Your example of the pizza store address data is a good one because it is midway between the music-on-CD example and computer-behavior (program) on-CD example.

Using the word “patentable” is tempting, but you should avoid it.
The real questions are:
1) is it “patent eligible” under section 101?
2) is it novel and nonobvious under sections 102 and 103 respectively?

The critical battle line as between inventor and society (represented by the US patent office) has historically always been drawn at the 103 obviousness question

The Bilski case (see topic of this IPWd post) can be seen as an attempt by anti-patent forces to establish a new, stop-the-patent beachhead line at the 101 line (think of potential stoppage lines, 101, 102, 103 being established in that order)

One might try to believe the anti-patent forces won because the US Supreme court ruled that Bilski’s invention is “too abstract”.

However, the US patent office is not a supreme court, it is an “administrative agency” and thus bound by administrative agency law (see 5 USC 706). Under administrative agency law, an agency cannot engage in arbitrary and capricious decision making. Merely saying, “this is too abstract” is arbitrary and capricious and thus in violation of administrative agency law. So the US patent office cannot do it even though the US supreme court can for the *one* case of Bilski v. Kappos

A hard disk is a machine component inside a computing machine just like a carburetor is a machine component inside a car.

A programmed disk is different from an unprogrammed disk.
A programmed disk bearing a computer program that makes the computer operate faster (or more efficiently, etc.) is an improved machine component and therefore “patent eligible”.

Here’s my issue with this argument: the ‘programmed disk’ (or CD, or FPGA) isn’t different in performance from the ‘unprogrammed disk’ as measured by the things you’d measure hard disk performance (rpm, read/write latency and throughput, MTBF, etc). The improvement came in the performance of the computer system with respect to some application, not to the computer system itself (which is unquestionably a machine).

Let’s say that the machine was responsible for arranging pizza delivery. Let’s say we have two instances of such machines, PMA which delivers pizza in 40 minutes, and PMB which delivers pizza in 30 minutes. A change is made to PMA (now PMA’) whereby it now delivers pizza in 20 minutes, a doubling of its former speed, and 33% faster than PMB, a significant innovation. Now let’s say that PMA (and PMA’) is my house, and the change that was made was to the phone number of the pizza shop, one that was 20 minutes closer to my house. The physical change that was made was to update the hard disk containing my list of phone numbers.

The addition that I made to your existing example was to view the whole house as the system to be measured, rather than the computer system in which the hard disk is installed. I think this remains within the spirit of the discussion, as the software changes don’t actually change the physical properties of the hard disk (rpm, magnetic technology, etc) or of the CPU (instruction set, cycles per second, etc) or any other physical element of the computer system, but only the effects of running some application in terms of that application’s performance. The application just happens to be pizza delivery rather than, say, web search.

I recognize that this wouldn’t be patentable because it’s obvious (at least), but you’re essentially arguing that there are household performance changes (‘business processes’) that are patentable based on what is written on my hard disk. You argue that this is true because the hard disk changed in the sense that some new, more useful bit pattern was written there. (Do I misunderstand?)

Let’s back track.

The USA patent law section 101 says that an “improved machine” is patent eligible subject matter.

More to the point:

Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.

A hard disk is a machine component inside a computing machine just like a carburetor is a machine component inside a car.

A programmed disk is different from an unprogrammed disk.
A programmed disk bearing a computer program that makes the computer operate faster (or more efficiently, etc.) is an improved machine component and therefore “patent eligible”.

This is no different than an improved carburetor that improves performance of car. It too is “patent eligible”.

We are discussing where the boundary between patentable and non-patentable is…

No we are not doing all of that. I was just trying to give examples that flesh out the different issues.
Perhaps this is where all the trouble starts, use of imprecise language.

There is one section in USA patent law, 35 USC 101, which deals with what types of subject matter are “patent eligible” (this being the coined term used by the Bilski Supreme Court, whereas beforehand we legal doers called it “statutory subject matter”).

There is another set of sections in USA patent law, 35 USC 102, 103, which deal with how novel and nonobvious the otherwise “patent eligible” thing/process is.

There are yet other reasons why a thing/process may be blocked from getting a patent. That would just turn the muddy waters deep dark. No need to go there.

The “Bilski” decision; which is the focus of the above IP WatchDog post is entirely limited to the 35 USC 101 question.

THERE WAS NO SOFTWARE mentioned in the Bilski patent application and no computer mentioned.

Those who are suggesting otherwise to you are misleading you –flat out.

You can have trade secrets, and patent rights, and copyrights and trademark rights and contract rights and more all out of one, apparently simple business venture.

Certainly. I would assume this carries down to the venture’s intellectual property as well. A narrower statement would be to characterize how many overlapping sets of rights apply to a single song or a ‘single article of manufacture’ (which I take by usage to be a term of art).

you do not at all remove the question of whether you can have a patent for an article of manufacture (i.e. a CD) whose only difference over the prior art is that it encodes a new rock ‘n roll song.

That’s good, because I think this is close to the question I am trying to isolate. Let’s take the four obvious cases:
1 – old media (article of manufacture), old song
2 – old media (article of manufacture), new song
3 – new media (article of manufacture), old song
4 – new media (article of manufacture), new song

You’ve highlighted cases 4 and 2 as examples of something potentially patentable. What they have in common is the new song, though you (also) point strongly toward the article of manufacture as being important My (perhaps still mis)understanding suggests to me that copyright would apply to the new song in cases 2 and 4, and that it is the new article that makes 4 (and 3) potentially patentable.

The other thing you are misunderstanding, and muddying the waters with, is the separation between “software” (the physical world kind) and mathematical abstractions.

Well, that’s the essential question, I think. There are very smart people on both sides of it. You’ve made up your mind about it, and I’m trying to get at your reasoning. Thanks for taking the time to elaborate. As I said, I would like to understand.

In its day, and since, N-R did speed up the machines it was used on (or built into),. I deliberately chose an example set in murky waters, to see if we could find it with our reasoning, and clarify the situation. We are discussing where the boundary between patentable and non-patentable is… and, hopefully, my original question of your claim that there is ‘no clear line of division between “software” and “hardware”.’

or am I misunderstanding something here?

Yes, you are misunderstanding something here.

In the domain of law, you can simultaneously acquire all sorts of “legal rights” in what appears to a lay person to be a “one thing”.

You can have trade secrets, and patent rights, and copyrights and trademark rights and contract rights and more all out of one, apparently simple business venture.

So by wishfully pointing at copyright law, you do not at all remove the question of whether you can have a patent for an article of manufacture (i.e. a CD) whose only difference over the prior art is that it encodes a new rock ‘n roll song. The same logic applies to an article of manufacture (i.e. a CD) whose only difference over the prior art is that it encodes a new and nonobvious machine process that a computing machine is going to perform.

The other thing you are misunderstanding, and muddying the waters with, is the separation between “software” (the physical world kind) and mathematical abstractions.

If the only (emphasis on the word “only”) inventive contribution you have is that of a purely mathematical algorithm (I’ll take it your N-R method is a numerical analysis method, yes?), then you cannot get a patent for that taken-alone act.

However, if you come up with a computer process that greatly speeds up the time needed by the computer to produce solutions (using N-R or whatever), then what you have is a “faster” machine. By law, such an “improved” machine is patentable. See again, US code section “35 USC 101″ (google it).

> Instead consider music,

I’m no lawyer, but it seems that music and its performances are well-covered under copyright law. It is my understanding that the copyright holders of ‘Happy Birthday’ would collect their royalties whether the medium were a phonograph record (33 or 45), an 8-track player, a CD, an MP3, or whatever the next medium is.
Unless they’re considering music patents, I’d prefer we avoid using examples from music as it muddies the water to my eye. Is that fair, or am I misunderstanding something here? (And don’t make me ask why don’t we just copyright software!)

So, rather than ask ‘What if Newton and Rhapson were in a rock band’, I’d rather return to ‘say Newton and Rhapson worked for IBM today. Would they have a case for patenting Newton-Rhapson?’
By the way, Newton-Rhapson is one of the earliest (but still used and taught) methods for finding the zeroes of a function. It was non-obvious at the time, it qualifies as an algorithm, and it could (and has been) programmed as software, in the last few decades usually late at night before the homework is due by junior engineering students in their numerical methods course.

I am really trying to understand how the law works, and how software patent defenders think

@pmorrison,

You’ve thrown in a whole ton of new stuff into the kitchen sink and, unfortunately, my insinkorator (waste disposal unit) hasn’t the time to crunch through all of it.

Please let me jump to a chase point that you did not yet raise -although you are converging on it.

“Software”, what is this thing? If you say it’s just recorded code which is recorded on a tangible medium of expression, well then of course your equations solution using Newton-Rhapson (whatever that is) written on a piece of paper would qualify under that definition. Pretty much “everything” could be called “software” under that broad definition; all books, all statues (of the Michelangelo type); in fact pretty much every physical object which by existential definition, is an “expression” of itself (the inverse of the “this is not a pipe” existential concept).

but let us not go that far out on the limb.
Instead consider music, let’s say Jefferson Airplane rock ‘n roll, recorded on a CD (compact disc) as digital code.
Is this software?
Well yes, in a sense it is. Code is code.

Let’s say a new rock n’ roll band records its new and nonobvious music on a CD.
Should *that* err (cough cough) “software” be patentable?
You might be surprised to learn that even we pro-patent fanatics will say that generally, no, the new music CD –even though it is a physical “article of manufacture” as we legal beagles refer to it, should not be patentable. But not for the reason you would hope for. Rather, because it is obvious to record music, even novel music, on a CD.

By the same token, it is obvious to record new software on a CD. However, this recorded-upon-CD (another physical “article of manufacture”) is a functional part of a functional machine (of the computer whose machine behavior is being transformed by the recorded-upon-CD).

*That* is something different.

Do we have language tools to cogently express that difference?
Well not really. Our language is not well enough developed to express many a strange concept. However, our failure to be able to express the difference between a CD containing new music and a CD containing new, nonobvious and useful “software” does not make the latter not-patent-worthy.

My time has run out for now.

I hope that gave you a small taste of how the “law” operates and what the underlying issues are.

Good discussion.