What's next: a patent for clicking OK?.
C|Net reports that Amazon as won a patent for hiding and displaying parts of an order form. While the ability to patent business processes could make sense in some cases, the whole patent situation is quickly getting out of hand.
In the software industry, patents are generally used for defensive purposes only: large companies like Microsoft, IBM, and others have an enormous portfolio of patents which are used to cross-license with other companies and to protect them from being sued because it's likely that the company suing them is violating one an existing patent, which provides great legal leverage.
This is clearly not what the people who created our patent laws had in mind. The idea was to reward and provide an incentive for innovation. Instead it has become prohibitively expensive for small innovators to file patents regularly, and it has become close to impossible for the patent office to give every patent application the scrutiny it deserves.
One of the problems with patents today is that unlike trademarks you don't have to enforce them to hold on to them. For example, Unisys did nothing to enforce its patent for Lempel-Ziv-Welch (LZW) compression for years. The LZW algorithm was incorporated in the GIF format, and after the GIF standard was well established, Unisys started knocking on doors to collect dues.
Of course, if everyone would start enforcing patents today, it would instantly paralyze our whole industry. This is proof that the lack of enforcement is hiding a much more fundamental problem with patents: As much as they can encourage innovation, they can make virtually innovation impossible.
Patents may work well for revolutionary ideas that change everything instantly, but that's now how innovation usually works in our industry: Software improves through a series of incremental changes.
For example, first there was the Escape key, then there was Undo, then there was Multi-Level Undo, and then there was Redo. Because enhancements are usually incremental, it's also likely that multiple people have the same idea at almost the exact same time… not because they are all geniuses but because it's the logical next step in innovation.
Every few years an invention comes along that everyone would agree deserves to be patented. The spreadsheet is a great example. How do we distinguish these major inventions from the small incremental enhancements? Without hindsight, it's impossible. The current system is clearly broken, but coming up with a better patent system seems like an impossible task. [Live @ Sax.net]