Over at IPCentral, Jim DeLong quotes a lengthy critique of the SFLC brief in the Microsoft v. AT&T case. The critique was written by one Greg Aharonian. A lot of it is the kind of legal inside baseball that I’m not really qualified to comment on, but there’s one theme that runs throughout the critique that’s just flatly wrong:
The first lie of Moglen’s brief is a big lie of omission. Nowhere in his brief does there appear the word “hardware”. It is unethical to talk about the patentability of software without simultaneously talking about the patentability of hardware, especially in light of hardware/software codesigns tools. And even using the word “hardware” is pointless unless you provide rigorous definitions of “hardware” and “software”. Moglen doesn’t. So when Moglen bases his software patent hatred on Benson: “The holding of Benson is properly applicable to all software, because a computer program, no matter what its function, is nothing more or less than the representation of an algorithm.” as well, he is arguing hardware patent hatred: “The holding of Benson is properly applicable to all hardware, because a digital circuit, no matter what its function, is nothing more or less than the representation of an [Boolean] algorithm.”
This is silly. I would be very interested to see the boolean algorithm that is equivalent to, say, an LCD panel. Some characteristics of hardware can be described as equivalent to software algorithms, but other aspects (such as, say, the ability to display information to the user) cannot.
Indeed, hardware has all sorts of intrinsic characteristics that software does not–how much energy does it consume? How hot does it get during operation? What materials is it made of? How big is it? Hardware engineers need to worry about these kinds of questions, and none of these questions can be reduced to a mathematical algorithm. In contrast, everything about software can be reduced to mathematics. It doesn’t make any sense to ask about the operating temperature, energy consumption, or color of string of 1s and 0s.
Now, it’s true that some aspects of digital hardware can be described as being equivalent to software algorithms. And it’s quite likely that in a world without software patents, the algorithmic aspects of hardware devices would not be patentable. But that’s a far cry from saying that hardware wouldn’t be patentable at all.
For example, suppose came up with a new manufacturing technique, such as silicon on insulator,that allowed them to produce chips that operated at a higher clock rate or lower power. I don’t see how you could possibly interpret Moglen’s argument to suggest that such an innovation wouldn’t be patentable. Silicon-on-insulator technology doesn’t have a precise mathematical description the way software does. Of course, you could build a software simulation of a SoI chip that would capture some of its characteristics, but the simulation would necessarily be imprecise. So there’s no reasonable sense in which SoI technology is equivalent to a mathematical algorithm in the same way that the latest version of Microsoft Office is.
Aharonian repeatedly insinuates that (among other things) Moglen is scientifically illiterate and has failed to do his homework. In my experience, when people deploy that level of bluster, it’s a sign that they’re overcompensating for the lack of strong arguments. People who know what they’re talking about and know they’re in the right don’t need to resort to name-calling: the facts speak for themselves.
The Technology Liberation Front is the tech policy blog dedicated to keeping politicians' hands off the 'net and everything else related to technology.