An open-source letter

The current flap between SCO Group and the Linux community brings back memories of summer 1995. It was an unforgettable time for me, in a bad way.

As vice president of strategy for Novell's Network Systems Group, I had been working for more than a year to help plan the integration of NetWare and Unix into a next-generation network operating system. Our engineering teams were a year into a three-year development plan, targeting 1997 or 1998 for the release of a breakthrough converged platform to be called NetWare 5.

That next-generation platform would have replaced NetWare 4's aging kernel, which was specialized for file and printer sharing, with an ultramodern, Intel-based, 64-bit-capable, TCP/IP (Transmission Control Protocol/Internet Protocol)-native Unix operating system equipped to support the industrial strength needs not only of network services but also of applications and application developers. NetWare's basic file and print services, along with its world-class enterprise directory, security and network management services, would be seamlessly integrated with a world-class, Internet-ready Unix application server.

At the time, Novell was the market leader in server operating systems by a wide margin through its NetWare and Unix product families. Novell pioneered the LAN (local area network) throughout the 1980s and early 1990s and then acquired Unix through the 1992 purchase of AT&T's Unix Systems Laboratories. Ray Noorda, Novell's co-founder, chairman and CEO, had moved to acquire Unix based on a growing understanding that NetWare's operating system was not designed to provide the kind of general-purpose platform necessary for the spectrum of fast-emerging enterprise application services.

In late 1993, Noorda decided to retire, and Novell's new CEO, Bob Frankenberg, carried on Noorda's vision. But through 1994 and 1995, Frankenberg and the rest of the executive team were also confronted with an increasingly aggressive marketing campaign out of Redmond, Wash., for Microsoft Windows NT. Although NT was, at that time, not capable of seriously fulfilling the roles required of an enterprise network and applications platform, the marketing force behind it was powerful.

Over a period of a few months, a combination of excessive fear, severe internal political battles and the flawed reasoning of a vocal group of Novell executives ultimately prevailed upon a frustrated Frankenberg to sell the licensing rights for Unix to SCO in 1995 for a paltry $100 million or so, just a couple months after then-Netscape Communications went public. Just as important, Novell decided at the same time to abandon any effort to compete with Microsoft in application server operating systems.

I was devastated. I fought that decision at every turn--even taking it to the board of directors--but ultimately lost the battle. Defeated, I decided to leave Novell later that year, convinced that the company had made a catastrophic error.

One of my peers in Novell was Darl McBride, who today leads SCO as its CEO.

Defeated, I decided to leave Novell later that year, convinced that the company had made a catastrophic error.

McBride was then leading Novell's Extended Networks Division, attempting to interconnect nontraditional computing devices through Novell's software. He was a good friend in those days. I've not spoken with him since.

Today, McBride is leading a charge against Linux based on three allegations:

• Certain portions of the code in several major versions of Linux incorporate intellectual property of Unix System V.

• The offending code in Linux cannot be identified publicly without effectively "open sourcing" the implicated intellectual property.

• The open-source movement is a communist affront to capitalism and should not be allowed to interfere in the profitable business of proprietary software. He thereby implies that it is un-American to support the open-source movement.

Given SCO's opaque tactics in this dispute, I cannot comment on the first allegation except to note that Novell asserts that its 1995 agreement with SCO did not transfer copyright and patent interests in Unix. So SCO's current legal position on the first allegation is ambiguous.

Setting that aside, I am in a position to take serious issue with SCO's second and third allegations.

SCO says that it won't identify which portions of Linux violate its rights because, in so doing, it would effectively release such code into open source. Yet at the same time, not only is McBride claiming that such code is already in open-source distribution but that SCO has distributed such code itself under open-source license in its own Linux platform. According to InternetWeek, SCO asserts that it never intended to release its proprietary code into open source. Its nontrivial labor and expense to release a full Linux operating system was apparently just a dumb mistake. At the same time, SCO Senior Vice President Chris Sontag says, "U.S. and international copyright law asserts you cannot inadvertently and accidentally assign your copyright to someone else."

OK, Sontag, fine. If you cannot inadvertently or accidentally assign your copyright, then there should be no problem in identifying exactly which portions of Linux allegedly violate SCO's rights. Simply issue a statement that identifies the offending code, stating clearly that the identification does not represent a release of rights into open source.

SCO's fallback response to this obvious truth appears to be that public identification of offending code would be akin to letting criminals wipe away fingerprints from the scene of their alleged crime. That is nonsense.

Nothing more than a visit to any one of several Web sites--or to a Fry?s outlet--is required to obtain original copies of the source code allegedly violating SCO's rights. SCO's real agenda in refusing to identify the offending code is quite clear: Prevent the open-source community from removing and rewriting implicated segments of source code. In effect, SCO is seeking to prevent the Linux community from correcting the alleged plagiarism so as to broaden and sustain its prospects for royalties. Thus, SCO is now an accomplice to the crime it alleges by refusing to allow the alleged perpetrators to clean up their act. Courts should succinctly reject such blatant and calculated extortion.

In effect, SCO is seeking to prevent the Linux community from correcting the alleged plagiarism so as to broaden and sustain its prospects for royalties.

To see the foolishness of SCO's posture, imagine that the creator of a copyrighted book alleged that another author plagiarized his work but refused to publicly identify the offending segments. Without public articulation, SCO's case is utterly ridiculous and should be summarily rejected on procedural, if not substantive, grounds.

SCO's third allegation is both more absurd and insidious. The Economist writes that McBride "found the entire free-software trend 'communistic'...'We don't get the whole free-lunch thing,'" he told the magazine.

Well McBride, the "whole free-lunch thing" is not exactly free because it involves the painstaking and passionate commitment of a lot of wise and ethical people who agree to share their work openly with others. And open source does have a particularly powerful precedent that a lot of people do "get."

The progress of science over half a millennium has been founded upon the notion of "open source"--a model in which the brightest minds contribute their work to the community for the betterment of humankind. Scientists across all major disciplines publish their insights in the public domain, climbing upon the shoulders of others and enabling others to climb upon theirs.

Well-heeled opponents of open-source software allege that it is incapable of yielding the quality, innovation, integration and completeness of proprietary software. To see through this utter nonsense, one needs only to imagine the world we would inherit today if the disciplines of physics, chemistry and biology would have been governed by proprietary interests. It is a gross understatement to conjecture that science today would be defective, primitive, chaotic and full of intellectual black holes if insights had been governed as private property for the last five centuries.

The model of open science is "communistic" in the sense of community ownership--or rather community stewardship. But innumerable highly successful organizations and institutions in America are founded upon the ideal of community stewardship--including our democracy itself.

The downfall of communism was due to state control by totalitarians--an attribute embodied by today?s commercial software industry far more than by the emergent open-source science of information technology.

Public and private organizations around the world are now choosing the fundamental character of information infrastructure likely to serve human civilization for decades to come, possibly for centuries to come. The stakes of this choice reach far beyond the trivialities of the case of SCO vs. Linux. They reach far beyond the market cap of Microsoft. Our choice will inevitably shape the very purpose, freedoms and ethics of the medium through which human beings will communicate, trade, govern, learn and play in the 21st century and beyond.

I hope we choose to build the world's information infrastructure upon the model of open science--a model that, along with democracy, has served civilization more effectively than perhaps any other social agreement in history.