutting down until 1998. I used it at university in the very early 1980s, and the same system was also marketed as a commercial timesharing machine by the university (and used for big data/compute jobs like seismic analysis by oil companies). That machine was first installed in 1978, and wasn't shut down until 1993 (but of course the hardware went through massive updates and expansion over that time, often with very little downtime).
Oral history from all the last sites at the time they shut down suggests that they only stopped using Multics because they couldn't get newer faster hardware and perhaps also because innovation in the OS releases had stopped.
Indeed U.Calgary had outgrown the hardware and was desperate for faster and bigger machines and switched over to IBM POWER machines running AIX.
Honeywell Bull just wasn't innovating in CPU design and keeping up with advances in super-microprocessors (though they did experiment briefly with trying to re-write Multics, first for a newer generation DPS6 mainframe, then for a massively parallel multiprocessor Intel x86 machine built by Sequent, and apparently there were other sales groups inside Bull that wanted to port to newer hardware too).
So, Multics wasn't a failure in any way shape or form, but Bell Labs certainly failed at following through on their initial commitment to help build it.
@Greg A. Woods - It wasn't a failure in terms of design. I remember it got a lot of admiration from my CS professors, saying it was the most secure OS ever built, even considering the current OSes we were using. It wasn't a market success, though. Honeywell was never that committed to selling it. ARPA had to twist its arm just to get them to take it on as a product to sell. I remember reading an article in the early to mid-90s, when Multics was cancelled, that said customers had to educate Honeywell's salespeople about Multics, in order to buy it, because they didn't even know they sold it! Doing some research on it, 77 units in total were sold in the approx. 20 years that Multics was on the market. Also, some of the people who worked on Multics went to work for companies that created variants of it under different product names.
Well, "market success" is a pretty meaningless comparison in this case because it's kind of on the scale of comparing a pickup truck to a bulk ore hauler.
It's quite true that Honeywell/Bull were not deeply committed to it, but that's also kinda like saying farmers in Iowa aren't really interested in growing avocados. Honeywell/Bull was quite a large company and they had competing internal organizations. In the end Honeywell/Bull more or less stagnated w.r.t. innovation in computers -- they completely gave up on designing fundamentally new systems at the time that IBM, DEC, Motorola, and Intel were forging ahead with remarkable new designs. From a third-party view the lack of enthusiasm for Honeywell modernizing GE designs after they bought them is basically the same way the DEC Alpha completely lost steam once Compaq bought them, and indeed the same as the Sun Sparc ended up dead once Oracle bought them.
The real story of the success of Multics is that it was never really expected to go anywhere as far as it did, and that extends out to the admiration I and many others have of it to this day.
As to "variants under different product names", well there was only about one (Apollo Domain-OS nee AEGIS), and it was far less ambitious and less capable (and arguably less successful too), though it did add a network filesystem as a central feature, and might have had a promising future had HP not bought Apollo.
There really hasn't been any other system since that's even come close to implementing a majority of the key innovations of Multics in anything close to a recognizable way.
@Greg A. Woods - I was fortunate to live during a time when companies were still experimenting with innovative computer designs, taking risks with what buyers would accept. I saw plenty of what I'd consider to be worthy knock-offs of original ideas tried and rejected by the market for products that were far less impressive. It was sad to see, but it gave me perspective that at least when it comes to computers, what succeeds in the market is not the best of ideas. Looking back at my own choices, I had rather mediocre taste at times, because I didn't know any better, and that's really the point. A subset of the population ventured out of what many took to be conventional thinking and habits, and thought they were really getting into something new by getting something that helped them optimize what they already did, at a price that wouldn't make waves, or they could afford.
As Lewis Mumford said, society doesn't advance through technology alone. It only advances when it develops and effectively propagates new symbolic languages that help us understand more powerful ideas.
Thanks for mentioning Mumford. I've read quotes from him in the past, but not really discovered his writing before. Having recently spent a bit of time listening to philosophers you've kindled an interest in reading some of his work. I wonder what he would think of the internet as it is evolving today. One of the topics I was recently exploring was the idea that maybe we haven't seen evidence of intelligence outside ourselves yet because all civilizations end up withdrawing into virtual realities.
Back in this world I keep dreaming of ideas of how I might try to revive some of the core principles of Multics and AEGIS in modern form, and how that might help change cloud computing.
@Greg A. Woods - An old popular theory was that maybe the reason we haven't seen intelligent life outside of Earth is that they all end up destroying themselves. This was during the Cold War...
The best theory I've heard so far about why no other intelligent life has been found is the "rare Earth" theory, that intelligence here is the result of some very unlikely circumstances: being the right distance from the Sun, the fact that we have a "double planet" system with the moon, which a consensus believes was created by a collision between Earth and another planet. The moon keeps the Earth's axis stable. What I understand we've found with other planets is that they stay on their axis for only so long before the action of their planetary cores changes the tilt such that what was at the poles is later at the equator, etc., making it very difficult, so the thinking goes, for intelligent life to emerge, since this would drastically change the climate that any life would experience. Secondly, the idea that our Sun is an "average-sized star" is actually a myth. It turns out it's a lot larger than average, and that even though it would be possible for Earth-sized planets to exist in an average-sized star system, the energy and atmospheric dynamics the planet would experience would be quite different. It would need to be closer to the star to receive the same level of radiation Earth does (makes sense), but as a consequence, it would not rotate on its axis at all. Instead, it would be like our moon is with Earth, where only one side of the planet would face the star. The other side would be in perpetual darkness.
This isn't to say other intelligent life is impossible, just very unlikely. The idea is it's a "goldilocks" phenomenon.
Mark you might be interested in a series of conversations by Robert Kuhn with a number of scientists about the subject of alien life and alien intelligence. The one with Jill Tarter was particularly interesting, as was of course Drake's conversation.