[ijidak]

I had to lol when I saw there was a patent for that.

Divide both operands by 2 was my first idea before loading the page. (I like to try that sometimes before reading the articles.)

I didn't think about the carry bit, but it seems like that would be a logical solution after 5 minutes of extra thinking.

I'm not sure how that's patentable.

That's insane to me.

But maybe there is more too it.

I didn't read the patent itself.

[staticassertion]

https://patents.google.com/patent/US6007232A/en

The patent is for a circuit design to perform that algorithm in a single cycle. The algorithm was never patented, nor could it be.

[not2b]

In practice it makes no difference, because digital logic designers haven't used schematic capture in a very long time. They most commonly write Verilog (or SystemVerilog, which is a superset), and it looks a lot like C:

logic [31:0] a, b, average; assign average = (a >> 1) + (b >> 1) + (a & b & 32'd1);

[WillSlim95]

Sure but any HW engineer before writing HDL will always draw up a circuit before writing the description. And given that the patent was issued in 1996 it was only like 6-7 years after Verilog became an open standard.

[staticassertion]

Sorry, I don't understand what you mean. It makes no difference to whom? It definitely makes a difference to someone writing that code, since that code is not patented.

[not2b]

What I mean is that the circuit diagram is a way to sneak it past a patent examiner, but the software/hardware distinction is completely artificial as both are designed in much the same way.

[tialaramex]

Several of these patent claims are for "an apparatus" because you're not allowed to patent ideas - but any realisation of the algorithm will necessarily be "an apparatus" so the effect is that in fact you can claim algorithms and that's exactly what this is doing.

[staticassertion]

I'm not a lawyer, so I'm confused. If you patent a hardware implementation of a software algorithm, on the basis that the implementation is novel, how would you stop me from writing that algorithm irrespective of how it executes?

[tialaramex]

Where in each claim do you see their proposed "hardware implementation" - the novel device you suppose they've invented - that we might distinguish from, for example, every modern microcomputer ? What I see is a description of the algorithm because of course the intent is to patent the algorithm and not some special hardware implementation nobody will manufacture because it's patented.

If you have some novel arrangement of gears and levers to propose, or some recipe of doping material A with chemical B and then doing so-and-such - that's a classical hardware patent, but the whole point of these "an apparatus" patents is that they aren't trying to describe a novel device anybody invented, they're describing the algorithm the patent is for the algorithm and phrases like "an apparatus" are just a wafer thin excuse for this to be waved through by people whose salary is paid for by the endless flood of such software patents.

As a parallel comment is explaining there hasn't for many years been any substantial difference between what seems to be very much "hardware" and what common folk think of as "software".

The patent application even spells it out, explaining that "A general purpose computer [...] can execute the invention in a single instruction cycle [...]". What "invention" do you suppose is being "executed" here? Is the general purpose computer... making a special circuit? Maybe your laptop is now selling DVD players to Best Buy? No. Their "invention" is just the algorithm, they successfully patented the algorithm.

Nobody would "stop you writing that algorithm", but if they wanted to, and if you've got enough money for it to be worth taking your money, while that patent was valid they could sue you for infringing on their invention and unless you've got deep pockets and proof you invented it first, you are screwed.

[staticassertion]

You don't have to invent a device to file a patent for it. They describe the device, the patent is for the device, the novelty is the construction of the device.

I honestly doubt that they could sue you for infringing on this.