August 21, 2005

Microsoft’s iPod patent?

By: Tim Lee

Will Apple have to pay Microsoft royalties on every iPod it sells? That was the prediction of some gleeful Apple-haters last week as the patent office rejected an Apple application for a patent on the iPod’s “scroll wheel” interface. Microsoft had beaten Apple to the punch, filing a conflicting patent several months earlier. But more sober commentators soon noted that the rejection was hardly a final decision. Analysts predicted that Apple would amend its filing to avoid the scope of the Microsoft patent, and Apple wouldn’t have to pay Microsoft a cent.

Still, it’s worth asking a more fundamental question: in what kind of crazy legal system is it even conceivable that Apple would have to pay royalties on a product it clearly invented? After all, Apple released the first iPod months before Microsoft filed its patent application, and no one seriously believes that Apple’s design for the iPod was copied from technologies developed at Microsoft. So why did Microsoft and Apple both feel the need to race to the patent office to patent a relatively banal aspect of the iPod’s design?

For more than two decades, the software industry has been engaged in a destructive patent arms race. It hurts the software industry generally and does particular damage to small developers who can’t afford large legal bills. Fortunately, there is a simple solution: Congress could put an end to the arms race by explicitly removing software from the scope of future patents.

Patents are designed for industries in which a great deal of research and development is required to produce a new invention. The pharmaceutical industry is a classic example. It costs hundreds of millions of dollars to develop and test a new drug, yet the finished product is rather simple. It can usually be described with a simple chemical formula.

Without the protection of patent law, it would be impossible for a drug company to recoup the initial investment. The moment the first pill is released, competitors would study it, determine its chemical formula, and produce generic competitors. Generics would hit the market long before the company that developed the drug had recouped its initial investment.

Software development is very different. The value in a software product does not lie in one “deep” discovery that requires millions of dollars of research to find. Rather, software consists of hundreds of interconnected components that perform different, relatively straightforward tasks. A word processor, for example, must have functions that handle user input, formatting, word counting, spell checking, printing, etc. What makes a software product valuable is not that any one of these features is particularly novel–any reasonably bright computer science student could implement any of them in a few weeks–but the way that the components fit together to form a complete product.

Apple’s patent filing illustrates this point well. It has a laundry list of features of the supposed invention, including the device’s size, the color of the backlight, and the fact that the scroll wheel makes a clicking noise when it’s turned. It’s not clear why that particular combination of features should be protected by a patent. While the iPod is certainly an innovative product, what makes it innovative is not the concept of a portable music player (there were many already on the market) or any particular feature (virtually all of them can be found in previous products). What makes the iPod great is the seamless integration of the many components into a single, elegant product. It is good design choices and attention to detail, not any one breakthrough “invention,” that makes the iPod a best-seller.

According to statute, only “non-obvious” inventions can be patented. One would think that the idea of using a scroll wheel to choose a song from a playlist would be an obvious idea. Unfortunately, the patent office has done an abysmal job of enforcing the requirement. An infamous example is the Amazon one-click patent. It protects the “invention” of storing a user’s registration information on a server so that he or she can purchase products with a single mouse click. As a result of the patent, every other e-commerce site must either license the patent from Amazon or ensure that its purchasing process uses at least two clicks to avoid the risk of a lawsuit.

A big part of the problem is that the patent examiners are usually not computer programmers. Programming is a complex craft, and what is obvious to an experienced programmer is very different from what is obvious to non-programmers. One-click technology probably seemed non-obvious to non-programmers in 1999, when the one-click patent was granted, because the technology involved was new, and therefore unknown to the general public. However, in light of the technologies available at the time–servers, web browsers, cookies–implementing one-click shopping could have been implemented by any smart computer science student in a few days. It was hardly a breakthrough.

However, it would be a mistake to view the one-click patent as an aberration due solely to incompetence at the patent office. Software “inventions” are rarely, if ever, “deep” discoveries like the creation of a new drug. Major software companies have been feverishly applying for hundreds of patents that mostly resemble Apple’s iPod patent: a mere laundry list of features, none of which are especially groundbreaking. The companies hope that the patent office will acquiesce to some of them, giving them more leverage in negotiations with competitors. Every large company is forced to join in the patent race lest they be caught without ammunition in a future legal battle with competitors.

Wouldn’t prohibiting software patents make it difficult for software companies to protect their investments? Hardly. The expensive part of software development is in getting the implementation details right. Those details are well protected by copyright law. If a company developed a product that worked in precisely the same way as a competing product, the competitor would have a strong case that the company illegally copied its code. If, on the other hand, a company made a product that offers the same functionality but was clearly developed independently, then that seems like good, old-fashioned competition.

Voiding software patents would allow the software industry to refocus its efforts on what it does best: developing great products. Amazon could focus on offering ever more products at lower prices. Apple can focus on developing the next great gadget. And Microsoft can focus on what it does best: ensuring that next year’s version of Windows is almost as good as last year’s version of Mac OS X.

Tim Lee is science and technology editor of Brainwash. He lives in St. Louis, MO, where he his helping to start a new think tank. His website is