« Rear View 2008 | Main | Cutting Farts »

January 3, 2009


"Abstract software code is an idea without physical embodiment." So opined the Supreme Court in Microsoft v. AT&T, 2007. The high courts in recent years have done what they could to denigrate software as unpatentable, most recently in a stunningly incoherent ruling at the CAFC In re Bilski. This is the result of both scientific and economic ignorance by the courts, and political brainwash by computer software corporations in this country, including, incredibly, Microsoft and Apple. As emergent Asian nations race to overtake the U.S. in every technological arena, the heavy patent action here is cutting off our leading-edge nose to spite our face.

The Economist (December 30, 2008):

To see the geography of the technology industry, crack open an Apple iPhone. Although the firm that sells it is American, it provides none of the physical innards. The components are almost entirely Asian: the screen is mostly from Japan, the flash memory from South Korea, and it was assembled in China. Apple's contribution is the design and software.

The biggest and most technically clever firms are American and European, but their predominance in research, innovation and production is being challenged by Asian companies.

While corporate R&D in America and Europe grew by 1-2% between 2001 and 2006, in China it soared 23%. China is now close to surpassing Japan in total research spending, from almost nothing a decade ago.

The most impressive growth has been in South Korea. In 2007 Samsung spent more on R&D than IBM. The company has jumped to second place in the number of patents granted by America's patent office (just behind IBM); a decade earlier it was not even in the top ten. South Korean firms spend more on R&D as a percentage of sales (6.5%) than European and Japanese firms (around 5%), and are catching up with American ones (about 8%). South Korea now has more high-tech researchers than Britain and Germany.

The starkest shifts are in computer services and manufacturing, where the roles of America and East Asia have diverged dramatically. The amount that American firms spend on research in computer services as much as trebled over the past decade. Japanese and South Korean firms, meanwhile, spend hardly anything developing services, and prefer to concentrate on more tangible, if less lucrative, hardware.

In 2003, Microsoft got patent religion and began a vigorous licensing campaign that has brought in over a billion dollars in the past five years. Given their revenues, it may resemble chump change, but it represents their crown jewels: software inventions. This past year, Microsoft finally rattled its patent saber, shivering the Linux penguin.

If software processes are not patentable, American software innovations are readily replicated. Without software patents, what prevents Chinasoft from replacing Microsoft?

Posted by Patent Hawk at January 3, 2009 9:58 PM | Patents In Business


Microsoft submitted an amicus brief with Dell asking that software patents not be issued unless tied to a computer and resulting in a transformation. So MS asked to lessen patent scope. So what is going to protect MS from Chinasoft? According to Groklaw:

"What? Copyright, trademark, trade secret, faux interoperability, standards capture, and dirty tricks are not enough? They worked for Microsoft when it was building its business, prior to State Street."


Posted by: Rob at January 3, 2009 10:45 PM

Hawk, I see that Covidien is moving, from The Bahamas to the Republic of Ireland. What makes you think that "What's good for Microsoft is good for the USA"? How many of those Microsoft patents have on them the name of an Asian, as inventor? Bit simplistic, isn't it, to assert that patents on business methods benefit the USA disproportionately? Given the mis-match between business methods and patentability, the reverse proposition has more traction. Now if it were pharma patents you were going on about...well, that would be different. But, go on, tell me that I'm urging this because.... I want to see the USA fail. Give me a laugh.

Posted by: MaxDrei at January 4, 2009 7:05 AM

We are quick to blame the courts, but perhaps the fault lies with patent owners and their counsel. We tech types think it is "obvious" that software should be a patentable aspect of an invention -- but then we use words like "language", "code", "algorithm", and other intangible-sounding words to describe our software.

I am curious how these cases (such as Bilski) were argued by patent-owners' counsel.

We need to be much more careful how we define software in our arguments. After all, a drawing is nothing but an abstract representation of an invented device -- and drawings are an accepted (even required) part of a patent application. The drawing is abstract, but the invention it is describing is concrete.

In a similar way, we need to tell the courts that software is simply a way of describing how to build a device that will behave in a specific manner. We should refer to software as part of the specification of our invention. Then we need to explain to the courts that this way of describing an invention is just like a blueprint, since it unambiguously specifies how to reproduce the invention. Like a blueprint, some details are left to the practitioner (such as choosing the exact diameter or thread pitch of a bolt) while clearly specifying other essential aspects of the new invention (such as the range of motion desired in a lever).

A mechanical engineer can look at a print and say, "I know how to build that!". A computer engineer can look at a segment of software and say the same thing. Someone can infringe a drawing-specified patent even if they do not copy the print exactly; by the same token, someone should be said to infringe a software-specified patent even if they write new software that utilizes the same essential principles.

So I am curious - does anyone know whether these arguments have ever been tried in a software patentability case?

Posted by: Carl Strathmeyer at January 4, 2009 4:01 PM

Or to make my argument more briefly:

A drawing is not patentable per se, but an embodiment built according to the drawing can be patentable.


Software may not be patentable per se, but an embodiment built according to a software specification may be patentable.

Posted by: Carl Strathmeyer at January 4, 2009 5:44 PM

what is going to protect MS from Chinasoft?
Market power. According to the CIA fact sheet the 20th Century was China's century. Now China will reap the rewards of its growth and stabilization during the 20Th Century. To that end, capital transfer between the U.S. and China must be fostered and abrogating patent rights assists in that endeavour. International companies (companies that are not dependent up the U.S. for revenue) can control Chinasoft and any other company in China by virtue of ownership in the same. Considering that China is a growing economy any international company will be more than happy to see revenue growth through their Chinese subsidiaries by the abrogation of nationalistic protectios, such as patents.

In fact, the EU is a means to that end. It is much simpler for international companies to manipulate the political machinations of a single political entity such as a Federal political body as the EU that a myriad of unique and often divergent political bodies such as the member states of the EU.

Look for the one world patent coming soon to a country near you.

Posted by: Kenneth Brooks at January 5, 2009 7:38 AM

I wouldn't worry about the ATT case doing much to uproot software claiming. The issue there was essentially that *disembodied* software (i.e., *not* embodied on a computer readable medium) was not physical and thus not a component under the export statute. For typical software claims, no one ever claims disembodied software. They claim software "embodied on a computer readable medium."

Bilski, on the other hand, is one to worry about.

Posted by: anonymous at January 5, 2009 8:28 AM

"A drawing is not patentable per se, but an embodiment built according to the drawing can be patentable."

This is one of the problems with the 2009 Patent Office: according to daily Office practice, a drawing is patentable per se. (Perhaps you meant to say "a drawing should not be patentable per se.")

Patent law 2009: If you can imagine it, you can patent it.

You can get a software patent just by imagining a flow-chart or an end-effect, and drawing or describing it ("what it does", or worse, "what it did"). The examiner will grant you a patent, even if you've never disclosed, and can't even write, a single line of code. The examiner will assume that making and using the invention, i.e., writing the code, is within the level of skill in the art, so it doesn't matter if you can build the embodiment or not. In fact, the examiner will assume that satisfying 112 ("teach how to make and use, possession, best mode) is within the level of skill in the art, if only you will disclose what your imaginary "invention" does.

There are so many "what it does" patents out there... they are unstatutory, in my mind. But unfortunately, since the mid 1990s, if you can imagine it, you can patent it - that's the de facto law of the land. And all you have to do is tell the PTO what you've imagined that a computer can do!

Posted by: NIPRA anonymous at January 7, 2009 10:02 AM