LWN: Comments on "US Supreme Court rules against software patents"

Gene, and the case generally

mathstuf — Mon, 30 Jun 2014 02:06:05 +0000

> Changing laws by court case is the long, slow method. The only way to get anywhere faster is to involve the (infamously dysfunctional US) legislature.

You might get the Congresscritters to *do* something, but it might have so many riders and loopholes so as to be ineffective. It could also turn on you and explicitly make the bad stuff legal (or the good stuff illegal) depending on lobbying. The courts may be slow, but I think I'd trust them more to get a better overall, long-term result.

Gene, and the case generally

kleptog — Sun, 29 Jun 2014 20:35:30 +0000

Sounds like the smart thing to do: rule on the cases before them and don't try to make decisions about issues for which no evidence has been provided.

Changing laws by court case is the long, slow method. The only way to get anywhere faster is to involve the (infamously dysfunctional US) legislature.

In the mean time, are there any running cases involving compression algorithms? Because they would get to the heart of the issue.

don't get too excited

Wol — Thu, 26 Jun 2014 10:47:05 +0000

Can I add drugs for Parkinsons to the list of "nothing new in years"? But there's plenty of new "variations on a theme".

The only real *new* drug is Levodopa, which was discovered in the 1960s. To the best of my knowledge ALL of the long-term drugs for Parkinsons are variants on it. But Levodopa is now deprecated (because it's out of patent ...)

And on this particular example I have first hand experience :-(

Cheers,
Wol

don't get too excited

roc — Thu, 26 Jun 2014 10:30:24 +0000

> The most notable example are anti-blood-clotting drugs - there are no new
> drugs after tens of billions of dollars spent on their development.

Hmm. What about rivaroxaban? That's been getting regulatory approval within the last five years; do you not classify that as "new"?

don't get too excited

Cyberax — Wed, 25 Jun 2014 22:25:35 +0000

> Actually, it isn't. Or rather, the problem is that the protection granted by current patents is pretty poor - if your patent lasts 17 years from publication, and FDA trials take 15 years, then your protected period is 2 years ...
It's close to 7-10 years. Typical time from the initial drug development to marketing is 10 years. Significantly extending the protection period won't generate much additional income, but it will destroy the generics market.

> Not off the top of my head. But I have a strong personal interest in this field, and all the stuff I've seen says that the majority of Pharma R&D is minor variations on old drugs.
Totally and insanely incorrect. Most of R&D is done on _totally_ _new_ areas. Tweaking old drugs is simply not profitable for large-scale R&D, because old drugs already work.

Sometimes companies foray into areas covered with existing drugs and most of the time it ends badly. The most notable example are anti-blood-clotting drugs - there are no new drugs after tens of billions of dollars spent on their development.

>Oh - and I get the impression you are in Eastern Europe?
Our company is distributed between continents, but our customers are almost all in the US.

>The proportion of R&D spent on blue-sky research is much higher in Europe that it is in America, as far as I can make out ...
Nope. Pharma companies trail only silicon companies in the amount of R&D - universally. Simply because pharma companies have to innovate or die: http://pipeline.corante.com/archives/2009/07/08/how_much_...

don't get too excited

Wol — Wed, 25 Jun 2014 22:07:05 +0000

>> then they get exclusive rights for say 10 years after the drug gets FDA (or equivalent) approval.
> That is called 'a patent'.

Actually, it isn't. Or rather, the problem is that the protection granted by current patents is pretty poor - if your patent lasts 17 years from publication, and FDA trials take 15 years, then your protected period is 2 years ...

> Can you provide actual examples of pharma research? I do, since we work with drug discovery companies.

Not off the top of my head. But I have a strong personal interest in this field, and all the stuff I've seen says that the majority of Pharma R&D is minor variations on old drugs. Oh - and I get the impression you are in Eastern Europe? The proportion of R&D spent on blue-sky research is much higher in Europe that it is in America, as far as I can make out ...

And if you work with drug discovery companies, does that mean you're being fed a filtered view? Or do you personally actively pick up on anything you come across in the specialist or general media?

One of my favourite examples - aspirin. Which, because it's LONG out of patent, has only recently made a comeback as a drug for treating heart attacks, despite its proven ability to save lives and aid recovery. If your first heart attack kills you, it's not much use, true. But many (30-40%) of heart attacks follow a double-whammy pattern, a minor initial attack and a fatal big one several hours later. Take aspirin after the first attack and the second never happens. And your chances of full recovery from the first attack are much enhanced.

And, following the antics of NICE (the UK drug prescribing authority), it certainly seems as though there is an awful lot of anti-generic pressure in the industry as a whole. NICE has a habit of pushing for generics (not surprisingly) and there is a massive pushback. That said, my wife is chronically ill, and we make a point of requesting specific drugs - that happen to be patented. Not because we believe patented to be better, but generics are subtly different and have been widely reported as causing havoc with the drugs regime. It's change we don't want, not being pro/anti generics. The big problem with generics is that every time you get a new supply of drugs, the specific version can change with all the subtle differences that implies :-(

Cheers,
Wol

don't get too excited

khim — Wed, 25 Jun 2014 20:24:55 +0000

I still change a + -b to a - b in my head.

LOL. This is exactly how negative numbers are usually “made” (Remember? “God made natural numbers; all else is the work of man”). You claim that “integer” is pair of natural numbers, then you can easily explain how all operations work (each iteger could be represented by many equivalent pairs, but this is not a problem: the same problem exist with rationals and it does not hurt us, right?). In particular −{a, b} is the same as {b, a}. Easy and simple…

don't get too excited

dashesy — Wed, 25 Jun 2014 18:18:26 +0000

Interesting, I remember problem with negative numbers too (subtraction was fine). Our teacher started the numbers by associating them with apples, I could easily grasp rational and real numbers (which were thought later than negative numbers), but I could not imagine having -3 apples. I still change a + -b to a - b in my head.

don't get too excited

mathstuf — Wed, 25 Jun 2014 14:32:58 +0000

Well, except for the drugs for rare diseases, there are usually off-brand versions which are a fraction of the price of the name-brand drug (and the same active ingredients). Be sure to ask if there's an off-brand version of things from your doctor when getting prescriptions.

don't get too excited

mpr22 — Wed, 25 Jun 2014 14:29:06 +0000

I believe the US-style drugs advertising setup is about getting someone who's on your competitor's drug (or a generic drug) to pester their doctor for a prescription for your New! Innovative! Shiny! (still in patent) drug.

Some other countries (my own among them) prohibit pharmaceutical companies from advertising their prescription-only drugs to the general public.

don't get too excited

mathstuf — Wed, 25 Jun 2014 12:19:18 +0000

Do you also have numbers for how much is spent on advertising? I remember how many ads there were for drugs years ago and I can only assume it has gotten worse. I guess the target audience is hypochondriacs since if the things they talk about are half as bad as they make them seem, why would have not already seen a doctor?

don't get too excited

Cyberax — Tue, 24 Jun 2014 21:26:53 +0000

> Stop spending millions on developing minor variations on drugs that are about to go out of patent?
This drug is completely new. It's also for a disease that is a major cause of liver transplants.

Yet it still took more than 5 billion to develop this drug. It's actually considered cheap these days for new drugs. Lots of companies spend more money without any results.

> Do a *real* risk analysis and allow dangerous drugs for serious illnesses to be trialled much more quickly and with far fewer self-defeating safeguards - if I stand a 10% chance of dying from the illness, and a 1% chance of dying from the drug, guess which one I'm going to choose!
Such regulations are already in place. Critical drugs can be fast-tracked by FDA and approved even if significant side-effects are present.

It still doesn't help a lot. Clinical tests are not just a formality to get an approval - a lot of drugs (50% is a commonly cited number) fail during the tests.

> And lastly, maybe allow companies to carve out research niches as "protected areas" - if a company does "blue-sky" research into a drug family, then they get exclusive rights for say 10 years after the drug gets FDA (or equivalent) approval.
That is called 'a patent'.

> But at the end of the day, while your figures may be impressive, you still haven't provided any evidence that that research is value for money. All the evidence says that, even measured in $Billions, most pharma research isn't!
Can you provide actual examples of pharma research? I do, since we work with drug discovery companies.

The amount of money spent on 'useless' research is pretty much zero. Almost all of the R&D is spent on serious diseases. And pharma companies actually spend more on R&D than just about any other industry.

don't get too excited

Wol — Tue, 24 Jun 2014 21:13:55 +0000

> So what are your suggestions?

Stop spending millions on developing minor variations on drugs that are about to go out of patent?

Actually STARTING to spend some money on sorely needed drugs for illnesses that don't affect westerners?

Or reducing the ability of the food manufacturers to sucker/advertise us into eating unhealthy diets (it's looking more and more certain that the low fat industry has got cause and effect the wrong way round - the diet industry is the *cause*, and obesity is the *effect*)?

And you said that things are only getting worse - maybe we should try and reduce this obsession (driven by the American legal industry) that everything "must be safe". LIFE ISN'T SAFE!!!

Do a *real* risk analysis and allow dangerous drugs for serious illnesses to be trialled much more quickly and with far fewer self-defeating safeguards - if I stand a 10% chance of dying from the illness, and a 1% chance of dying from the drug, guess which one I'm going to choose!

And lastly, maybe allow companies to carve out research niches as "protected areas" - if a company does "blue-sky" research into a drug family, then they get exclusive rights for say 10 years after the drug gets FDA (or equivalent) approval. More to the point, other companies aren't allowed to copy-cat research into similar drugs - they have to carve out their own niche.

But at the end of the day, while your figures may be impressive, you still haven't provided any evidence that that research is value for money. All the evidence says that, even measured in $Billions, most pharma research isn't! Okay, sofosbir may be a completely new drug (which would be protected under my "carve out a niche" rules), but equally (the wikipedia article gives no clues) it could just be a minor variant on an existing drug.

Cheers,
Wol

don't get too excited

Asebe8zu — Tue, 24 Jun 2014 21:03:12 +0000

Did some searching and found that these ideas originate from Jean Piaget who studied this extensively. He found that the ability to perform formal operations as he called it develops in the early teens. However, it does not necessarily develop at all. Some later studies indeed found this ability in only a fraction of high school seniors.
See http://www.edpsycinteractive.org/topics/cognition/piaget....

Cheers

don't get too excited

Asebe8zu — Tue, 24 Jun 2014 20:44:59 +0000

Thank you for the explanation. I will try to remember this when I explain things to my son.

don't get too excited

Cyberax — Tue, 24 Jun 2014 19:31:02 +0000

We work from folks in drug discovery.

Here's an example. This drug ( http://en.wikipedia.org/wiki/Sofosbuvir ) costs about $150 to manufacture. Yet its retail price is about $80000.

You'd think the markup on each bottle of pills is $79850? Wrong, it's only about $20000 - about 30% of price. The rest goes towards financing the expenses incurred during the drug development.

The problem is, it took around 5 billions of dollars to discover and test this drug. And it's only getting worse with new drugs.

So what are your suggestions?

don't get too excited

Wol — Tue, 24 Jun 2014 18:31:38 +0000

I believe you're correct - patent law does make that distinction.

But the patent lawyers believe that using a computer (a previously patented device) to carry out a list of instructions (an unpatentable "recipe", or program) creates a new, patentable machine.

Notwithstanding patent and case law to the contrary ...

Cheers,
Wol

don't get too excited

Wol — Tue, 24 Jun 2014 18:29:14 +0000

Unfortunately I don't have any links. If you can find any information on childhood brain development and at what age children are typically capable of learning certain concepts, then I'm sure you'll find it.

One of my strong memories (which influences the way I teach my grandson) is of my father trying to teach me about negative numbers. It was just completely over my head. And the way I used to read his "Understanding Science" magazines because they had lovely pretty pictures and I loved reading and the language they used, but a lot of what they were trying to teach I couldn't understand.

Then, at the right time at school, all these things were taught, and I remembered back to all this stuff, and I just knew what the teacher was teaching. A classic example is valency. I believe it usually took a week or two for the topic to really sink in (I know various people in my class had horrible trouble with it). It took me 30 seconds, to remember back to the articles in those magazines, and say to myself "Oh!! THAT'S what it was trying to say!!".

So when I'm explaining something to my grandson, I never worry if he understands me or not (of course I want him to! :-) But if he's interested and understands, he'll come back at me, and if it's over his head I trust he'll remember and come back to me when he's ready.

Cheers,
Wol

don't get too excited

Wol — Tue, 24 Jun 2014 18:17:51 +0000

> Patents on drugs are useful.

So pharmaceutical propaganda would have you believe ...

I think the evidence is that even there they do more harm than good ...

Cheers,
Wol

don't get too excited

Wol — Tue, 24 Jun 2014 18:15:37 +0000

Just because the matter is invalid subject matter, doesn't stop the European Patent Office from granting invalid patents.

It never stopped the US Patent Office from granting invalid patents, either.

Cheers,
Wol

don't get too excited

niner — Tue, 24 Jun 2014 17:16:49 +0000

Citation needed.

don't get too excited

Cyberax — Tue, 24 Jun 2014 17:12:18 +0000

> I just assumed it is. Because it sure as hell is in the European Patent Convention

Yet there are tons of software patents in Europe. Mostly on codecs and similar stuff.

don't get too excited

Cyberax — Tue, 24 Jun 2014 17:11:05 +0000

Patents on drugs are useful. As are certain mechanical patents.

don't get too excited

Asebe8zu — Tue, 24 Jun 2014 11:43:09 +0000

This is interesting.

It reminds me about the distinction between freedom of expression and expressions you don't approve of.

Do you have links to information on this transition to abstract thought?

don't get too excited

Tjebbe — Tue, 24 Jun 2014 10:01:08 +0000

> Probably this is just the way of the world to tell me, that ALL patents are rubbish, and the concept of patents itself is just some blatant rent-seeking effort which should have been erased from this planet in the 18th century.

They are, and they should.

A few countries have had a some bright moments in history and abolished patents (the Netherlands, for instance, but they later reintroduced them. I can only guess because of international pressure).

don't get too excited

Seegras — Tue, 24 Jun 2014 08:53:26 +0000

>> It says so in the law: Mathematics are not patentable.
> The law says no such thing. Go and read it, please.

Oh hell. I did. I know there's some ruling that says so, but indeed, it's apparently not in the US patent law:
http://www.uspto.gov/web/offices/pac/mpep/mpep-9015-appx-...

I just assumed it is. Because it sure as hell is in the European Patent Convention:
http://www.epo.org/law-practice/legal-texts/html/epc/2013...

"2 The following in particular shall not be regarded as inventions within the meaning of paragraph 1:
a) discoveries, scientific theories and mathematical methods;
[...]
c) schemes, rules and methods for performing mental acts, playing games or doing business, and programs for computers;"

> So the question is whether software running on a general-purpose
> computer is a new machine. SCOTUS ruled that it is indeed so.

Ah yes, the fabulous Moby Dick Support Device: http://seegras.discordia.ch/Blog/the-moby-dick-support-de...

don't get too excited

nybble41 — Tue, 24 Jun 2014 03:10:24 +0000

> ... even if the hardware is a general purpose reprogrammable part, the whole machine when it operates on a particular task includes the state of the electrical signals, etc. which is not an intangible.

On this much, at least, I think we can agree. While the data (including the software) is representation-agnostic and thus intangible, an operating computer is a tangible thing, including the states of all the signals inside it at any given time.

Fortunately, the court seems to agree with me that this makes exactly zero difference in regard to patent-eligibility. A claim for an abstract business method on a general-purpose computer is a transparent attempt to claim the abstract business method; the use of a computer to speed up evaluation of the business-method algorithm is incidental and obvious. If future courts show the slightest tendency toward consistency, they'll apply the same logic against attempts to patent other kinds of algorithms.

don't get too excited

raven667 — Tue, 24 Jun 2014 01:44:38 +0000

And if I understand correctly patent law makes a clear distinction between idea and implementation which is how they claim that something can be patented when it is run. The idea of an algorithm being distinct from an actual implementation running on hardware doing work.

Gene, and the case generally

Wol — Mon, 23 Jun 2014 21:03:42 +0000

> but the court is pretty clear that there are plenty of software patents that should exist - it just doesn't say where the line is

Is the court clear on that? As I understood from previous rulings discussed on Groklaw, the court couldn't get its head round "all software patents are patents on maths" (or didn't try) and as such was simply saying "some software patents MAY be valid, but ones like this aren't". It couldn't get its head round "all software patents are like these ones, therefore they are all invalid".

It wasn't (sadly) closing the door on software patents, but it wasn't legitimating them either.

Cheers,
Wol

don't get too excited

Wol — Mon, 23 Jun 2014 20:56:43 +0000

The point is, I think (and I'm sure people will come up with counter-examples), that if the *same* *concept* has multiple physical implementations, then that concept is abstract.

For example, the byte 11010001 may be represented in voltage as +5,+5,0,+5,0,0,0,+5. Or it may be represented in magnetism by NNSNSSSN.

Thing is, we're separating the *idea* from the *implementation*. And no disrespect to the readership here, but in my interest in "how people learn", I came across the interesting facts that (1) people aren't capable of separating concept from implementation (ie they don't understand the abstract) until they are about 14. Even more interesting, (2) it may be OVER HALF of people *never* *make* that transition, ie the majority of people are incapable of abstract thought!

Unfortunately, I get the impression that a fair few people here haven't made the transition ...

And what do you do when (as I think is the *normal* *real* *life* *behaviour*) that byte above is represented as SNSSSNNS ? Done that way because it seems to be less error prone. But what's a 1 and what's a 0?

Cheers,
Wol

don't get too excited

raven667 — Mon, 23 Jun 2014 18:39:43 +0000

This is largely a semantic discussion with a lot of opinions which are hard to pin down to anything concrete. The hair I am trying to split is that there may be a fundamental difference between an ethereal "idea", say written down in a book, and an actual implementation in a machine, even if the hardware is a general purpose reprogrammable part, the whole machine when it operates on a particular task includes the state of the electrical signals, etc. which is not an intangible. So sound in the air or graphite on paper is not a machine, electrochemical signals in a brain might be but we consider people separately and try not to punish thought-crime even though they are made out of stuff.

> Changing the inputs or the software does not transform it into a different machine. It *replaces* other special-purpose computing devices; it doesn't *become* them.

I don't think that is an unassailable point, that is an assumption and depends on semantics and belief, others with different assumptions could legitimately come to different logical conclusions. If you want to understand other people and other arguments you have to be able to separate out what is real beyond debate and what the assumptions are.

don't get too excited

nybble41 — Mon, 23 Jun 2014 15:04:00 +0000

> bits actually are electro-magnetic states of real materials

Bits may be *represented by* electromagnetic states of real materials, but that isn't what they *are*. The meaning we ascribe to bits is independent of how they're represented, whether through electrical charge, magnetism, phases of matter, optics, or even mechanical cogs and gears. A bit is an abstract logical 0 or 1; the physical form is an irrelevant implementation detail.

Under your interpretation, words like "intangible" or "virtual" or "abstract" would have no meaning whatsoever, since all ideas are ultimately represented in some physical form, whether that be sound waves in air, graphite patterns on paper, or electrochemical signals in someone's brain. The distinction isn't between things that have a physical form and things that don't, but rather between things where the physical form is important and others where it is merely one of many equivalent representations.

> The fact that our computers are transistor based and not gear based like a fire control computer or made of brass like the difference engine does not change this fact.

Indeed. The software remains pure math and not patent-eligible regardless of the specific makeup of the general-purpose computer that evaluates it. Whether made of brass gears or transistors, it remains a machine specifically designed to evaluate arbitrary inputs according to software instructions--*any* inputs, and *any* software, all with the same machine. Changing the inputs or the software does not transform it into a different machine. It *replaces* other special-purpose computing devices; it doesn't *become* them.

don't get too excited

raven667 — Mon, 23 Jun 2014 04:24:55 +0000

I don't think bits are intangible, virtual or only existing in the realm of ideas, bits actually are electro-magnetic states of real materials, the complexity of the machine around them and the fact that we as people ascribe them meaning does not change the fact that everything which happens on a computer, on the internet, happens in the real world, not some imaginary realm of pure math.

This keeps coming up where people who should really know better keep treating computers like magic wizard boxes rather than the real physical machines they are. The fact that our computers are transistor based and not gear based like a fire control computer or made of brass like the difference engine does not change this fact.

don't get too excited

Cyberax — Mon, 23 Jun 2014 02:32:10 +0000

>It says so in the law: Mathematics are not patentable.
The law says no such thing. Go and read it, please.

The relevant passage is:
> 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.

So the question is whether software running on a general-purpose computer is a new machine. SCOTUS ruled that it is indeed so.

don't get too excited

dlang — Mon, 23 Jun 2014 00:26:02 +0000

> The software that computes the cure time of rubber based on temperature measurements cannot be patented because it is merely an idea. A device that measures the temperature of the curing rubber and turns off the heating element once curing is calculated to be complete can be patented because it does something (turns off the heat).

If you read the Alice decision, you will see that it's more than just "turning of the heat", it's that the process did something that hadn't been possible before, measuring the temperature inside the mold on a continuing basis to calculate when the curing was completed.

There are other cases where the SC has rules that merely using a computer to set off an alarm when something crosses a threshold is not patent eligible.

don't get too excited

fest3er — Sun, 22 Jun 2014 23:25:03 +0000

No, on a computer, you have intangible, virtual things called 'bits'; your 'object' is naught but a collection of bits. Intangible, virtual things are ideas. Ideas cannot be patented. Only tangible, physical things that do something real and physical or cause something real and physical to happen can be patented.

Print out a program's source code. Lay it on the floor. Wait for it to do something or cause something to happen. You'll die first.

Oh, wait. It's source code. So compile it. Place the drive (rotating or flash) that contains the compiled program on the floor and wait for it to do something. You'll still die waiting.

Like words printed on a piece of paper, software is an intangible idea. You cannot touch it. Like a novel, you can only imagine it. Without something real and physical to follow its instructions, software can do nothing.

Patents apply to real, physical, tangible things that do something real and tangible. Copyrights apply to virtual, non-physical, intangible ideas that can only be imagined. Software lies in the realm of ideas. And like all ideas, software can be copyrighted. Not patented.

The software that computes the cure time of rubber based on temperature measurements cannot be patented because it is merely an idea. A device that measures the temperature of the curing rubber and turns off the heating element once curing is calculated to be complete can be patented because it does something (turns off the heat).

don't get too excited

Seegras — Sun, 22 Jun 2014 20:53:38 +0000

> No, the "slide to unlock" *switch* is a physical device which was already
> patented some 300 years ago.

As mentioned [citation needed]

> The "slide to unlock" *widget* is software form of a mathematical
> function mapping a sequence of numerical inputs from a touch input device
> to a set of RGB pixels on a display device plus a boolean "unlock"
> output, with the net effect of *simulating* a "slide to unlock" switch.

Alright, you tell me, this is illegal to patent in the first place, because it's math:: http://www.uspto.gov/web/offices/pac/mpep/s2106.html

The physical device itself is also obvious, and covered by prior art. Namely there's an Egyptian hieroglyph that appears at the latest in 1290 B.C. depicting it.

don't get too excited

Seegras — Sun, 22 Jun 2014 20:47:24 +0000

> The original slide switch was patented around 300 years ago in England.

Ah, by the way, [citation needed]

don't get too excited

Seegras — Sun, 22 Jun 2014 20:43:22 +0000

> The original slide switch was patented around 300 years ago in England.

What??? Which complete moron granted that? It's more than 3000 years old!

http://seegras.discordia.ch/Blog/patents-on-bronze-age-te...

Probably this is just the way of the world to tell me, that ALL patents are rubbish, and the concept of patents itself is just some blatant rent-seeking effort which should have been erased from this planet in the 18th century.

don't get too excited

Seegras — Sun, 22 Jun 2014 20:38:43 +0000

> However, this ruling does not invalidate all the software patents.

I frankly can not understand what is even the issue there. It's illegal. Point. It says so in the law: Mathematics are not patentable. I don't understand why anyone could squabble about it.

Unless you want to define Pi=3.