I just came across an interesting article that appears to be from Francesco Nidito—although I’m not sure about the provenance—concerning the Plan 9 operating system. The gist of the post was that Plan 9 was a brilliant piece of software but that it didn’t solve any real problems.
The most salient example of this was Plan 9’s fulfillment of the Unix promise that everything is a file. Unix broke that promise as soon as networking was introduced and the socket interface was implemented. Plan 9 provided a much more Unix-like networking interface that made networking connections look like files. The problem, according to the post, was that no one considered the socket interface a problem. Most users didn’t consider Unix broken in any significant way.
I have a couple of problems with the post. The first is the assertion that Plan 9 failed because AT&T didn’t have any experience in selling software. That was true—perhaps even more true—when Unix was introduced, yet Unix survived and thrived. It had all the same problems concerning incompatibility and the need to reprogram existing applications that Plan 9 had yet was successful.
Far from never hearing about it, I actually had a running Plan 9 instance and contra the article it had at least two excellent code editors, Acme and Sam. What it did lack and what I consider its fatal failure—and my second problem with the post—was its lack of a Web browser. If it had had one, Plan 9, like Linux, could have gained traction in the nerd community and leveraged that to more a general acceptance. Being able to browse the web is the sine qua non for any workstation OS these days so Plan 9 was going nowhere without one. Oddly, the Plan 9 faithful never considered this a problem worth solving and as a result Plan 9 lives on mainly as Plan 9 From User Space, a port of many of Plan 9’s tools and libraries to other operating systems.
It’s too bad. Plan 9 was a better Unix and it’s program from anywhere paradigm was a powerful and compelling feature. It’s too bad it didn’t succeed.