Peter Wayner

The news is a major watershed for the language because Metrowerks is the premier compiler manufacturer for the Macintosh, and the Macintosh was once a Pascal-only club. Many of the early versions of the MacOS were written in the ’80s in Pascal because it was the clean, crisp, entirely modern language of the day. The language was also wowing the PC world at the same time, in part because a man named Philippe Kahn hunkered down in a garage and wrote TurboPascal — a programming tool that was cheap, fast and, most important, stable.

Pascal took a very mathematical approach to creating software: It forced the programmer to specify the “type” of each piece of data. That is, the programmer had to spell out whether variable “x” held some text, an integer or a real number. This didn’t prevent the programmer from doing something stupid like dividing by zero, but it allowed the compiler to stop the computer from doing something really stupid — like trying to multiply the word “rabbits” by 23. After Pascal’s birth in the early ’70s, the mathematicians in university computer sciences grooved on the complicated hierarchies created by “type-checking” a program.

The language began to fade quickly by the early ’90s — in part because Microsoft embraced C and its descendant, C++. In the early days, C was considered a low programming language that was especially useful in dealing with the grungy parts of an operating system, like the printer drivers. Pascal’s mathematical pretensions gave it a gloss of class but also drove the bit-banging programmers nuts.

C began to evolve and grow out of its blue-collar roots by embracing many of the features that supposedly gave Pascal the edge. Later versions of C and C++ provided good type-checking to catch errors, as well as elaborate structures for object-oriented programming. Meanwhile, Pascal’s development splintered as people argued over the right way to do things. It’s almost as if the Pascal community acted like the French by defending the purity of their language — while the C++ world acted like the English, promiscuously absorbing words from other languages.

Now, Java and C++ are replacing Pascal in schools, where it gained an early following as a very easy to read and teach language. In 2000, the College Board’s Advanced Placement Computer Science exam will move from Pascal to C++. Many universities now have Java courses as part of their curriculum.

Naturally, there are still redoubts of Pascal lovers who use the language — just as there are still some fans of the Soviet Union holed up in Cambridge, Mass. Kahn’s TurboPascal has metamorphosed into Delphi, a product that still has a sizable following and good support. The Macintosh programming mailing list was buzzing this week with contributors’ plans to start a grass-roots effort to build their own compiler.

Old languages and old political movements never really die. In recent years, languages like COBOL experienced a strange sort of renaissance, as the Y2K problem created a new demand for COBOL hackers. If the war in Yugoslavia is bringing back once-extinct political alignments like isolationist Republicans and hawkish Democrats, who knows what the future will bring for Pascal?

The problem with the war against hackers is that most of what the hackers are supposedly doing would be trivial if it weren’t happening on the Internet. The typical hacker attack on a Web site isn’t much different from scrawling graffitti on the outside of a building. Many attackers are just poking around — like suburban teenagers who hop a fence to jump into a pool.

All of this would be great theater and a nice distraction from the war in Kosovo if it weren’t inspiring some serious reprisals in the courts — and some ominous inflation in sentencing that could wind up affecting everyone who uses computers in his or her daily life.

Wars on hackers are usually followed by calls for legislators to “do something!” and campaigns for new laws to crack down on the bad guys. The problem is that “doing something” often produces laws that treat the same action much more harshly in cyberspace than in “meatspace.”

The archetype of the demon hacker is Kevin Mitnick, a young man who has spent more than four years in jail waiting for his trial. When he was arrested, Monica Lewinsky was in her last year of college. During this time, Mitnick and his attorneys have jousted with government lawyers in endless pre-trial maneuvers that seem to have ended recently when Mitnick decided to plead guilty, probably hoping to receive a sentence that would be limited to time served. But even that deal is uncertain and taking forever to evolve; meanwhile, for Mitnick it’s just prison without a trial and with no bail.

Many, no doubt, see the crackdown on folks like Kevin Mitnick as a great deal for society: Information can be stolen just like anything else; surely the thieves who traffic in such goods should be locked up, just like car-jackers and muggers.

But there’s also a hidden danger. The precedents that the courts set now for dealing with demons like Mitnick will also apply equally to everyone who follows. And it’s not clear that the world is ready for Mitnick-like sentences for the crimes he might have committed, which remain murkily defined.

Think about it: Someone who reads another person’s Rolodex is just a snoop, but someone who clicks through somebody else’s Palm Pilot is hacking a computer database.

It’s easy to see just how slippery the calculus of evil gets on the cutting edge of technology. 2600 Magazine, The Hacker Quarterly, recently posted letters from computer manufacturers like Sun and Motorola estimating their losses to Mitnick’s alleged theft of computer source code. After Mitnick’s arrest, he was said to have stolen billions of dollars of information. Some companies calculated their loss by simply listing the hundreds of millions of dollars in development cost of the software affected — that is, the cost of all the programmers, their computers and other overhead. Other companies were a bit more careful and noted that the value was difficult to judge, but that recalls of products like cell phones could be costly.

The problem is, the price tag of information is almost impossible to determine. If Mitnick did take a copy of these companies’ source code, the companies weren’t denied the use of it, as when a mugger steals cash. Mitnick’s lawyers seem ready to point out that the companies involved didn’t bother to announce an official price on what they lost to Mitnick — something that the Securities and Exchange Commission requires public companies to do if the losses are significant enough. That would have required strict accounting measures.

To make matters even cloudier, in the meantime, Sun Microsystems began giving away the source code to its operating system to students around the world. In other words, if Mitnick had only waited a few years, enrolled in a university and asked nicely, he might have been a poster boy for Sun’s charity instead of a prisoner. Today, Sun is even circulating the source code to products like Java in hope of recruiting customers and snagging bug fixes. The company is practically begging people around the world to come take a look at its code.

This big change in the customs and attitudes of the software industry
strains the arguments against hackers. If giving away the source code is now a “good thing” for corporations, did Mitnick and the other hackers do a smaller good thing by grabbing it ahead of time? Is Mitnick now a bit closer to being a Robin Hood instead of a demon? If Linux triumphs, will children be told tales of the dark days when the Sheriff of Notingham sat on the boards of all of the corporations and forced them to keep their source code proprietary so
only the nobles could enjoy its bounty? Is it true that begging forgiveness is always easier than asking permission?

Such questions may be impossible to answer, but they illustrate just how confusing it can be in the nether-netherworld of information’s hall of mirrors.
As a commodity, information is fundamentally different from objects, and society has always graced it with special respect. The journalists who printed the stories about the allegedly racist words that appeared on a secret audio tape of Texaco employees looked like crusaders. But if it had been a digital tape, the reporters could be painted as hacking data compiled by a Texaco employee on Texaco time.

In the long run, society is going to have to think differently about hackers and the crimes with which they are charged. Taking information when it’s printed on paper is not always bad, and there’s no reason we should change this rule just because the information is stored on a computer disk. The intent of the criminal and the extent of the malice has always played a crucial role in our system of criminal justice. Many owners of things will forgive a theft if the “borrower” merely returns it unharmed. Crimes like trespassing are rarely prosecuted if someone just hops a fence and does no damage.

Computers and the Internet continue to frighten people, but prosecuting hackers runs the danger of setting nasty precedents that will begin to snare regular people, not programmers. Many convicted hackers are released from prison only to be denied the ability to use a computer or the Internet. In the past, this made it impossible for a person to get work as a programmer; today, they can’t even push the order screen at McDonald’s. After all, it’s hooked up to a central database — who knows what havoc a hacker could wreak while punching up an order of fries?

One of the best ways to put this all in context is to take yourself back in time 100 years to the turn of the last century, when auto racing was just beginning to roar across the scene. The machines were grand in size and sound if not in speed — Emile Levassor won the 1895 Paris-Bordeaux race with his four-horsepower jack rabbit that covered the distance at an average speed of 14.9 mph. Feats of technical prowess like that frightened the world, and by 1903 the French government was shutting down auto races — or restricting the death-defying machines to a bearable 20 mph.

A few decades later, James Dean became a rebel automobile hacker who scared parents around the globe. Today, he’s just another cutie pie
competing with Hanson for poster space on dorm room walls. One era’s demon is another’s icon. Is teen idol the next stop for Kevin Mitnick?

But the latest batch of patents that focus on the Internet aren’t anywhere near as impressive. In fact, they look as if any schmoe did “invent” them — by taking some everyday occurrence and adding to it the phrase, “with a computer network.”

Consider U.S. Patent 5848161, which describes the flash of genius that hit two Canadians and an American: They “invented” the practice of locking up the data traveling over the Internet between the customer and the store — that is, they use encryption functions to hide credit card account numbers from prying eyes.

Or consider patents 5191573 and 5675734, created by Arthur Hair when he lived in Pittsburgh. He claims to have invented the concept of “selling electronically … through telecommunications lines, the desired digital video or digital audio signals” — in short, pay-per-view over the Internet.

It’s not really fair for me to single out these three patents, because there are many more like them. Plus, it’s hard to summarize the scope of a complex legal document in a short paragraph. (And no doubt some readers will want to point to similar patents I’ve been granted over the years: They’re not exactly a cure for polio, either — they don’t even make the short cut of my risumi.)

The problem is that the patent system wasn’t meant to be just an opportunity to pad the resume of a person or corporation. Patents are meant to reward inventors of products by giving them the sole right to control who uses their invention. This is usually enforced by another invention popular in America, the lawsuit.

Most people don’t really begrudge the inventor of something truly new, novel and useful the right to force royalties from everyone using the invention. This sort of quid pro quo is what the Founding Fathers imagined when they deemed that inventors would get exclusive rights for a limited time (the current period is 20 years) in compensation for disclosing the invention to the public; after that time, the invention is free to everyone. This is a pretty good deal for the public if the invention is something like a vaccine for a big disease.

But no one is happy when a bright person grabs a patent on something that doesn’t seem particularly new or novel — especially when he or she demands royalties. This is just what some of the “inventors” of the Internet patents are doing, to the consternation of many. They’re sending out letters demanding payment of royalties and backing the demand with the threat of a lawsuit. Half the people who get the letter are wondering, “Is that really patentable?” — and the other half are kicking themselves for not filing the patent first.

The good news is that most of the truly silly patents will fall by the wayside, and the patent system still works for protecting serious and important ideas, providing ample reward for the folks who do something like cure AIDS. The bad news is that the U.S. Patent and Trademark Office isn’t particularly helpful, and the process only works when fueled by plenty of cash.

To understand how we ended up where we are today, you must understand the patent process — and how clever inventors and their patent lawyers are preying on its weaknesses. The system begins the instant an inventor says, “Eureka!” and starts to scribble down a description of the invention. This description is known as the “claims,” and it is a structured list enumerating all of the important parts of the invention. The inventor tries to make this description as broad as possible in order to claim the maximum intellectual-property turf.

The patent office has the job of examining the claims and canceling all but the good claims. About this, the law is pretty explicit. For a claim to be good, it must be new, it can’t be “obvious” and the inventor must really be the first one to invent it.

Clearly, one person’s “obvious” is not the same as another’s, so this term generates many fights. But an inventor has many different types of appeals available, and the inventor’s lawyer can make life difficult for the patent examiners — who are often judged by how many applications they process each year. The potential patent holder argues and argues and argues with the patent office until the arguments over “obviousness” fall by the wayside. This explains why so many people are thinking, “That’s so obvious.”

These arguments between the inventor and the patent office aren’t entirely worthless because they often narrow the scope of the claims dramatically. In many cases, the patents aren’t really as all-encompassing as they might seem because the negotiations have limited the breadth of the claims.

For instance, patent 5675734 — one of Hair’s patents for online pay-per-view — doesn’t really apply to all sales of audio or video over the Internet. One claim requires that the signal be copied into a “sales random access memory chip which temporarily stores a replica of the coded desired digital video or digital audio signals purchased by the second party.” If your Web site doesn’t have a “sales random access memory chip” or some equivalent, then the patent doesn’t apply to you. Patent 5675734′s claims also specify that money is involved. That is, a person must provide “a credit card number … so the second party is charged money.” If there’s no money exchanged, then the patent probably doesn’t apply.

The give-and-take between the patent office and the inventor’s lawyers often narrows the focus to something that seems non-obvious. But establishing who was the first person to think of an idea is a trickier problem. In theory, a patent application must contain a good summary of all of the “prior art” that is out in the world — that is, a summary of all of the similar ideas that were previously invented. The new patent only covers the new ideas that weren’t previously invented, and the scope of previous invention is defined by this search of the prior art.

In practice, no one can do a perfect job with this process, because no one can possibly search all of the world’s prior knowledge. So virtually everyone does a half-assed job — they poke around for a bit and then give up. More diligent people will look around a bit longer to find examples; the fly-by-night inventors will spend much less time. But no one has time to search the world as thoroughly as it should be searched.

The Patent Office tries to check the search and conduct its own, but is has a limited budget for searching, and it’s not allowed to talk about the applications. That means the examiner can’t call up IBM and say, “Didja ever think of selling videos over the Internet?”

In practice, the job of searching for prior art falls upon the shoulders of the other companies being sued for violating the patent. They often have the most incentive to find it, and they’re usually much more resourceful than either the inventor or the patent office. (New online resources like IBM’s Intellectual Property Network are helping make this process a bit more accessible.)

It usually isn’t hard to come up with some credible prior art. I wrote a book in 1995 called “Digital Cash” and it described several systems that were pretty close to 5848161 (the encryption for credit cards idea). The application for 5848161 was filed on May 16, 1996, so my book preceded it. That should be a potent weapon for anyone threatened by this patent. But there are also complications: I might not describe exactly the same system as patent 5848161, so the book might not disqualify all aspects of the patent. Plus there are thousands of loopholes and grounds for argument in the system. Nonetheless, the material in the book can’t be claimed as an invention by someone after the book is published.

Andrew Milne, an engineer for N2K, is evaluating what patents 5191573 and 5675734 mean to his company’s plans for selling music over the Internet. He’s already been doing research looking for past products and services that might qualify as prior art, and he’s uncovered a wide range. One Wall Street Journal article from 1981 that he cites describes a system for downloading digital music from a Western Union satellite. There are so many other examples, he says, that it’s hard to know where to begin. Still, he is not sure how his company will proceed — the final decision lies with the lawyers who will analyze the results of his survey.

Scott Sander, the president of Sightsound.com, the owner of the Hair patents, says he’s interested in what Milne turns up, but isn’t too worried. He says that his company has been negotiating with many major content, software and computer companies since 1993. Many of these have conducted their own search for prior art and given up.

“Do I assume that people will try vigorously to avoid giving us a seat at the table? ” he asks. “Absolutely. But that’s what patents are for — to give little guys like us a seat at the table.”

Many patents fall by the wayside because of good searching by people being threatened by a lawsuit. This might not be the best way to run a system — it forces people who are ultimately innocent to defend themselves against charges of infringement — but it’s the way the system has evolved. Searching is just too complicated and expensive.

Greg Aharonian, a San Francisco patent agent specializing in high-tech patent searching, publishes an e-mail newsletter about patent issues that often takes incendiary positions. He feels that the government, or some of private industry, should fund a better database that maintains some order about computer-related patents.

“You talk to, say, the biotech or the chemical people — they don’t really complain as much about the quality of the patents in terms of stupid stuff being patented,” he says. “The reason is that they have access to two very good databases: Chem Abstracts and Medline. These are databases that have additional content. They have staffs that add value. If you have a new idea and it’s in one of the databases, then you’re not going to get a patent.”

Still, he also points out that ideas in computer science are particularly prone to shape-shifting, and it’s difficult to keep all of the terminology straight. “Do you consider a Cobol module type of software technology that’s 30 years old an ‘object’?” he asks. “I’ve heard some people make the argument that Cobol’s been ‘object-oriented’ since day one. Would I consider an old Cobol program a reference on an object-oriented patent in the ’90s? It’s hard to tell.”

There are deeper philosophical problems that also haunt the system. The law has always forbidden patenting “laws of nature,” and that prohibition could cover anything from the principles of mathematics to ways of doing business. But many people are getting patents in these areas by formulating their claims as machines for doing the same thing. In other words, no one would be granted an overall patent for exchanging coins for candy bars, but it would be perfectly acceptable to get a patent on a new kind of vending machine.

The rise of the digital computer and the Internet has blurred distinctions even more, because computers and automation are now part of all businesses. That’s why no one ever thought of getting a patent on selling records for cash, but someone wants to patent a mechanism for selling audio files over the Internet.

Part of this problem might be remedied if the patent office were able to say that to take a past practice and conduct it electronically or online is now “obvious.” That is, shipping audio signals by digital signals is pretty much equivalent to shipping them by vinyl disk — an inventor would need to come up with a more significant improvement than that.

This will probably happen over time when one of these cases is litigated. If a court makes such a ruling, all of the older patents will lose some of their value because they won’t be as enforceable.

Many big computer companies are developing their own defense against the vagaries of endless patent litigation: They simply cross-license their patent portfolios with each other. It’s not unusual to discover that two mammoth competitors like Xerox and Canon have cross-licensed their entire portfolio of patents to each other. This saves them endless fighting, while it keeps out the little competitors.

Anthony Clapes, IBM’s former assistant general counsel and author of the book “Softwars,” downplays the sinister aspect of this trading. “They’re not doing it to be unfair or conspire,” he says. “In the Cold War, there was a certain amount of making available information about what was going on on either side. They had the red phones. There was a certain amount of pressure being released by providing information through back channels. That’s what this is like. I’ll cross-license, you’ll cross-license, and we’ll get enough freedom of operation.”

He also downplays the disadvantage to smaller inventors. “In theory and in philosophy, I don’t think there is anything that favors the larger,” he says, but adds, “It’s just that life favors the larger entity.”

But Aharonian, the patent agent, has a different view. He feels the big companies care little about patents because the value of even a great patent is only about $20 million, which adds little to their balance sheet. The little companies don’t have the ability to influence Washington.

He says, “No one wants to spend a year or two arguing whether the patent is valid. If we want to encourage the small people and the start-ups, we’ll have to clean up the patent system. Those people aren’t being served.”

Sander of Sightsound.com agrees with this pronouncement. “During the turn of the last century, when technology was really exploding, people were putting forth the same arguments that the Patent and Trademark Office had lost the ability to keep pace with that much innovation,” he says. “It’s essential for companies to protect their intellectual property in times of great change.” Otherwise, he adds, “It all ends up in Redmond, Wash.,” in the hands of the biggest companies like Microsoft.

Normally, no one would much care about such a technical detail. Most people confuse the term “kernel” with the honorific given to people like Colonel Sanders, and they don’t care about their software as long as it runs and doesn’t crash on New Year’s Day 2000.

But Linux is different these days — it’s caught in a hype storm that makes everyone want to understand what makes it tick. It’s a free operating system developed by a loosely knit coalition of developers who’ve donated their time and effort to building it. The source code is given away for free in regular periodic updates (this week saw the release of the long-awaited Linux 2.2 kernel); programmers grab a copy for themselves, use it and occasionally donate their own fixes to the process.

Now Linux has reached a critical mass where it provides many of the same services as commercial operating systems like Windows, but at no charge — and many people are wondering whether it could dethrone Microsoft. If it does, it will be because of lots of people like Rob Newberry who are making small contributions to the big effort.

The kernel is the core of the operating system responsible for making sure that different parts of the computer communicate with each other. It’s the computer’s combination basement, boiler room, kitchen and loading dock — and only the bravest programmers actually enjoy the heat of working there.

The roster of contributors to the Linux kernel has become a kind of pantheon of respect — not unlike the academies in various intellectual fields. Linux software is well-built, in part, because a team of dedicated programmers tests new code and guards the kernel with a fair degree of pride. New software is only introduced when it makes a good contribution.

The work Newberry did won’t affect many people, and most won’t even understand what it does. “I contributed some changes to the Appletalk stack that’s in the Linux kernel. That makes it easier for a Linux machine to offer dial-in services for Macintosh users. If they wanted to use Appletalk over PPP, the support wasn’t really there,” he says.

What he’s really saying is that he’s a tinkerer and he wanted his Macintosh to speak with Linux machines. He was able to make this happen because Linux comes with all of the source code available. Once he did so, he donated his code for all of the Mac-Linux users who might follow him.

That’s not really a big deal in the whole scheme of things: Mac users are a small minority, and there still aren’t many Linux users compared with the Windows hordes. But when these little efforts are repeated time and time again, they add up. There are now thousands of people who’ve done similar things. And this is why some observers are certain that Linux will eventually be a major competitor for anything from Microsoft, Apple, Sun, Compaq or IBM.

The odd thing is that Group Logic won’t be making any money off of Newberry’s work. Naef says he is “technically” Newberry’s boss and thus responsible in some way for making sure that their company keeps making enough money to stay in business. Yet here he was, bragging about the free work his employee was giving away. That says a lot about how the Linux movement is able to thrive.

Naef is pretty straightforward: “It obviously makes him kind of happy. People get a lot of credibility in the hardcore programming community if they’ve got some cool project that they’ve done. It’s like a badge of honor,” he says. Newberry worked on the project in his own time, but the work overlapped with some of the projects he does for Group Logic.

Newberry is deferential because he doesn’t want anyone to think that he’s claiming he did anything special. “On one hand, I’m excited to be talking to you,” he said. “But I do think that there are a lot of other people who’ve made more significant contributions to the kernel. And I wouldn’t have been able to do it if it wasn’t for the other people who’ve done this before.”

He speaks about the project in much the same way that Kevin Costner taught Tim Robbins to speak about joining the Major Leagues in the movie “Bull Durham.” He says, “I write software because I just love doing it. I get a certain amount of satisfaction from the work, but I get a certain amount of satisfaction out of helping people. Improving Linux, and especially its integration with Macs, has been a pet project of mine for some time.”

But there’s also a certain amount of self-interest lurking in the generosity. Information is a strange commodity that’s traded much more frequently than it is bought. In many cases, the most valuable information can’t be paid for with money — it can only be accumulated by trading equally important information. Newberry’s gift to the Linux movement means that he now has better access to other experts’ time and energy.

Naef is quite sure this helps Group Logic, and says, “He’s plugged into that community and mailing lists a lot more. There’s code out there that can be incorporated into computer projects. It can cut your development costs if you can find stuff you can use.” In other words, Newberry’s connections may point him toward other free software that might save Group Logic days, months or even years of development time.

The two also admit that work for Group Logic must take priority: If Newberry wished to donate free code to the Linux kernel and that could somehow hurt the company’s revenues, fights could follow. “We’ve pretty much had to define this up front. We have to make sure he doesn’t conflict with the best interest of the company,” Naef says.

To illustrate the point, he suggests: “If he worked for Cisco and put some border routing protocol in the Linux kernel [which would make it easy to replace a Cisco router with a cheaper Linux-based box], then they would probably be really upset about it.”

Alan Cox is the lead programmer responsible for Linux networking enhancements; he ushered Newberry’s code into the kernel. Explaining what drives people to contribute to the project, he says, “I’ve often had code for cited reasons like ‘It was bugging me that …’ ‘This fixes a problem I’m having,’ ‘I wanted to use my XXX.’ We also get code from research projects.”

More companies are catching on to the game and asking their programmers to make sure that their products work well with Linux systems. Many device manufacturers now contribute drivers written by their internal programmers. Cox says that these “donations” are becoming more common as companies want to make sure they can sell devices to Linux users.

But none of this hidden self-interest changes the fact that the Linux project now has a buzz that attracts talented people. Adding new revenues to the Group Logic bottom line might benefit Naef and Newberry, but it doesn’t animate their voices or light up their faces. There’s a pride and even a little personal glory involved in contributing to Linux that can best be described by the word that keeps creeping into these programmers’ sentences: cool.

While most people may see two sumo-sized companies battling for superiority, the Java realm is evolving differently — becoming a strange blend of small start-ups, jealous corporate partners and bigger monoliths bumping, jostling and pushing for control. While everyone cautiously agrees that Sun’s rigid standardization is an important feature, there’s plenty of griping about who pays whom for what, and many small companies are running in, looking to capture some of the crumbs.

That may explain how the JavaWorld magazine award for the best Java virtual machine went not to Sun but to a small Berkeley, Calif., start-up named Transvirtual — a company that does not even pay Sun’s Java licensing fee.

Transvirtual is a great example of a company trying to carve out a niche where it might win over some of Sun’s business. The company makes a version of the virtual machine — the translator that lets cross-platform Java software code run within a specific operating-system environment — that requires a smaller amount of memory than Sun’s. And since Transvirtual is pursuing an open-source strategy, desktop users can download a free copy from the Transvirtual site, along with all of the source code.

The company isn’t doing this out of pure generosity: It hopes to attract attention and bug fixes by releasing the version for desktop systems for free, and plans to make money by selling another version — one that’s tuned for appliances like cell phones or TV set-top boxes.

“Our business is making Java for embedded devices like cell phones or Web phone,” says Transvirtual CEO Tim Wilkinson, “for companies that want to put Java in some sort of a box that isn’t a PC. The desktop version isn’t right for them because it’s too big.”

Miles Jones, head of marketing for Transvirtual, says that licensees of the commercial version of Transvirtual’s Java machine are getting something for their money: “Our custom edition offers its own AWT graphics library [which provides user interface tools], is portable to literally any platform, offers a very small footprint [and] runs more than twice as fast as the desktop version.”

Sun, at least in its public pronouncements, isn’t too worried about Transvirtual. Jim Mitchell, Sun’s vice president for architecture and technology for Java, says that he’s not too familiar with the company. “There are lots of clean-room versions,” he says — referring to the industry practice of cloning a language or a chip by employing programmers who work from specs to duplicate a product. Sun encourages such efforts, Mitchell explains: “We didn’t do this to give up a revenue stream, but to encourage the Java language.” He points out that Sun embraces many clean-room versions and supports them once the companies pay Sun for a license to the technology.

But that’s not a step Transvirtual has chosen to take. Wilkinson says that he has no problem paying Sun to test the compatibility of his version of Java, but he’s hesitant to buy a license because of its constraints. “I think that their relationship with Hewlett Packard shows that Sun really wants to have control,” he says — referring to the tumultuous battle between Sun and HP that led HP to start its own cloning program for Java. For now, Wilkinson wants to go it alone.

Transvirtual is clearly betting that its greater openness will draw users and customers to its software. Sun has recently made some high-profile moves in the direction of open-source code, but according to Wilkinson, “Sun’s open source goes something like: You can go and get their code, you can play with the code and you can bug-fix it. If you’re an individual or a university, then it’s no big deal. But if you want it to be commercial, you’ve got to go back, make sure it passes their specification and license their brand from it.”

Transvirtual users don’t have to license anything; they got the code for free and they can use it as they like. On the other hand, they are constrained by terms drawn from the Gnu Public License, the original “free software” contract: If they release a product, they must also include the source code. This can be just as constraining for companies that want to keep their source code proprietary.

Sun’s Mitchell argues that all of this freedom is nice, but it runs the risk of compromising Java’s big feature — the ability to run consistently on many different operating systems and platforms. He says that there’s little Sun can do to stop independent clean-room operations like Transvirtual, so its goal is to provide good services and support to those clean-room operations that have obtained Sun’s license.

“[Transvirtual has] never come to us to get the compatibility test,” Mitchell says. “It’s a really hard thing. I would almost guarantee that they don’t have a compatible clean-room application. It may be pretty close, but it’s so hard to get compatibility.”

In the end, he predicts that Transvirtual will come join Sun to ensure perfect fealty to the Java specifications: “They’ll have a certificate to prove it. They’ll get the code. The certified clean-room implementation will always have a time-to-market advantage and a compatibility market advantage.”

Transvirtual is not the only company trying to build its own version of Java. NewMonics, for instance, is concentrating on producing a version that provides guaranteed performance for applications requiring predictable responses. Microsoft also continues to build its own version.

The fact that small companies are filled with people willing to bet some of their time on a technology is usually a good indication that it is maturing and winning acceptance. Java is now a common tool in many large companies, where its users are intensely interested in achieving the best performance possible.

That leaves Sun with a difficult decision. It wants to keep Java homogeneous and uniform, so developers will experience a bug-free existence as they move their software from platform to platform. But it can’t stop cloners from building clean-room versions, and users like the results of competition and choice. Perhaps Sun will be able to find a middle ground and eventually work with companies like Transvirtual, who are both partners and competitors.

Jones seems to think that this orchestrated competition is a viable solution. He says of proprietary approaches like Microsoft’s Windows CE, “I think in the appliance marketplace, proprietary systems will lose. There’s no point in hitching your star to one company. Once you do, you become completely dependent on them. With a standard system like Java, you’re assured of fair treatment
because you can acquire compatible solutions from multiple vendors.”

The victory dance is likely to be short-lived, however — because the content companies are about to discover that they’re also the new law’s main victims. These companies may love the part of the law forbidding illegal copies, since they’re the main copyright holders. But the restrictions on tools are going to be a major headache for the artists in their stables. Content manipulation tools on the market — like Adobe Photoshop, Avid’s video editing software or many sound-editing programs — will become less useful and begin to disappear, and the ones remaining will become more expensive.

Fritz Attaway, general counsel for the Motion Picture Association of America, says everyone in his organization is celebrating several long years of hard work to convince Congress that the legislation was necessary. When asked about the negative effects of the bill, he said, “I don’t really see a downside.” But real life laws rarely have such simple Hollywood outcomes.

How can the content industry be both victor and victim in this crazy case of regulation gone wild? Think about how artistic works are created: First, artists sit down with some tools; then they think deeply about the human condition and grouse a bit about how the world doesn’t understand their perfect vision; then, when the deadline looms, they produce a work of some kind. The tools they use may be paint brushes, pencils, paper, videotape machines, film and a zillion other media — including computers.

Every one of these tools can be used to make illicit copies. But the last is the one that upset Hollywood and the content conglomerates. The lawyers for the content rights holders realized that computers make it easy to produce perfect copies. Meddlesome kids were starting to use computers to make digital copies of songs over the Internet in much the same way that kids used to make cassette tapes.

Of course, the content industry lawyers didn’t stop to think that their own artists used the same tools as those punk kids. In fact, the content mills might even do more cutting and pasting than the rest of the world put together.

“Films are released first in theaters, then after some time they’re released on home video, then pay-per-view, then pay channel, then free TV. That’s the distribution chain that lasts anywhere from two to four years depending on the film,” Attaway says. “The reason that the distribution to the home comes so late on the distribution chain is because there’s no way to date to prevent copying.”

The content industry’s solution is to push Congress to make laws against the business of making tools that can make copies. Now, thanks to the new law, anyone who makes a tool with the “primary purpose” of circumventing a copyright protection mechanism will face fines of “not less than $200 or more than $2,500 per act of circumvention, device, product, component, offer, or performance of service, as the court considers just.” People who do this for commercial gain face fines of up to $1 million and up to 10 years in prison.

Hollywood is clearly taking aim at people like the garage electronics wizards who produce devices that descramble premium cable television services. The gadgets, which are often sold illicitly, make it possible to get HBO, Showtime or other expensive signals without paying for them. As the Internet evolves, there are bound to be more scrambling systems put in place, and Hollywood wants to be ready to prevent digital brigands from circumventing them.

But in this process Hollywood never really bothered to think much about who is really doing the most “circumvention” — or what a tool for “circumvention” really is.

The best way for most readers to grasp this dilemma is to use the basic cut and paste or copy and paste features built into all Web browsers and computer systems. You can test this out by copying the text of this article and pasting it in, say, the window of a word processing program. (Don’t paste a copy in your e-mail program and send it to a friend, though — that would be violating the copyright.)The big content combines in Hollywood would like to believe that there’s a way to distinguish between good cutting and pasting and bad cutting and pasting. This is probably a side-effect from watching too many of their own features from the 1950s, in which the characters’ morality is established in the first two minutes of the film.To solve this problem, the content conglomerates are pushing technical solutions that essentially lock up artwork. Editing devices are supposed to look for a particular embedded no-copy mark and refuse to copy if such a mark is found. (Such marks are often dubbed “digital watermarks,” and the term is apt because they are usually barely detectable.) Some of these features are already included in many lower-end editing tools and videotape decks. Only the “professional” ones that cost substantially more can slip around the restrictions. Adobe Photoshop, for instance, will now scan files to look for an embedded watermark that might identify the copyright holder. Today, it just provides the information; in the future it might simply refuse to let the artist do anything with the file.As the new law’s restrictions begin to be built into all levelsof technology, the big copyright combines in Hollywood are going to start discovering how onerous the new limitations can be. Imagine some artist at Walt Disney is given the assignment of creating a box for holding the videotape version of some recent film, like “Pocahontas.” In the past, the artist could take a copy of the DVD version of the film, stick it in a computer and cut and paste an image of Pocahontas from there.No one knows what will happen in the aftermath of the new bill. Devices and software are prohibited if they are “primarily designed or produced for the purpose of circumventing protection afforded by a technological protection measure that effectively protects a right of a copyright owner.” They’re also prohibited if they have “only limited commercially significant purpose or use other than to circumvent protection afforded by a technological protection measure.” It is not clear what the software companies will do, but they may remove the feature that lets people take out still frames from DVD movies. The big content companies would probably be overjoyed — this would stop those pothead kids from snipping out a picture of Pocahontas, sticking a joint in her lips and posting it on a Web site. Sure, a few technically savvy folks in Peoria would have trouble using recordable DVDs to distribute home movies, but that’s not Disney’s problem, right? In their eyes, the primary purpose to cutting and pasting from DVD movies is to interrupt the revenue stream of the major conglomerates. In essence, the bill is banning the creation of tools for picking the electronic copyright locks the content factories will stick on their products. Everyone knows that locks work fine in most cases, but there are always cases where locksmiths need to circumvent them. The bill seems to prohibit the creation of these devices.Attaway and others feel that this won’t be a problem. Of course, he says, there will be tools for legitimate circumventing, “if people are authorized.” But how will these tools manage to distinguish between the good circumventers and the bad ones?Attaway says he doubts this bill will have any effect on the marketplace for editing tools. “I don’t think anyone on the content side thinks that those types of devices will be prohibited,” he says.At the same time, one of the MPAA’s sister industry groups, the Recording Industry Association of America, is continuing to press to stop the distribution of the Rio, a digital music player manufactured by Diamond Multimedia. Although the Rio can only play back digital MP3 recordings, it must copy the music data from the Net to its internal memory in order for you to listen to a track — and the association was trying to stop its distribution even before the new law came into effect. While Attaway doesn’t see the lawsuits coming, who knows what others will try to do with the new legislation behind them?All of this litigation and regulation is bound to make the marketplace for editing software more restricted. Fewer small companies will enter the arena with neat tools for fear of being sent to prison for 10 years; others will struggle to devise and implement new kinds of tools for automatically distinguishing between the right and wrong kinds of cutting and pasting — and between good and bad users.This will drive up the prices of editing software dramatically. First, the greater complexity will make life harder for the programmers who write the software, who will now have to integrate copyright juggling mechanisms into all levels of the editing tools; second, the smaller “legitimate” marketplace will mean fewer sales over which to amortize development costs.And who knows how that poor graphic artist at Disney is going to get a picture of Pocahontas to use for the package? Maybe Disney will keep a special computer that acts like a vault for storing versions of its properties that aren’t copy protected. That graphic artist may be able to get access after going through the proper channels — but the artist for McDonald’s working on a cross-licensing deal will have a harder time of it.Maybe Disney will try to buy special, super-secret tools for unlocking the copy-protected images. These master keys are going to have to be even more carefully guarded than non-copy-protected content. If they slip into the hands of the pirates, so much for the copy protection system. It’s likely that the content conglomerates won’t even know why their development costs will skyrocket in the future. The companies are vast organizations in which the attorneys pushing for more and tougher laws are disconnected from the artists struggling with the tools they need to create. The MPAA, the RIAA and the other content-industry lobbies are going to crow about their legislative successes. In the meantime, the graphic artists, the film editors, the sound editors, the commercial creators, the marketing staff and practically all of the employees on the content assembly lines are going to be walking a few extra hallways, signing a few more forms, taking a few more data-juggling steps before they can get their work done.Another way to think about this is by looking backwards through time at how new copying technologies spurred the development of the content industry itself. In the Middle Ages, monks copied sacred texts by hand. The movable-type printing press led to a profusion of books. Now, after the development of the photocopy machine, the fax machine and the Internet, a wider selection of books is more widely available to more people than ever before, and mega-bookstores proliferate. As copying became easier and easier, in other words, the content conglomerates only grew bigger and more profitable. Why would they want to turn back the clock now?