USPTO Asks For Input On Software Patents

New submitter MouseTheLuckyDog writes "The patent office is reviewing its policy on software patents and is asking for feedback (PDF). Groklaw reports that the USPTO will be hosting a pair of roundtable sessions in February, during which the public will have the ability to attend and put forth their viewpoints. From the article: 'It's obvious the USPTO realizes there is serious unhappiness among software developers, and they'd like to improve things. Software developers are the folks most immediately and directly affected by the software patents the USPTO issues, and it's getting to the point that no one can code anything without potentially getting sued. I don't wish to be cynical, though, as that's a useless thing. So maybe we should look at it as an opportunity to at least be heard. It's progress that they even thought about having a dialogue with developers, if you look at it that way.' If you can make it to Silicon Valley on February 12 or New York City on February 27, go and make your voice heard."

Comments

I Would Like To Suggest "Accountability"

[Press2ToContinue]

I would like to see a new law on the books: "wrongfully or negligently issuing a patent", to be applied as follows:

In the case where a patent is declared invalid, I would like to see the issuing patent office and/or examiner held responsible for damages done....

And to reimburse the patent applicant for:
1) the fees charged for granting the patent
2) legal fees incurred by the patent holder in attempting to defend the patent before it is struck down

And to reimburse any party who is financially damaged by the patent office having wrongfully issued a patent, such as
3) to any company which licensed the patent: any license fees paid out to use the patent
4) to any company which was sued for infringing on the patent: court costs and damages

Patents are applied for in good faith. If the recipient can be irreparably damaged due to negligence or other actions which wrong the recipient, shouldn't there be legal recourse?

Do you think the USPTO might hold "inventiveness" and the "obviousness" tests, and the search for prior art to a much higher standard? Do you think they might have the motivation to remedy any weaknesses in the system and keep on doing so?

Accountability anyone?

Re:I Would Like To Suggest "Accountability"

[RobertLTux]

what i would suggest is pure software patents be BANNED (and all currently active software patents voided)

now if software is some part of an actual physical product (ie something that would go THUD is dropped) and is an intergral part of said physical product then you can have a patent on the entire setup.

also there should be a rule of "must infringe on all parts" for a patent to be violated (dropping out clauses that don't apply ie claim for water use when the infringement is land use if there is a land use clause)

Re:I Would Like To Suggest "Accountability"

[Hentes]

now if software is some part of an actual physical product (ie something that would go THUD is dropped) and is an intergral part of said physical product then you can have a patent on the entire setup.

I never understood this argument. If the software is purpose-built for your hardware, then there's no use in copying it without said hardware. Here in Europe a similar precedent gets misused to push all kinds of software patents. Getting a patent on the hardware part only should be enough.

Re:I Would Like To Suggest "Accountability"

[h4rr4r]

How could they sue?

They had the patent until the law changed, so nothing I see new there. Lots of things are one way until the law changes then another.

Re:I Would Like To Suggest "Accountability"

[spire3661]

If this were true,the entire American Populace could sue Congress for extending copyright on works ex post facto. Public domain status is payment for a limited monopoly. Once Congress changed the rules, WE THE PEOPLE lost a good chunk of the social bargain. The door swings both ways.

Re:I Would Like To Suggest "Accountability"

[morgauxo]

Accountability yes, damages no.

Think of your own job. Whatever you do, how much can it cost the company when you screw up? Do you use any expensive equipment? Does your performance affect the decisions of any large account customers? Do you handle large quantities of product or highly valuable products? Can you afford to replace them? That would be the equivalent of taking damages from the individual examiners.

Damages from the issuing office might make a little more sense but ultimately that would just be punishing the tax payers.

What we need is a change in accountability. Performance is measured mainly on how many patents the offices issue. They are constantly being flooded with applications and the idea is that the patent offices need to keep up or this will harm the economy by slowing businesses down. It's also believed by politicians that the number of patents granted to US companies is somehow a valid measure of how our technical industry is doing vs other countries. That attitude needs to go!

Instead, hold patent offices and examiners accountable for quality of patents granted or not granted. This could be measured by percentage of patents invalidated in court and percentage of rejects that succeed later. That last one might be tricky to measure but it would be important too. Otherwise, maybe keep looking for quantity but also look for quality. It should affect their job reviews, raises, promotions and such just like happens for any other kind of worker. If they are really bad... they get fired.

This will mean huge delays in getting a patent granted given the current quantity of applications. That flow needs to be controlled! One sollution might be increasing the cost to apply for a patent so that only worthy ideas are worth attempting. The problem with that is a price that keeps larger companies in check shuts out individuals and small companies entirely. Instead.. I propose an application fee that gets larger based on how many applications you have in the system. Large companies still might get around this by having their individual employees apply for the patents and then transfer them back to the company. Some rules will be needed to prevent this. Maybe something stating that any contract requiring that an employee transfer patents from the employees name to the company are not valid. The company must apply for the patent itself or the employee might use it as leverage for more money or even sell it to someone else.

Re:I Would Like To Suggest "Accountability"

[Anonymous Coward]

The problem is that there is no single person whose fault the bad patent is. Consider if A (arbitrarily wealthy) and B (not) independently invent and locally produce X, but only A files for it, unaware of B. B resolves not to apply for a patent. Patent examiner C, also unaware of B, grants it. Company D hires B and begins producing Y, a variant of X that 'infringes'. After a few years, A's patent is granted and he (now a fairly large player) becomes aware of D (a very large player) and (because Y was released after he applied for the patent on X) sues. D, due to their ace-in-the-hole of B, cheerfully goes to court.

As far as A is concerned, he invented something, patented it, and is producing it.
As far as B is concerned, he invented something, but didn't bother patenting it. Doesn't matter, as his prior art should be enough to protect him against infringement claims.
As far as D is concerned, they are safe because of B's guarantee he won't patent it.

So then A sues D, after a long battle the patent gets invalidated, so who is to blame? Not A, he invented and patented the damn thing legally. Surely not B, he's taking the Benjamin Franklin (I think) approach. If B isn't to blame, I can't see a real reason that D is. Do you seriously expect C to scour the entire country himself to find B?

The correct answer would seem to be that if A and B came up with the invention, the patent should be invalid on grounds of obviousness, not prior art. But just because Newton and Liebnitz both wrote about calculus doesn't make their advances obvious. If math were patentable, and Liebnitz invalidated Newton's patent on integration, would you hold C responsible for not considering calculus obvious? Nobody in my example, I believe, "deserves" to lose a few billion. Various parties were simply unaware of each other.

Re:I Would Like To Suggest "Accountability"

[Shetan]

Doesn't the U.S. Federal Government have Sovereign Immunity?

Re:I Would Like To Suggest "Accountability"

[Steauengeglase]

I'll settle for the patent office taking that "non-obvious" bit into consideration.

Re:I Would Like To Suggest "Accountability"

[StripedCow]

What I would suggest is the following. Currently, the USPTO gets paid for every patent which they validate. This is so obviously perverse that it actually hurts.
Now instead of letting that money flow into the hands of the USPTO, it should go somewhere else, perhaps it should go into education.
This, imho, would already be a huge improvement.

Let's take it one step further. Let the patent office PAY for each patent they validate. A second government could then pay them back based on the societal impact of the patents they approved (measured, say, 5 years after validation).

Thus, with this two-tier model, lack of societal impact means a loss for the USPTO.

This means that, under this model, the USPTO will not so easily approve simple things such as "one click shopping" because they might lose on it on account of a lack of social impact. Similarly, patent trolling will be actively barred by the patent office (no product means no social impact). However, a patent for a new medicine may be approved.

Re:I Would Like To Suggest "Accountability"

[suutar]

Patents are applied for in good faith.

Your optimism is refreshing.

Re:I Would Like To Suggest "Accountability"

[Sarten-X]

I understand how a software keyboard can emulate a physical keyboard, but how about a webcam, or a graphics card, or a mobile phone.

Just like we've done many times with sound cards, old graphics cards, and serial devices. A piece of hardware is emulated, and it's connected to a non-infringing piece of real hardware. If the real value is in the software (as it would be in a software patent replaced by a patent on purpose-built hardware), then any similar hardware will do, and the "protected" software isn't really protected any more.

Re:I Would Like To Suggest "Accountability"

[CanHasDIY]

Khaaaaaaaaaaan!

Ban them!

[eksith]

A program/software/instructions for a computer, whatever you call them, should be covered under copyright, not a patent. Algorithms should be treated as works for art. Functional (or imperative or whatever) art, but creative works nonetheless.

The end.

Re:Ban them!

[jedidiah]

Computer programs are already treated as copyrightable works. So there would be no new unintended consequences.

Re:Ban them!

[Ikonoclasm]

The code and algorithms are already copyrighted. Every creative work in the US is automatically copyrighted, so that concern is moot.

Secondly, algorithms can't be patented. The law explicitly forbids the patenting of math. You can only patent an implementation of the math.

Strictly speaking, software should be unpatentable. Software is, at its most fundamental level, pure math. However, much like gene patents, the US courts have decided to conveniently ignore the law where it states that Nature can't be patented (math and genes fall under Nature) in order to allow industry to prosper. Unfortunately, by granting individuals temporary monopolies, the USPTO has insured that those industries have become legal minefields that are stagnating out of fear of litigation.

Re:Ban them!

[WaywardGeek]

Copyrights and trade secrets protected the software industry just fine before the USPTO opened the flood gates on software patents in the early 90's. They should stick to the original intent of the constitution, and protect the free flow of ideas by banning patents on mathematical algorithms (which includes software, IMO). They should not overturn the patents they've granted - that would harm the companies that filed them - but going forward, patents should cover something more than what can be executed in any mainstream computer language. If I can violate your patent simply by writing C code, it should not be patentable.

Software patents have resulted in:
The Open Invention Network
Peer to Patent
Oracle suing Google over Java
37 Android related patent suits
Nearly killing RIM
Linux patent suits ...

I'm afraid we're at the point where the anti-software-patent people warned we'd be. Small companies live in terror of being sued over any software they write. Big companies waste billions of dollars in court. Coders like me intentionally "code dumb", to avoid accidentally using a patented software idea. It's a terrible waste, and it makes me very sad to see America throwing away it's software innovation lead in this way. Thank God software patents weren't around when we wrote so much of the software that still powers the world. If they were, we'd all still be renting time on IBM mainframes. Just imagine a world where Donald Knuth patented all his ideas.

Re:Ban them!

[WaywardGeek]

There are other problems besides the expense. EDA companies have patented lots of very obscure algorithms, rather than keep them trade secret. If you take away the patents, then their competitors will be free to use their ideas, which could be fatal to some companies.

Unfortunately, it's a moot point. I went and read the poster's links, and the USPTO has no intention of considering any broad limits on software patents. They're just seeking feedback on nonsense like whether XML can be used in a claim. These meetings wont be worth anyone's time.

Re:Ban them!

[WaywardGeek]

Requiring source code wouldn't do much to discourage software patents.

I'm listed as an inventor on 22 patents so far. Several of them are software patents. I didn't file any originally, until a competitor patented a key software algorithm we'd used for years at a previous company - an algorithm we could not function without. Now I encourage that we patent all company-critical software ideas. Anyway, I've been filing them since the early '90s. My early patents all included full source code, and one has many thousands of lines from our commercial product - enough to actual run the algorithm on real designs. If you have a new logic optimization idea, have you fully disclosed how to implement it if you haven't included a netlist parser? Early software patents included full source code, which gradually was reduced until now a bit of pseudo code in the patent text is generally enough. The old requirement for source code didn't seem to slow patent applications.

Every patent lawyer knows that you can't violate a patent in your head. The same is true if you write an idea down on paper. You also can't violate a patent by writing it down in Notepad on a computer, or even if you save it, and share it with friends. However, if your idea is a sequence of steps a computer could execute, and you "run" it, chance are high that you've just violated somebody's patent. Being a software developer in America is a bit like being an illegal immigrant. You just keep quite, keep your head down, and hope that nobody bothers looking at what you're doing too closely.

My input on software patents...

[jkrise]

Input -----> Process -------> Output

100000 applications for software patents -----------> USPTO ------------> All rejected.

All software is, by definition, math. And all math, by definition, is not patentable.

Re:My input on software patents...

[jkrise]

Try rephrasing that, is what I'm saying.

I feel it will not work with the nutcases in the USPTO; so absent summary rejection of ALL software patent applications, and voiding of past software patents granted, nothing will happen. Instead of focusing on the USPTO which is a gone-case, I suggest a different approach when patents are used in litigation, to solve the problem. This is what I posted in Groklaw, on a related discussion:
--------------------
The PTO has a limited amount of time to inspect each patent (I believe it's around a day per patent).

This is the exact problem, and in the rest of your post, you have detailed how to DEAL with the problem. What I am advocating is, how to solve the problem? The fact that the PTO has just 1 day to inspect a patent, implies that innovation is happening at a rapid pace these days. So a roughly 2-decade monopoly on a patent in these modern days; is totally not justified, since entire businesses and ecosystems are impacted by such long term monopolies.

Consider that there are about 2 billion Windows devices worldwide, in about 2 decades. In just 2 more years, it is projected that there could be more than 2.5 billion Android devices, surpassing Windows devices.

So a patent that cripples Android for 2 decades means ENORMOUS incalculable harm to the progress of science and arts, which is the raisson d'etre of patents.

So the cure to the PTO having just 1 day to inspect a patent application, is to drastically reduce the number of applications, rather than hastily issuing dubious patents, re-examining and rejecting them, and further re-examining and validating a limited number of claims.

To reduce the number of patents filed, a severe penalty has to be levied on a patent found to be invalid on re-examination; when such a patent is asserted in a case. If a company faces the prospect of a $10bn penalty, compared to a $1 bn damages compensation; it will think a 100 times before using the patent in a court. Additionally, it will also reduce the need and motivation to apply for a patent in the first place, thus allowing the PTO a lot of time for review and examination of a vastly reduced number of applications, which are bound to be genuine, rather than frivolous.

Keep It Simple Stupid

[presidenteloco]

No software or algorithm patents.

If you really want to keep something exclusive, keep it hidden, call it a trade secret, and sue anyone who leaks it.
Unless you are Einstein, someone else will think of it fairly soon anyway, because it's obvious to those at the leading edge of whateever specialty, so keeping it a secret may be bad social form but is not really harmful.

NYC Feb 27 - I'll try to make it - Slashdot meet?

[MetalliQaZ]

Perhaps Slashdotters in the areas around these meetings would like to get together to plan, practice, and eventually travel to these meetings? Beers/Sodas after the meetings are suggested.

show us one good software patent

[johntromp]

Has the USPTO presented specific examples of what they consider to be excellent software patents? That should help focus discussion...

Horse has bolted

[Joce640k]

they'd like to improve things...

Too late for that, the damage is done. The patents they already issued are enough to destroy the software industry for the next 15 years at least.

Re:Questionable summary

[Grond]

software patents are illegal, since software is mathematics.

Good thing "software patents" don't actually claim software per se but rather methods of programming or using physical computing devices.

Consider this highly simplified claim: "A method for displaying a line on a computer screen, comprising using Bresenham's line algorithm to draw the line on the computer screen." Suppose that the patent applicant was Bresenham himself and the method was new.

This claim does not claim mathematics. The algorithm itself can, of course, be reduced to a proof, per the Curry-Howard correspondence. But "a computer screen" is neither software nor mathematics. No amount of thinking about mathematics well ever result in a line appearing on a computer screen. You need hardware to do that. And since the computer screen is an essential element of the claim, mathematics alone (such as the aforementioned proof) will never infringe it.

If Congress wants them to be legal, they have to write a statute to do so.

What in the patent statute prohibits software patents? How is new, useful, and nonobvious software not an improvement of a machine (namely a computer), per 101? How is new, useful, and nonobvious software not a a new use of a known machine, per 100 and 101? It seems to me that the plain language of the patent statute clearly encompasses software, and that it is the judicially-invented "abstract idea" exception that potentially prohibits them.

Re:Questionable summary

[Grond]

If the computer screen is an essential element, then I don't need a patent license for my free software that draws lines using Bresenham's algorithm, because I'm not using Bresenham's patented computer screen!

The patent doesn't claim that the screen is itself new or patented. The patent claims the method of displaying a line on any computer screen. It's no different than a patent on, say, a method of catching mice using a bucket of water, a ramp, and a piece of bait dangled above the bucket. It doesn't matter where you get the bucket, the ramp, or the bait. The patent covers the method of using these pre-existing objects together to catch mice.

You can't have it both ways. If a physical object that I'm not selling or giving away is essential to the patent, then my free software is not covered by the patent.

I'm not trying to have it both ways. This is a long-settled area of the law, codified in the statute. The end users would be directly infringing the patent, and you would likely be liable for indirect infringement. Imagine a patented mechanical device that is held together with screws. If someone sells all of the parts, minus the screws, plus instructions on how to put it together and where to buy the screws, that's indirect infringement. In this hypothetical you're distributing the software knowing that end users will use it in an infringing manner with their computer screens (which, again, could be any kind of computer screen).

Re:Arrogance

[fritsd]

I haven't heard a single valid argument for why software is any different than any other discipline.

Really? Oh.

Here are a whole bunch of them, each one carefully reasoned out and commented on:
http://www.groklaw.net/staticpages/index.php?page=Patents2

"Software Patents
Here are some of the articles Groklaw has published on software patents, particularly in support of the claim that software is mathematics and hence unpatentable subject matter."

Re:Software: Patents or Copyright

[steve79]

You are incorrect, sir. You do not need to release the source code. This was true back in like the 1980s (or common practice then, anyway). Nobody release code for their software patents these days b/c it is simply not required. You must describe the general algorithm in sufficient detail to enable one of ordinary skill in the art, without undue experimentation.... this is a far cry from source code.