Recent disputes over the authorship of Linux are missing an extremely obvious point. Has nobody noticed?
commentary Recent disputes over the authorship of Linux are missing an extremely obvious point. Has nobody noticed?
You sort of have to feel sorry for Richard Stallman. Poor old RMS, sitting in his dingy office in the comp sci building at MIT, issuing proclamations about the difference between free software and open source software, and insisting that everyone call Linux "GNU/Linux", worrying that the efforts of the GNU Project might be forgotten. But perhaps he has a point.
|It took Tanenbaum three years to write Minix, and he had had access to the Unix source code while he was doing it, the report explains. Torvalds wrote Linux in only six months, without any access to the original Unix. What is he, supergeek?|
You may have been following the story that broke in late May, about a report published by Ken Brown, president of the Alexis de Tocqueville Institution, claiming Linus Torvalds didn't write Linux. Instead, it suggests he rewrote the code from Minix, a Unix clone designed by Andrew Tanenbaum at Vrije University in Amsterdam. It took Tanenbaum three years to write Minix, and he had had access to the Unix source code while he was doing it, the report explains. Torvalds wrote Linux in only six months, without any access to the original Unix. What is he, supergeek?
Why are companies licensing Unix source code "if it is as simple as writing it from scratch with little help or experience?" the report postulates. "Is it possible that building a Unix operating system really only takes a few months -- and, oh by the way, you don't even need the source code to do it?"
Naturally, the response from the Linux community was hostile, and not surprisingly it was suggested the Alexis de Tocqueville Institution is a Microsoft front. This is the classic Linuxhead reasoning that goes "this Institution said something bad about Linux, Microsoft doesn't like Linux, therefore Microsoft must have paid them under the table to say it". But subsequent statements from Tanenbaum and Torvalds, among others, reveal the report to be sensationalist fact-free nonsense aimed at either raising the profile of the Institution, damaging Linux, or both.
However, the report is 100 percent correct on one point: Linus Torvalds did not write Linux. This is where Richard Stallman comes into the picture. For years Stallman has been campaigning for people to say "Linux" when referring to the kernel that was originally written by Torvalds, and to say "GNU/Linux" when referring to Linux the operating system. Unfortunately for Stallman, we suspect the horse has already bolted on this one; most people use "Linux" to refer to the entire operating system, no matter how historically inaccurate it is.
When you install Linux on your PC, you're not only installing the kernel -- the bit that tells the software how to talk to the processor, memory, and hardware -- you're also installing compilers, editors, Web servers, e-mail programs, and a lengthy list of other applications. Torvalds hardly wrote any of them. When Torvalds started distributing Linux in 1991, a large proportion of the software he bundled along with it was written by the GNU Project.
GNU was started by Stallman in 1984 with the vision "to develop a complete UNIX style operating system which is free software". Between 1984 and 1991, Stallman and his colleagues had developed a fairly complete library of Unix-style software. What it lacked was a workable kernel, and this is what Torvalds took six months to write in 1991.
So when the study suggests it was impossible for Torvalds to have written Linux in the time he did, it completely ignores that Torvalds didn't write Linux. Or at least he didn't write GNU/Linux. The GNU Project had seven years of work on its Unix style OS by 1991, which would seem to indicate a few more person-hours were involved in GNU/Linux than Tanenbaum's three-year effort on Minix.
It astounds me that in all the punditry following this report, no one has thought to point out this obvious mistake. But then, considering the number of Linux enthusiasts who weren't even born in 1984, it's no wonder this all seems like ancient history.
This article was first published in Technology & Business magazine.
Click here for subscription information.