Software Patents are Broken


If there has ever been evidence that software patents, at least as practiced in the United States, are broken, this is it: Today, the US Patent Office published Microsoft's patent application 0040181571:

Reducing unwanted and unsolicited electronic messages by preventing connection hijacking and domain spoofing

This application nicely demonstrates what is broken about such patents:

Banging Nails is Obvious, No Matter Where in the House You Bang Them

Patents aren't supposed be able to claim things that are obvious to anyone skilled in the art. Plowing through this patent you can extract that they are claiming many basic Internet concepts like: "look up a list of authorized IP addresses". In my experience with patents this sort of thing is abundantly prevalent.

Patents fail software because software is a very new field and is still developing. What is "obvious to one skilled in the art" hasn't had time to become implemented everywhere yet. Just because no one hasn't used XML to describe mail authorization yet doesn't mean it isn't completely obvious to any skilled computer programmer. And it shouldn't be patentable.

Claim Everything, Applogize Later

Software patents are invariably written with starting claims that cover the most broad idea imaginable. Usually so basic so as to be obviously not patentable. The the claims are slowly built up, one tiny idea at a time. Surely it is a patent holder's favor to have as many claims as possible -- that way the judge and jury can feel like they are being "fair" to both sides when they throw out half of them... even if you knew those were bogus. And what's better, you may end up with a bonanza if some of them survive!

Which Mueseum Houses the Prior Art?

The filing date on the application is October 10, 2003. There is so much prior art on exactly the claims made, I'm not even sure where to begin. I'm talking right-on-target exactly-what-is-claimed public disucssions and submitted internet drafts. Just one example: Sender Permitted From: A Convention to Identify Hosts Authorized to Send SMTP Traffic

In the 19th century, getting a paper published meant that your ideas could be found, and count as prior art. Today, especially in software, journals don't count for much. The vast bulk of the ideas and developments are documented in the hugh swell of the web. But just because it doesn't get cataloged by university librarians, doesn't mean that it isn't prior art. Apparently, the US Patent Office has decided that since there is just so much to look through, they aren't going to look at all.

Priests of an Ancient Language

Patents are written in an arcane legalese dialect that doesn't correspond to what any one skilled in the art of software speaks. Above all disciplines, software has several highly developed methods of expressing precisely what is meant. From programming languages to the developed style of technical standards, we have learned to say what we mean.

Writing software patents in the style of the 19th century only leads to difficulty in interpreting them. It serves to increase lawyer's time, since they are the only elite entrusted to read and write this sacred language. It also serves companies well who benefit by enabling the interpretation to be made in the courtroom, after a legal challenge has been made.

To promote the Progress of Science and useful Arts

It's why we have patents, right there in Section 1, Article 8, Clause 8 of the U.S. Constitution. For the public to grant the temporary monopoly on an invention, the public gets the invention to be publicly disclosed. But for the vast majority of software patents, the public isn't getting anything since the ideas were already being developed, or are obvious. Granting software patents simply rewards the companies with biggest teams of lawyers for being first to write up ideas in the arcane language and rites that is patent application. It is a government granted monopoly for already rich companies.

If the current trends continue, the huge numbers of software patents sought by large corporations will have have quite the opposite effect from promoting progress: The very patent above has caused a great outcry and upset for a dedicated group of people trying to stop e-mail spam. They have had to alter their technical course away from what is best and innovative because of the existence of a patent. They have done this even though to many the patent claims are either obvious or claim things that were developed by the community prior to the patent application: It is simply too risky to face litigation.

The last half century has shown that software doesn't need the disclosure of patented ideas in order to make progress. It is time to stop the practice that only benefits privileged corporations, and chokes the progress of (computer) science and useful (programming) arts.