Sudo You?

Whenever I write about patents, trademarks, and copyrights, I'm always careful to state two things up front. First, I'm not a lawyer. It is quite possible that I am missing some part of law that makes my opinion invalid. I try to understand the technical underpinnings of the patents and see what it means to the computer industry and economy at large. I'm not trying to be an attorney.

Second, I am completely in favor of intellectual property protection. I am not one of those folks who believes that patents are evil and that all software should be open source. The fact of the matter is, intellectual property protections provide motive to continue to innovate. They protect the small inventor from having their life's work pulled out from under them by a deep pocketed company. Same goes for copyright and trademark protections. History shows that if people can't benefit from their work, they'll do something else and we all lose out on the richness of life.

Patents are monopolies granted to an inventor in exchange for adding to the useful knowledge of the world. Without them, the world would be full of virtually permanent monopolies as inventors strive to keep inventions secret rather than disclose them. It is also well understood that without the time based monopoly, many inventors would not recoup their investment in innovation and wouldn't bother inventing in the first place. When it costs nearly US$800 million to develop a new drug, patent monopolies are the only way to recoup the costs and make a profit.

That's why I took it with a grain of salt when I saw the initial commentary on the new Microsoft Patent ( number 7,617,530 ) issued November 10, 2009 and originally filed in April of 2005. You see, there are a lot of people who hate the idea of software patents. In their eyes, no software patent is valid. I'm still not sure where I stand on software patents nor am I a Microsoft hater, so I tried to turn a critical eye to the patent. Once I read it (and read it and read it and read it...) I came to a very firm conclusion: What was the USPTO thinking? This is so obviously wrong that I can't imagine how this got through.

The patent is entitled “Rights Elevator”. It describes “systems and/or methods” to allow a computer user to elevate their rights from a lower, standard user account to higher level administrative rights. If this sounds familiar, that's because it is. It has existed in UNIX systems since at least the 1970's. In UNIX and Linux we use a command called su, for superuser, to obtain the rights of another user with higher level rights. There is even a short hand version called sudo which runs a command or program once with elevated rights.

All patents have to pass certain tests for them to be granted by the US Patent Office. For a patent to be granted the claims must describe an invention that is novel (new), useful, and non-obvious to practitioners of the inventor's art. These tests are important. Without them a lot of inventions would be granted patents that should not. That would, in turn, inhibit innovation.

It is hard to argue that this patent passes at least two of the three tests. I agree that it is useful. The ability to briefly elevate rights to install software or copy files to a restricted directory has been proven to be good method of balancing security with the need to do certain important functions. I have a Unix book from the late 1980's that tells a sysadmin how to do that using existing UNIX commands like su. Therein lies the problem. It has already been proven to work because it already exists. That kind of kills the novelty of the invention doesn't it. The patent makes the argument that remembering a user name and a password is too difficult. Part of what is says is innovative is not having to remember a user name. That's a load of hooey. How hard is root to remember?

Let's be nice though. The Microsoft patent also includes a component to select a higher level account with a GUI and ask for a password. You have to admit, given the prevalence of graphical operating systems today that seems like an obvious addition. Wait! Did I say obvious! That seems to run afoul of the non-obvious test. In fact, this is something that most Linux distributions do today. They even reference Ubuntu, Debian, and Red Hat in their prior art list. How is this method any different from what is already done in Linux and Unix systems.

To be fair, this method of elevating rights temporarily with a graphical interface may not have been used when the patent was filed in 2005. I don't think that's true but I'll give the USPTO the benefit of the doubt. The method outlined in the patent, however, doesn't move far beyond sudo (which is also referenced in the prior art listing). Certainly not far enough to claim novelty and non-obviousness. It doesn't take an expert in software and operating systems to see that, never mind someone practiced in the art of system administration.

This method is so ubiquitous that everyone does this. Everyone except Microsoft that is. Windows, in all it's forms, has always required you to either have administrator rights or log in as someone with those rights, when that was possible at all. The Windows Vista UAC allowed you to override built in restrictions not elevate your rights temporarily. The UAC never even kept you from doing something. It just nagged you that it was bad. Windows 7 finally catches up with the rest of the world and Microsoft is trying to patent it. Talk about making lemonade from lemons.

How did this one get by? There could be a lot of reasons including overworked patent examiners. The patent should be overturned and likely (hopefully) will. In the meantime the patent office needs to do something. Maybe independent panels who don't work for a vendor. I'm not sure. All that patents like this do is throw fuel on the fire for people who want to eliminate patents, especially software patents.

This one should never have been granted. But then, I'm no lawyer nor do I play one on TV.