Brief History Of Linux (Part 4)
Feature written by on Saturday, May 27, 2000
We're back with the next "brief" installment in our series on the history of Linux. We apologize for the delay in publishing this fourth installment, but it takes time to make this stuff up. Sit back, grab a caffeine-laced beverage, and follow along as we trace the development of the Linux Cathedral.
The GNU Project
Meet Richard M. Stallman, an MIT hacker who would found the GNU Project and create Emacs, the operating-system-disguised-as -a-text-editor that everyone loves to hate. RMS, the first member of the Three Initials Club (later joined by ESR and JWZ), experienced such frustration with software wrapped in arcane, forbidding license agreements that he embarked on the GNU Project to produce and share free software.
His journey began when trying to get a printer to work with his system. The printer's proprietary drivers simply would not function properly and RMS was unable to fix the problem without the source code. RMS discovered the fine print for the printer's driver software, written in Flyspeck-3, that no other end-user had ever to read, or even noticed. This license stated:
You do not own this software. You own a license to use one copy of this software, a license that we can revoke at any time for any reason whatsoever without a refund. Since we did not sell you this software, we are not responsible in any way if it contains bugs or doesn't work, or if it causes millions of dollars of damage. You may not copy, distribute, alter, disassemble, or hack the software. The source code is locked away in a vault in Cleveland. If you say anything negative about this software you will be in violation of this license and required to forfeit your soul and/or first born child to us.
The harsh wording of this license shocked RMS. He would be unable to use the printer because the drivers were locked behind a reinforced wall of legalese. The computer industry, RMS reasoned, was in it's infancy, which could only mean one thing: it was going to get much, much worse.
He had a dream that night... a horrible, terrible nightmare set in 2020 in which all of society was held captive by copyright law. In particular, everyone's brain waves were monitored by the US Department of Copyrights. If your thoughts referenced a copyrighted idea in any fashion, you had to pay a royalty. To make it worse, a handful of corporations held fully 99.9% of all intellectual property rights.
Coincidentally, Bill Gates experienced a similar dream that same night. To him, however, it was not a horrible, terrible nightmare, but a wonderful utopian vision. The thought of lemmings... er, customers paying a royalty everytime they hummed a copyrighted song in their head or remembered a passage in a book was simply too marvelous for the budding monopolist.
RMS, waking up from his nightmare, vowed to fight the oncoming Copyright Nightmare. The GNU Project was born. The idea was simple: develop a system with built-from-scratch free software covered by a viral copyleft license. Such a system would slowly infiltrate the computer industry and put a stop to marketers and lawyers.
The plan called for a kernel, compiler, editor, and other tools. Unfortunately, RMS and his fledgling GNU Project became bogged down with Emacs, the editor, that the operating system kernel, HURD, was shoved on the back burner. Built with LISP (Lots of Incomprehensible Statements with Parentheses), Emacs became bloated in a way no non-Microsoft program ever has. Indeed, for a short while RMS pretended that Emacs really was the GNU operating system kernel.
Over the years RMS and his crew continued to hack on Emacs and waste time in doctor mode, but HURD remained in vaporous pre-alpha development for years.
Of course we all know what happened next. Linus Torvalds, thanks to alien intervention as documented in the previous installment, throws together a Unix clone from scratch and places it under the GNU viral license. The Revolution begins.
Right place, right time
Linus Torvalds certainly wasn't the only person to create their own operating system from scratch. No, we're not talking about Tim Peterson, the person behind MS-DOS, since DOS isn't an operating system. Other people working from their leaky basements did create their own operating systems, however, and now they are sick that they didn't become an Alpha Geek like Torvalds (or at least a Beta Geek like Alan Cox).
Unlike these other failed projects, Linus had one advantage not many else did: Internet access. The world was full of half-implemented-Unix-kernels at the time, but they were sitting isolated on some hacker's hard drive, destined to be destroyed by a hard drive crash or thrown out into the trash can. Thankfully that never happened to Linux, mostly because everyone with Net access could download a copy instead of paying the $50 shipping charge to receive the code on a three-foot stack of unreliable floppy disks.
Indeed, buried deep within a landfill in Lansing, Michigan sits a stack of still-readable 5-1/4 floppies containing the only known copy of "Windows Killer", a fully functional Unix kernel so elegant, so efficient, so easy-to-use that Ken Thompson himself would be jealous of its design. Unfortunately, before the system could be distributed, the author's mother threw out the stack of floppies (along with a Babe Ruth rookie card) in a bout of spring cleaning. The 14 year old author's talents were lost forever as his parents coerced him into attending Law School.
He should've patented the idea
While 1999 was the year of the Linux Portal Gold Rush, the first ever Linux portal was actually founded in 1992. A small newsletter published by Lars Wirzenius, titled "Linux News", was distributed via FTP, Usenet, and e-mail. With the exception of flame wars and gratuitious spelling errors, this ancient (pre-Web) newsletter featured the same type of content as today's Slashdot, Linux Today, and Freshmeat.
Issue 3 (October 1992) contains what may very well be the first published Linus interview. Some quotes from this issue:
"I doubt Linux will be here to stay, and maybe Hurd is the wave of the future (and maybe not)..."
"I'm most certainly going to continue to support it, until it either dies out or merges with something else. That doesn't necessarily mean I'll make weekly patches for the rest of my life, but hopefully they won't be needed as much when things stabilize." [If only he knew what he was getting into.]
"I've planned the 1.0 release for a long time, and I've always waited just a bit longer. Right now my final deadline is "before X-mas" [Apparently Linus was unintentionally using the Microsoft vaporware tactic of giving a release date without actually giving a release date. Is that Christmas '92 -- or '94, the year when 1.0 was actually released? What works for Microsoft sometimes works for Linux, too.]
"World domination? No, I'm not interested in that. Galactic domination, on the other hand..."
"Several people have already wondered if Linux should adopt a logo or mascot. Somebody even suggested a penguin for some strange reason, which I don't particularly like: how is a flightless bird supposed to represent an operating system? Well, it might work okay for a Microsoft product or even Minix..."
"I would give Andy Tanenbaum a big fat 'F'."
The Snowball Effect
Back in early 1991 Linux was just some magnetic fields sitting on some Finnish student's hard drive. By the mid-1990's the Linux community was burgeoning as countless geeks fled Redmond monopolistic oppression, Armonk cluelessness, and Cupertino click-and-drool reality distortion fields. Even as early as 1991 there was an informal Linux User Group in Finland, although its primary focus was Linux advocacy and help, not drinking beer and telling Microsoft jokes as most LUGs do today.
Kernel development continued at a steady clip, with more and more people joining in and hoping that their patches would be accepted by the Benevolent Dictator himself. To have a patch accepted by Linus was like winning the Nobel Prize (without the prize money, at least), but to face rejection was like being rejected from Clown College. The reputation game certainly sparked some arguments and good old-fashioned flame wars.
One of the most memorable crisis was over the behavior of the delete and backspace keys -- a legacy problem that persists to this day just like MS-DOS' 640KB nobody-will-ever-need-more limit. A certain faction of hackers wanted the Backspace key to actually backspace and the Delete key to actually delete. Linus wasn't too keen on the proposed changes; "It Works For Me(tm)" is all he said. Some observers now think Linus was pulling rank to get back at the unknown hacker who managed to slip a patch by him that replaced the "Kernel panic" error with "Kernel panic: Linus probably fscked it all up again".
That secretive Silicon Valley startup known as Transmeta is, according to our Vast Spy Network(tm), really a conspiracy within a conspiracy. "Trans" is a seldom used slang word in Finnish that means "cover-up", so the name Transmeta is literally "meta-cover-up". Or at least that's what our Chief Linguist tells us, although he's been wrong about such things before ("Windows" is not really an old Native American word for "gullible white man staring on to hourglass" as he previously claimed).
On the surface, Transmeta was a startup that hired Linus Torvalds in 1996 as their Alpha Geek to help develop some kind of microprocessor. Linus, everyone found out later, was actually hired as part of a low-budget yet high-yield publicity stunt. While other dotcoms were burning millions on glitzy marketing campaigns nobody remembers and Superbowl ads displayed while jocks went to the bathroom, Transmeta was spending only pocket change on marketing. Most of that pocket change went towards hosting the Transmeta website (the one that wasn't there yet) which, incidentally, contained more original content and received more visitors than the typical dotcom portal.
Microsoft relies on vaporware and certain ahem stipends given to journalists in order to generate buzz and hype for new products, but Transmeta only needed Non-Disclosure Agreements and the Personality Cult of Linus to build up its buzz. When the secret was finally unveiled, the Slashdot crowd was all excited about low-power mobile processors and code-morphing algorithms -- for a couple days. Then everyone yawned and went back to playing Quake. It's still not entirely clear when Transmeta is actually supposed to start selling something.
But does Transmeta intend to sell anything? Long-time Humorix readers know the answer: Transmeta is really a front for an illegal Finnback smuggling operation, while also acting as the US branch of the sinister Helsinkian Underground. It's all a meta-conspiracy. We here at Humorix have found that our readers are getting a little sick of far-fetched conspiracy theories, so we'll move on now.
Meanwhile, back in Redmond
Microsoft's position as the 5,000 pound gorilla of the computer industry didn't change during the 1990's. Indeed, this gorilla got even more bloated with every passing Windows release. Microsoft's continued success prompted countless MBAs and PHBs to shell out megabucks for content-free books hoping to learn Bill Gates' secrets. As the Alpha Marketer, Bill Gates could spin flaws into gold, but he really had no secret. His business strategy was simplicity itself:
- Pre-announce vaporous product.
- Hire monkeys (low-paid temps) to cruft something together in Visual Basic.
- It it compiles, ship it.
- Launch marketing campaign for new product and remind peons just how innovative Microsoft is.
- Repeat (GOTO 1).
With such a plan Microsoft couldn't fail. That is, unless some external force popped up and ruined everything. Such as Linux and the Internet perhaps. Both of these developments were well-known to Bill Gates in the early and mid 1990's (a company as large as Microsoft can afford a decent spy network, after all). He just considered both to be mere fads that would go away when Microsoft announced some new innovation, like PDAs -- Personal Desktop Agents (i.e. Bob and Clippit).
Chairman Bill explained in an internal memo (don't ask how we obtained this document), "Linux and the Internet are both non-profit anarchies dominated by kids. Real people -- in other words, people with money -- aren't going to mess with these things. Users don't want anarchies, they want pre-digested content and controlled environments. They want Windows and the Microsoft Network experience."
Free, Open, Libre, Whatever Software
Eric S. Raymond's now famous paper, "The Cathedral and the Bazaar", set the stage for the lucrative business of giving software away. In CatB, ESR likened the software industry to an anarchistic bazaar, with each vendor looking out for himself, trying to hoodwink customers and fellow vendors. The produce vendor (i.e. Apple), for instance, felt no need to cooperate with the crystal-ball seller (Oracle) or the con artist hocking miracle drugs (Microsoft). Each kept their property and trade secrets to themselves, hoping to gain an edge and make money fast. "With enough eyeballs, all bug-ridden software programs are marketable," ESR observed.
ESR contrasted the "caveat emptor" Bazaar to an idealistic Cathedral model used by free software developers. European cathedrals of medieval days were built block-by-block with extensive volunteer manpower from the surrounding community. Such projects were "open" in the sense that everybody could see their progress, and interested people could wander inside and offer comments or praise about construction methods. "Those medieval cathedrals are still standing," ESR mused. "But bazaars built in the 14th Century are long gone, a victim of their inferior nature. Of course, the same fate will hold true for proprietary software."
CatB is credited by many (especially ESR himself) as the primary reason Netscape announced January 22, 1998 the release of the Mozilla source code. In addition, Rob Malda of Slashdot has also received praise for the decision because he published an editorial ("Give us the damn source code so we can fix all of Netscape's annoying problems ourselves!") about the subject a few weeks earlier.
Of course, historians now know the true reason behind the landmark decision: Netscape engineers were scared to death that a large multi-national corporation would acquire them and crush Mozilla. Which indeed did happen much later, although everybody thought the conqueror would be Microsoft, not AOL (America's Online Lusers).
The Netscape announcement prompted a strategy session among Linux bigwigs on February 3rd to find ways to sell the concept of Cathedral-style development to businesses and venture capitalists. They decided a new term to replace the confusing 'free software' was needed; some rejected suggestions included "Free Source", "Ajar Source", "World Domination Source", "bong-ware" (Bong's Obviously Not GNU), and "Nude Source". We can thank Chris Peterson for coining "Open Source", which became the adopted term and later sparked the ugly "Free Software vs. Open Source", "Raymond vs. Stallman" flame-a-thons that persist to this day.
Not the end
We're not finished with this Brief History of Linux quite yet. Check back eventually for the next installment, in which we make up some more stuff about the founding of Slashdot, the Linux Gold Rush, and the continued journey towards that state-of-mind known as World Domination.
To be continued...