Comment by beejiu | Hacker News Reader

Comment by beejiu | Hacker News Reader

Parent Full thread

Skimmed through this, but couldn't quite understand why the appeal was allowed given all the evidence pointing to the opposite in the judgement.

Is a summary that, because the ANN could be implemented in hardware, it is not necessarily strictly software, and therefore patentable?

Edit: There's a more digestible summary here: https://www.gov.uk/government/publications/examination-of-pa...

> 8. The judgment notes that an ANN may be implemented in hardware, for example a physical box with electronics in it (see [14]), or in software where the ANN exists as a computer emulation, for example where a conventional computer runs a piece of software which enables the computer to emulate the hardware ANN (see [18]).

> 9. The judgment compares hardware and software implementations of an ANN (see [32]-[62]). For hardware implementations, the judgment notes there is no program to which the program exclusion of s.1(2)(c) can apply (see [43]). For software implementations, the judgment holds it is appropriate to look at the emulated ANN as, in substance, operating in the same way as the hardware ANN it emulates. If the hardware ANN is not operating as a program, then neither is the emulated ANN (see [56]).

> 10. Thus, *the judgment concludes an emulated ANN is not a program for a computer* (see [58]). As a matter of construction, the claimed invention is not a computer program at all (see [61]). So, the computer program exclusion is not invoked by the claimed invention ([61]).

Isn't that a bit silly...? Almost any software can potentially be implemented in hardware and be "emulated".

The Comptroller-General of Patents has an awesome job title and, imho, was right on this (and the judge is wrong).

It's very strange. It's also not clear to me why you couldn't construct that any software is a hardware emulation, and therefore use that as a claim of prior art. Seems to be a fundamental contradiction.

I'm not sure how the UK avoided software patents before now based on this argument though. As in:

- literally implement ANN in hardware

- patent it

- someone emulates your patented design in software (or even hardware of some parts slightly abstracted to make "emulation" of your patent fast).

What happens now? Either all emulatable hardware patents are useless, or there's a backdoor into software patents (synthetically put it in hardware first).

Software runs on hardware, specifically, computers. There's no non-hardware way to run software. The court is seriously confused and shouldn't be analyzing anything more complicated than a toaster.

Don't rely on logic in matters involving patents.

I'd rather say, don't rely on logic in matters involving judges.

I will patent this.

I will patent that. The judges being bereft of both intelligence and abundantly full of property and pomposity, I do ordane that both parties and judiciary are far outside of the f*king boat in which they are engaged in rowing.

Turn off the machine and show me the circuit? It is not there. It should be copyrightable... But since created by machine, it should not be. Scrooge should come early for them. I shall argue on.

The only case I could potentially see software not being hardware implementable is when it is self-mutating, but even in this case ... FPGAs?

I can't follow the train of thought.

Humans are implemented in hardware. Any "software" algorithm that could in theory be executed manually by a human can be implemented in hardware.

The only kind of software I can imagine that (probably) is not implementable in hardware are uncomputable algorithms such as a halting problem solver.

Good catch, also in the same vein, maybe something related to busy-beaver numbers / computations?

Yes it's ridiculous. Literally all working software can be implemented in hardware. In fact all working software is implemented in hardware. There's no other way to run software than in hardware.

And God made pencil testing for a reason; to sell pencils.

Much like the EU patent convention that also excludes computer programs as such, programs for generic computers routinely get patented. It seems it is the opinion of esteemed patent judges that "as such" means it only applies to programs that do nothing other than exist as computer programs, in some pure, Platonic sense. If they actually do anything, they are no longer "as such" (but if they do nothing, that is not a patentable invention anyway, software or not).

In other words, it is the learned and wise decision of the honored judiciary that the entire sentence excluding computer programs is there just for decoration. Like flavor text on a Magic card. As a layperson I naturally defer to their expert judgement.

> For hardware implementations, the judgment notes there is no program to which the program exclusion

Puzzling argument. This logic can apply to every single computer program (i.e. it can be implemented in hardware), which means the exclusion will never apply, which means the judge's reasoning is wrong.

But he has a "Sir" in his title. That is guaranteed to ensure maximum accuracy and flawless logical thought.

I've no problem burning some karma on this one, the "BUSINESS AND PROPERTY COURTS OF ENGLAND AND WALES" are basically a shop to "buy a judgement" from somewhere sounding authentic. Absolute scam.

To whoever "won" the appeal, well done, you have a software patent, "good luck"; in his opening article he even points out to you in his opening statement this can't be applied to program for a computer. Money well spent eh.

You are right, judges can not be wrong, but they can misdirect themselves.

Someone should tell them that any program for a universal Turing machine could be implemented as an ordinary Turing machine. That any computer program could be implemented as an ASIC.

Computer chips are designed from software Programs.

It used to be a language called verilog. (According to wikipedia based on ADA, pascal,C and Fortran)

https://en.m.wikipedia.org/wiki/Verilog

NAND and NOR gates are sufficient for any computation, so you can invalidate any software saying it's form of emulation of hardware?

Section 12 seems to contain the real reason why this is a technical contribution:

> The judgment accepts an argument that moving data outside the computer system, in the form of a file that is transferred, provides an external (outside world) technical effect.

> When coupled with the purpose and method of selecting the file’s contents, this fulfils the requirement for a technical effect which avoids the computer program exclusion.

Sounds like these sorts of "decision matrix"-type ANNs are going to now be patentable if there is some kind of application of "special sauce" that can be used to select files or some other external component.

That's not really special sauce though. Using data from an external sauce is foundational computing. :(

Only impure software uses I/O. This judge is clearly versed in the highest arts.

Thanks for the link to the Statutory Guidance, which provides some very helpful commentary, and I feel that points 12 and 13 are of some interest:

12. The judgment accepts an argument that moving data outside the computer system, in the form of a file that is transferred, provides an external (outside world) technical effect (see [73]-[74]). When coupled with the purpose and method of selecting the file’s contents, this fulfills the requirement for a technical effect which avoids the computer program exclusion (see [76]).

13. The judgment further holds that a trained hardware ANN is capable of being an external technical effect which prevents the computer program exclusion from applying (see [78]). The judgment holds there ought to be no difference between a hardware ANN and an emulated ANN in this respect.

If I am not mistaken (though there is a good chance I am), the judge could have made a narrow judgement in favor of the plaintiff on the basis of point 12 alone. Inventions that use a suitably-programmed computer in a process that produces an "external technical effect" are not prima facie ruled out (though they do have to satisfy a requirement of sufficiency of description, which, as the judge noted in his opening remarks, is not being ruled on here, together with (I think) non-obvious novelty and utility)

Point 13 may have much broader implications, as it asserts that any trained software ANN is, in itself, an "external technical effect", and presumably patentable if it satisfies other requirements such as sufficiency.

The first issue with this is, as you and others have pointed out, that the justification given for this claim (the possibility of a functionally-equivalent implementation in hardware) works equally well for any computer program - ANN programs are not different in this regard. If followed to its logical conclusion, this principle would seem to completely nullify the "program for a computer" exclusion.

A second point that might be made here is that the difference between a trained ANN (whether implemented in software or hardware), and itself in its untrained state, are data - specifically, the weights. This might imply that the intellectual property issue here is a matter of copyright, not invention.

Finally, and perhaps most broadly, if a trained ANN is, in itself, an "external technical effect", does that open the door to the patenting of ANN training methods? (Maybe they already are?)

So if I implement a software version of the ANN with one small change, my software version is not an emulation of the hardware ANN, so the patent does not apply?

the doctrine is that only things that are 'novel' and 'non-obvious' (to a 'person having ordinary skill in the art') can be patented, and any accessible published material (among other things) is considered 'prior art' for determining this. So unless your 'one small change' fundamentally alters the behavior of the 'black box' in a novel way, one would probably say that it is not novel and pretty obvious for a person having ordinary skill in the art to emulate a slightly modified version the hardware in software. of course, vector search on music also seems pretty obvious to me. I don't actually know how this plays into infringement specifically rather than trying to patent something though.

> For hardware implementations, the judgment notes there is no program to which the program exclusion of s.1(2)(c) can apply

This is complete nonsense, but not surprising because the judge probably has no idea what a program is. In France, a few years ago, when discussing the matter of Web cookies, some authority described cookies as “little programs” ¯\_(ツ)_/¯

Reply on news.ycombinator.com