Stockfish: Our lawsuit against ChessBase

[jkominek]

I understood that the U.S. Patent Office doesn't do much more than just file the patents (after checking that it is not for a perpetual-motion device). Whether they might be invalid because of prior art, is up to the courts to decide, when someone decides to challenge the patent.

The USPTO does examine novelty and non-obviousness. In recent times they have become stricter and the quality of their searches has improved too. But it is true that the grant of a patent does not mean it will survive in court. I believe the stricter line started with the KSR v. Teleflex SCOTUS opinion (which rejected the mechanical teaching-suggestion-motivation test for obviousness as applied by the USPTO and the lower courts, which basically ignored the role of common sense).

Actually, since 2012 the most promising way to attack a granted US patent is by requesting inter partes review at the USPTO (which are dealt with by the Patent Trial and Appeal Board of the USPTO). The PTAB uses the "broadest reasonable interpretation" of a claim when examining for novelty and non-obviousness, whereas the federal courts tend to use a more narrow interpretation (which renders the claim more likely to be upheld but less likely to be infringed). A former (rather patent maximalist) chief judge of the CAFC has referred to the PTABs as death squads.

I happen to have two good friends knowledgeable on this topic, one of whom is a patent attorney and the other is a patent examiner, and while I don't know much myself I can confirm, second hand, that KSR was a pivotal ruling in US patent law. It provided the examiner with a greater toolkit, if you will, for asserting the obviousness of an application in terms of logical combination of previous inventions (prior art), resulting in issuing a "103 rejection". If the attorney argues in response, the examiner may cite KSR v. Teleflex in support of their position, which the attorney knows would likely be upheld in the courts. The normal continuation is to amend one or more of the claims.

It is unusual for a patent application to be a one shot ordeal. Almost all undergo several rounds of back-and-forth. The overall acceptance rate from start to finish is somewhere in the range of 40 to 60%.

[dkappe]

89% sounds high to me. Maybe it is 89% of the originally filed claims?
Indeed it is very common for claims to be dropped or further restricted during prosecution, which for a third party can be just as good as the application being refused.

You’re right. It’s 88.9%.

[syzygy]

I don't remember anyone ever making the argument that, while the box that houses the computer contains artistic expression, the computer itself doesn't, therefore other computer manufacturers could simply make an exact copy of IBM PC computers and just put them in a different box.

If there are no patents in the way, they very certainly could. This is the point of patents.
But in practice different manufacturers will source components from different suppliers and will look for ways to cut costs, etc. Also they may want to differentiate their products from the competition.

And obviously there were patents in the way. IBM was and still is record holder patenting.

https://www.engadget.com/2015-02-06-mar ... oneer.html

Dr. Mark Dean, an African-American computer scientist and engineer, spent over 30 years at IBM pursuing the Next Big Thing. He was chief engineer of the 12-person team that designed the original IBM PC in the early '80s, earning him three of the nine original patents for that device.

Sure enough:
US4528626 Microcomputer system with bus control means for peripheral processing devices
US4575826 Refresh generator system for a dynamic memory
US4598356 Data processing system including a main processor and a co-processor and co-processor error handling logic

So the manufacturers of IBM-PC compatibles had to design around these three and the other six patents (or license IBM's technology).

[towforce]

You shouldn't be able to get a patent for a harder/sharper/cheaper nail

So you are advocating that company B should be free to copy company A's invention of a harder/sharper/cheaper nail. Understood.

That's a straw man argument: patents should be for useful ideas that aren't obvious, which doesn't apply to making nails harder, sharper or cheaper.

THE FACT: companies very rarely make exact copies of each others' products - even when there's no "human artistic expression" involved.

MY OPINION: companies should not be allowed to make exact copies of other companies' products.

If it is true that an NN (or its weights file) do not have protection, then the remedy would be to build an application that converts NNs into methods/functions that would be incorporated into the software, which does have protection. It would be better for everyone*, though, to simply protect NNs from being copied.

*except for people who want to be allowed to copy NNs

Once again, though: the fact that manufacturers don't generally make exact copies of other manufacturers' machine parts, even though it would be both cheaper and faster to market to do so, implies to me that between ourselves we're missing something here.

[smatovic]

[...]
THE FACT: companies very rarely make exact copies of each others' products - even when there's no "human artistic expression" involved.
[...]

???

https://en.wikipedia.org/wiki/Schuko

https://en.wikipedia.org/wiki/Edison_screw

CTRL+F patent

we're missing something here.

Do you actually follow those links given in this thread?

--
Srdja

[towforce]

[...]
THE FACT: companies very rarely make exact copies of each others' products - even when there's no "human artistic expression" involved.
[...]

???

https://en.wikipedia.org/wiki/Schuko

https://en.wikipedia.org/wiki/Edison_screw

CTRL+F patent

we're missing something here.

Do you actually follow those links given in this thread?

--
Srdja

It's not clear to me what point you're making, but join me in examining these two products (Schuko plug and Edison screw):

* they began as patented products

* they later become widely used standards

* you haven't shown me an example of a company making an exact copy of another company's products. No doubt multiple companies make products that work to the same specification - but in each case, they will have worked out exactly what to build for themselves. Therefore, there will be detail differences between products from different manufacturers (if you find two products that seem exactly the same but by different manufacturers, it's likely that they were made on the same production line but had different names printed on to them: this is common).

[smatovic]

[...]
* you haven't shown me an example of a company making an exact copy of another company's products. No doubt multiple companies make products that work to the same specification - but in each case, they will have worked out exactly what to build for themselves. Therefore, there will be detail differences between products from different manufacturers (if you find two products that seem exactly the same but by different manufacturers, it's likely that they were made on the same production line but had different names printed on to them: this is common).

Seriously? Now you are nitpicking about the wording "exact copy"? On a specification level? On an fine mechanics precision level? On atomic level? If I file a patent and give a license to an US manufacturer and one to a Chinese manufacturer, guess what, it will be the same product, different quality maybe, but as you say, not an "exact copy". When you buy a CPU with a specific brand name, guess what, there are different steppings, the first charge is different from the second, not an "exact copy", if I compile SF with clang and GCC, guess what, not an "exact copy", and so on and so on and so on

***edit***

and, as analogy for the digital space this does not hold, cos contrary to the material world you CAN make "exact copies" in the digital space, at least on bit-level.

--
Srdja

[syzygy]

* you haven't shown me an example of a company making an exact copy of another company's products. No doubt multiple companies make products that work to the same specification - but in each case, they will have worked out exactly what to build for themselves. Therefore, there will be detail differences between products from different manufacturers (if you find two products that seem exactly the same but by different manufacturers, it's likely that they were made on the same production line but had different names printed on to them: this is common).

Are you serious?

Did you forget what you were trying to argue before? You are simply incoherent.

Does admitting that you got it wrong hurt you so much? How do you think you will ever improve as a human being?

[towforce]

OK - everyone who doesn't understand "exact copy", please read carefully and completely understand if you can.

* Facebook and Google both have NNs that recognise objects in images (for simplicity, let's say that they have one each)

* the NNs both have exactly the same functionality - so if a patent existed for an NN that recognised objects in images, at least one of them would be in breach of it

* although both companies have an NN that recognises objects in images, the two NNs were built independently, and they are not exact copies of each other

* if a third company ("Thievery Ltd"), makes a copy of Google's NN, and includes it in an app that they sell to customers in the Play Store, then this will be an exact copy

OK - now suppose that "Clocks UK" manufacture a mechanical clock motor, which will not be seen by customers, contains no "human artistic expression", and will be sold to other manufacturers who will put them in artistic cases and sell them:

* the basic machine is well known (coiled power spring, winding mechanism, cogs to get the fingers to rotate at the correct relative speed, bouncy spring for timing, escapement wheel, and catch mechanism which lets the escape wheel move a little, then catches it (the part that makes the clock tick)

* company B see Clocks UK's product, and decide that they'd like to be in that market as well. They have a choice:


Exact Copy

Tool up to make exact copies of every component in Clocks UK's product. This will be quick and cheap, and will greatly enhance their business outcomes.


Design And Build Their Own Components

The components will have different dimensions, different thicknesses, different shaped cog teeth (especially in the escapement catch mechanism), and the components will be made from different materials.

The results of this choice will be:

* different product with the same functionality

* business outcomes won't be as good: development and testing will have cost more, and the time to market will have been slower

Anything still not clear?

[syzygy]

Anything still not clear?

I have no idea what point you are now trying to make.

I have no friendly words for your behaviour.