Interesting People mailing list archives
more on Patents -- What is the distinction?
From: David Farber <dave () farber net>
Date: Mon, 16 Aug 2004 15:48:40 -0400
Begin forwarded message: From: Brian Thomas Sniffen <bts () alum mit edu> Date: August 16, 2004 3:04:18 PM EDT To: dave () farber net Cc: tom_gray_grc () yahoo com Subject: Re: [IP] Patents -- What is the distinction? You're absolutely right that you should be able to patent your machine, including the software inside. What most people complaining about software patents are talking about would not interfere with that -- they (well, we) have a problem with a different type of software patent. As a couple of examples, there was a patent on using the RSA algorithm to secure data. That was fine. There's a patent on use of the LZW algorithm to compress data -- somewhat less fine. There's a patent on using XOR to draw a cursor on a display. That's completely not fine. And there's a patent on remembering a customer so that he can buy things just buy telling you he wants them -- the infamous Amazon "one-click" patent. That's so far from fine it's coarse. What's the distinction here? Well, there are two. The Amazon patent, and many other very, very bad software patents, are patents on business methods. There's a booming industry in taking any traditional method of commerce, slapping a web interface on it, and patenting it. And this is legal -- now that organization is the only one who can use that technique, for years and years. The one-click idea is obvious. There were plenty of ways of implementing it, and plenty of people who'd done pieces of it. But now, not only does Amazon have a monopoly on one-click purchases, but on many systems like that or incorporating pieces of it. These patents on business methods in software are a real, serious problem. And they share a characteristic with the other troublesome software patents: they're obvious. A patent on use of XOR for drawing a glyph that's supposed to be obvious is right up there with a patent on using a lever to lift heavy things. It's a normal, expected part of building a graphics system. It's so basic that it's definitional: a cursor *is* a quickly-drawn obvious glyph. XOR *is* the maximum contrast operator, and *is* simple to calculate. Patents like this are scattered throughout the current software development universe. Rederiving LZW is a standard undergraduate information-theory problem. That is, a bunch of undergrads will be given LZ and told to improve it to solve certain problems, and the vast majority of the class will come in the next morning with LZW. But until it expired very recently, somebody (UNISYS, more or less) has a monopoly on use of that algorithm for compression, granted as a reward for developing something innovative and non-obvious. There are over two hundred similar patents which encumber any reasonable implementation of a computer operating system. It's not possible to build a reasonable software system without infringing on many of these patents. That strongly implies that they're not obvious, and many never were. What's the solution to this problem? There are several possibilities, none of which are harmful to real innovators: * Reduce the amount of time a patent lasts. * Examine applications much more closely for obviousness and innovation. * Rescind patents when their art matures enough to consider their ideas obvious. * Eliminate business-method-in-software patents. For example, require that the technique be globally novel, not merely novel within the field of software design. But none of those hit the real core of the problem: software construction is all about abstracting ideas and methods of operation. Code reuse is common and vital. But the industry's still in its nascent stages. Patents on techniques that weren't innovative in 1988 are still restricting what people can build; that became basic, standard technique over a decade ago. The only way to fix that problem is to remove the ability to patent ways of manipulating bits with other bits. -Brian David Farber <dave () farber net> writes:
There have been some recent postings to the IP list about software patents. Frankly I cannot understand the point behind these posts and others that I have seen. I have designed hardware and I have designed software. I cannot see any real difference in the design of each. In fact, in designing a solution to a problem, I am often faced with the choice of developing a custom circuit or of choosing one of many processor/software design architectures. The choice between hardware and software does not depend on anything inherent in either technique but in various aspects, such as cost, space availability etc., that depend on the current technological context. Sine patents do not protect a design in itself but the combination of a solution and a perceived problem, I cannot see any reason why there should be a distinction between a hardware and a software-based design in this regard. If I design a piece of hardware that contains a controller which sequences circuit behavior through some custom microcode, should I not be able to patent this design. It is both hardware and software and both neither. Why should patenting not apply to software-based designs? Tom Gray __________________________________ Do you Yahoo!? Yahoo! Mail Address AutoComplete - You start. We finish. http://promotions.yahoo.com/new_mail ------------------------------------- You are subscribed as bts () alum mit edu To manage your subscription, go to http://v2.listbox.com/member/?listname=ipArchives at: http://www.interesting-people.org/archives/interesting-people/
-- Brian Sniffen bts () alum mit edu ------------------------------------- You are subscribed as interesting-people () lists elistx com To manage your subscription, go to http://v2.listbox.com/member/?listname=ip Archives at: http://www.interesting-people.org/archives/interesting-people/
Current thread:
- more on Patents -- What is the distinction? David Farber (Aug 16)