Ask Slashdot: Reducing Software Patent Life-Spans?

seattle_coder writes "Many have advocated for the elimination of software patents. The arguments generally are that software patents are handed out too easily, and that they're too difficult and expensive to fight. Some say that patents just plain don't make sense for software, which is such a dynamic technology. Given that the standard patent lifetime is 20 years, and software changes so rapidly, is the life-span the problem for software patents? Would reducing the software patent lifetime to 5 years or even less be the thing to do?"

Comments

This doesn't solve the problem

[Anonymous Coward]

No, because the process to get a patent can easily stretch five years. Also that doesn't solve some of the other fundamental problems with software patents, such as software being math.

Re:This doesn't solve the problem

[Odinlake]

Would reducing the software patent lifetime to 5 years or even less be the thing to do?

I think that would solve the problem, say, in the same way that we could stop obesity by poisoning all refined sugars with arsenic.

Better to just abolish them altogether then.

I've filed two (fairly insignificant) software patents for my employer. Of course this is one way they fund their research and can afford having smart people (and some others like me) sitting around and coming up with fancy stuff. The process isn't inherently bad (imho). The problem both with patents in general but with software patents in particular is that it's so difficult to distinguish between good and bad patents. One thing I think is really bad is that it is so easy to patent the obvious solution to a problem that just hasn't been considered until now because (e.g.) some necessary technology hasn't been around. But how can we make a rule to disallow patenting of the obvious? 10 years from now anything could seem "obvious". Another "bad" thing is too general patents. But a flip side of that coin is that if your patent isn't general enough, someone might fine a tiny little thing to change and thereby circumvent it with hardly any effort at all.

So all in all I would probably be for abolishing software patents or at least making them far more restrictive until it becomes clearer how they can be well used. But this isn't because they are inherently bad, but because we haven't figured out how to define good boundaries yet. Within specific fields I think it might well be possible to have useful software patents.

Better to eliminate them altogether

[Anonymous Coward]

We shouldn't be able to patent software for the same reason we can't patent mathematics. Copyright protection is sufficient and suitable for software.

Re:Better to eliminate them altogether

[blair1q]

You can't patent math, but you can patent a process that applies it.

In other words, if you come up with a theorem, you can't prevent me from coming up with a new theorem that uses or requires your theorem.

But you can patent any new process for converting matter or data into other matter or data that depends on the truth of that theorem. If my theorem includes your theorem (rather than just requiring it to be true to prove mine is true), then I would owe you royalties because any process derived from my theorem would have to include processes derived from your theorem.

The trick is to determine all the processes dependent on your theorem and claiming them before prior art undercuts you.

E.g., you can't patent 1+1=2, but you sure could patent the abacus, and everything that works like an abacus, if you'd been there.

Re:Better to eliminate them altogether

[Jane Q. Public]

"A patent on software coupled with a specific underlying machine is about as far as patents should be allowed to go, and only because of the modern reality of industrial processes and control equipment."

That argument was shot down over 100 years ago, in court cases regarding to player piano rolls that controlled machines... the pianos.

The courts ruled (quite properly) that the "software" -- the rolls that controlled the pianos -- were simply expressions of written music, and therefore the appropriate law for protecting them was copyright law, not patent law. They reasoned that a written work is a written work, no matter what physical form it may take, and regardless of whether it controlled a machine... a piece of punched paper telling a machine what to play did not fundamentally differ from a printed piece of paper telling a human musician what to play. It was exactly the same music, only the physical form had changed.

Recent years have brought nothing new to the table. There is no real difference between a piece of software (which is ultimately written by human beings) telling a computer what to do, than an English translation of the software telling a bunch of people with pencils and paper what to do. The only real difference is speed... nothing fundamental has changed at all.

Re:Better to eliminate them altogether

[CastrTroy]

I have no problem with software patents so long as they want to divulge full source code as part of the patent, so that we don't have to redo their discovery when the patent expires. It's seems a little odd that they can get a patent, and then not release the source code in order to make the patent useful. There is no reason that the same piece of source code should get patent, copyright, and trade secret protections.

Re:Better to eliminate them altogether

[robbak]

Seems like you've got my point :D

You can knock on a drug, if you have molecular knuckles. "Manufacturing process" seems to be how we slid down into this mess to start with.

Well, maybe I wasn't being completely serious, at least for a general patent rule. In IT, however, it applies.

Re:Better to eliminate them altogether

[Dwonis]

No, it's not. De-compiling to source then re-compiling using a different compiler will result in different (machine) code, the copyright won't match. The only way to protect software is with short term patents...

It doesn't work that way. As it turns out, judges are not stupid. Copyright law already applies to translations of a work, and for software, there's the abstraction-filtration-comparison test.

if you just spent 5 years full time building a mission critical embedded RTOS for a commodity router you'd want to be paid for it.

In my experience, if you're building anything of value, you have customers and/or investors lined up while you're building it. Patents really don't make much of a difference.

Re:Better to eliminate them altogether

[dgatwood]

Basically, in such a situation, you get copy-cat trolls instead of patent trolls. How would you propose to solve this problem, if patents are eliminated all together?

You seem to be under the mistaken notion that this is something that someone needs to protect against.

Writing good software is not trivial. If somebody can knock off your "invention" in six weeks of coding, then your "invention" can't possibly be sufficiently non-obvious or non-trivial to be worthy of patent protection anyway. And if it takes them a couple of years, by that point, your software should be two years ahead, which means that it should not be possible for them to realistically catch up.

In short, as Bill Gates once said, "Innovate or Die." That's the software business. If you are resting on your laurels, you aren't contributing anything useful to society, and you should not be able to continue profiting from it for an extended period of time.

Re:Better to eliminate them altogether

[TheRaven64]

The real argument against software patents is that they provide nothing of value. The purpose of a patent (in theory) is to encourage disclosure, so that other people can implement the same idea without having to reinvent it independently. When was the last time you heard of a software developer wanting a solution to a problem and doing a patent search? Developers at big companies are explicitly told not to do this, because it increases the company's liability if they are caught infringing a patent.

Do software patents encourage innovation? There's little evidence for that. There are two sets of software patents that people often cite as evidence that they do: RSA and H.264. There was a sufficiently large market for secure communications that public key cryptography would have been developed by a company interested in selling products based on it sooner or later. Even if they hadn't, then after a few years the official secret in the UK would have expired and GCHQ would have published it (as did happen, although interestingly the fact that it had been invented decades earlier did not count as prior art).

H.264 is easier, simply compare it with OpenGL. On the implementer side, a lot of the players are the same. The difference is that OpenGL is developed as a royalty-free standard so you see cheaper hardware OpenGL implementations than hardware H.264 implementations. Take a look at the list of MPAA members. Then a list of BluRay manufacturers. Then a list of mobile SoC manufacturers. Then the (admittedly much smaller) list of browser makers. All of these have a financial interest in the existence of a useable, open, video standard. Do you really think none of these would have put any money into video CODEC R&D without software patents? Google was willing to spend $133M for a royalty-free video standard...

a word of advice...

[Lead Butthead]

the one that's holding all the cards isn't going to ask for a new hand. the broken patent system serves the interests of large corporations, and they'll fight tooth and nail against any changes that is against their interest. in summary - dream on.

You're framing it wrong

[cultiv8]

is the life-span the problem for software patents?

No, the problem is that software is an algorithm and shouldn't be subject to patent law anyways. It's like trying to patent a mathematical formula.

Forget Patents, what about copyrights?!

[telchine]

I dunno about patents so much. I think used properly they are useful. Unfortunately they are more often used to stifle innovation than they are to encourage it like they were intended to do!

My main worry is copyright. I think copyright law needs dramatic alteration! To me a Rudyard Kipling work that has been out-of-copyright for some time and bastardized by Disney to create "The Jungle Book" is more worthy of protection that a 20 year old computer game like "Tai-Chi Tortoise"

The first has a great deal of artistic merit, and will be recided in its original form for hundreds of years. The second will be in copyright for many years, despite the fact it has slipped from memory already!

There should be some distinction between a work that has potential monetary worth 75 years after its creation, and something that has no worth 5 years after creation.

Re:Forget Patents, what about copyrights?!

[robbak]

There should be some distinction between a work that has potential monetary worth 75 years after its creation, and something that has no worth 5 years after creation.

I have a different opinion. Any work that still has relevance 20 years after it's release has become an important piece of cultural property that _desperately_ needs to be in the public domain, and the property of all.

Re:Forget Patents, what about copyrights?!

[cpt kangarooski]

The sole determinant for whether a work should be in copyright is whether it will benefit the original creator for it to remain in copyright.

Not only no, but Hell No!

The issue is whether it will benefit the public for it to remain in copyright. Copyright exists for the sole purpose of serving the public interest, by promoting the progress of science. It does this in two ways: First, by encouraging the creation and publication of the greatest number of works which otherwise would not have been created and published; Second, by limiting the scope and duration of copyright as much as possible.

It doesn't really matter what authors want, except insofar as we're essentially trying to 'bribe' them with the offer of a copyright, and it needs to be the least bribe possible that still gets them to do what we want, viz. create and publish works. Of course, at some point there is an issue of diminishing returns: It might be impossible to get a particularly recalcitrant author to dash off a few lines without promising him the moon. In that case, much as we might like having that work available, the price is too high, and we'll just have to pass. So long as the public gets the most bang for its buck, as it were, who cares whether authors like the deal. It's not as though they have a lot of alternatives.

Copyright would be just fine if the term was limited to something sane.

No, it would take more than that. In fact, I'd say that the number one reform we need is to stop automatically granting copyrights; we need to require registrations and impose some strict formalities. A system of short overall terms comprised of even shorter initial and renewal terms would probably be number two, and a broad exception for non-commercial infringement by natural persons would probably be number three.

Re:Forget Patents, what about copyrights?!

[cpt kangarooski]

Thanks.

It was never intended to generate monetary wealth. It was intended to generate cultural wealth by providing monetary incentive.

Not even that; it's intended to generate cultural wealth for all by providing a possibility of a monetary incentive. The genius of copyright is that whether a particular copyright has actual economic value depends on the author, the publisher, and the receptiveness of the public. Most works, frankly, are economic flops. But so long as authors who otherwise would not create and publish works are drawn to do so by the chance that they could exploit their copyright for money, things keep rolling along.

Registration, renewal, short terms, and formalities improve things greatly:

Requiring authors to register their works in order to get a copyright (beyond a minor level of protection for manuscripts), along with a token registration fee (more for the purposes of separating the wheat from the chaff in the author's opinion, rather than to fund the system), we avoid copyrights on a bunch of works that even the author doesn't think are worth protecting. And who are we to second-guess the author?

Short terms and frequent renewals (probably again with a token fee) let us weed out flops more quickly than if we just granted a single long term to everything and waited for it to expire. If a copyright holder doesn't think that a work is worth renewing, again, who are we to second-guess them? Let the work enter the public domain earlier than it might have otherwise.

Only a handful of works have a copyright that's worth a non-trivial amount of money more than a few years after publication, or sometimes after even less time. (When was the last time you bought yesterday's daily newspaper? Or a bestselling novel from 1911?) Let's reform the system so that we quickly and efficiently weed out the others and get them in the public domain. This will solve a lot of the problems. It'll be easier to decide what's to be done about Mickey Mouse when it's just Mickey Mouse, and not a million other things being dragged along for the ride.

Re:Forget Patents, what about copyrights?!

[Rakishi]

No, it would take more than that. In fact, I'd say that the number one reform we need is to stop automatically granting copyrights; we need to require registrations and impose some strict formalities. A system of short overall terms comprised of even shorter initial and renewal terms would probably be number two, and a broad exception for non-commercial infringement by natural persons would probably be number three.

Congratulations, you're just killed 90% of open source software. And 90% of works created by individual artists and 90% of works put online by individuals.

Those are all entities who cannot afford to copyright all their works. It's simply not economically feasible especially if they create many small works. Nonetheless they benefit from having copyright protection as it prevents excessive outright copying.

Large corporations, who are the real source of problems, would simply pay the to them trivial fees for copyright and get on with their day. Trolls would probably also copyright the works of others, who didn't have the money to do it themselves, and then sue those who use them. That would be a fun lawsuit to see.

Not a great solution

[cjonslashdot]

Yes, the lifespan of patents is a big problem.

But in software, things change so rapidly that patent protection for even five years is an eternity: by then, it is game over.

The fundamental problem with software patents is that companies patent simple ideas. The Amazon one-click purchase patent is a prime example. These kinds of ideas should be considered "obvious" by the USPTO, but unfortunately these kinds of things are routinely patented. The result is that there is a minefield of patents around every simple idea, every basic thing that one can do in software. Anyone who wants to create a startup company around a software product is at great risk, and instead of investing their time and energy into product development they now have to invest it in legal research. That is not a very good state of affairs for an industry that thrives on innovation.

If patents are to be allowed to exist for software, the bar for what is not obvious should be much, much, much higher than it currently seems to be.

Better analysts perhaps?

[pixline]

Just stop granting random patents because the people who decide on 'em can't understand what's written in a paper.
Let real tech people judge and - eventually - grant a *limited* patent, but stop giving away things because people can't read a proposal..

The problems

[ciaran_o_riordan]

Making them 3 years would solve many many problems.

But, the TRIPS agreement says patents have to last 20 years.

However, the TRIPS agreement doesn't say that software has to be patentable. So countries could declare that software isn't patentable, and then create some new legal thingy called "petents", and say that petents last 3 years and that software innovations can be petented.

This would be hard work because some countries (USA for example) push the idea that TRIPS requires software patents.

http://en.swpat.org/wiki/TRIPS

Really, shortening the duration would be as much work, and there's always the risk that the monopolists will find some other nasty clause to stick in to make 3-year petents really harmful.

Let's just go for abolition. It will take time, but it's the only practical solution.

http://en.swpat.org/wiki/Why_abolish_software_patents

No, the problem is: software patents are stupid

[jopet]

Anyone who has ever programmed just a little knows that programmers invent algorithms every day and reuse somebody else's algorithm every minute. It is just a greed-fuelled idiocy to patent random bits of what programmers invent as part of their work. It is like patenting a mathematical proof or a law of physics.
Then there is another matter: when are two algorithms identical? When the code is exactly the same, including the naming of variables? No matter the variable naming? When they generate the same machine code on some machine? When they compute the same results (but this is undecidable formally)?

Software patents are a stupidity to make lawyers richer and to make the life of developers more miserable.

Patents are on Inventions, not ideas

[dwandy]

There's numerous reasons why patents on software are invalid. The most difficult to understand for a non-technical person might be that it can be reduced to math, which is already not patentable; so be allowing patents on software you are allowing patents on math. And demonstrating this in court for any scenario could be somewhat difficult.
But the simpler answer (from the start) should have been that you are not supposed to be able to patent an idea; you must patent an invention (implementation). In software it's either source code (covered by copyright) or it's an idea (not patentable).
The simple question from the patent examiner or judge for a software patent is a request to see the implementation. Since there's no implementation in "a way to navigate the web using a touch interface" it should get tossed. On the other hand, the specific implementation could (I suppose) be patented, but since the implementation is already covered by copyright, why bother with the (inferior) patent...?

The original point of a patent..

[sstamps]

..is NOT to reward someone who invested nothing other than a few neuronal sparks coming up with an idea, but to reward someone who risked a significant investment in time, money, and materials to be able to recoup that investment plus make a profit in exchange for sharing the fruits of that investment with everyone.

The VAST majority of software and business-method related patents are nothing that someone competent with ordinary skill in the field could not come up with as a common-sense solution to a specific problem, WITHOUT said investment OR risk.

That said, the problem is that bringing the original intent of patents back into the system is not in the best interests of those who decide who gets elected (it isn't the voters, natch) and, thus, will not happen without a revolution of some kind which circumvents their control.

Even civil disobedience won't get us very far because, ultimately, most people could care less that they are controlled like sheep, and believe that their masters are the "good guys" and can "do no wrong". As a result, get used to it until the system explodes and has to be replaced by something else, hopefully better, but such is not guaranteed. It is the time-honored tradition of the human race since time immemorial to let the decay of civilization outrun our ability to contain or mitigate it, to ultimately consume and destroy us.

I just wish it would hurry the hell up so we can start anew.

New angles

[Dragon_Hilord]

Instead of looking at this with that same crappy "algorithms and software are math" (apologies, but it hasn't worked before, and it certainly won't now (unless we maaaaybe reduce EVERYTHING to 100% pure mathematics)) view, let's start looking for new ways of explaining it to the masses - they don't get it. The fault is not entirely their own cause of their ignorance, we haven't properly educated them. What do you do when people don't get the idea in political situations?

Protest.
Civil Disobedience.
Lobbying.

I'm pretty sure if 1000+ hackers showed up on the doorstep of the whitehouse equipped with laptops running as much patent-violating software as they could possibly muster - that'd be a pretty good start. The only question is who has the gull to show up? Who has the balls to write that code? So far, the one thing the hacker community (and other code-centred groups yes) seems to lack is some kind of motivating factor. Perhaps, and this is just a thought, the right to code what we please?

I'm almost tempted to organize something of these lines myself, problem is I can already smell the -1 rating this is going to earn for "stating the obvious" or "trolling". Can't say I didn't try.

The issue isn't patent length...

[copponex]

The real issue is that patenting has moved from protecting the method of solving a problem into solving a particular problem.

For instance, I cannot patent mowing a lawn. I can patent a lawn mowing machine, even one that exactly copies the pattern created by another lawn mower that also runs on gas and uses a rotary blade to do it. As long as my lawn mowing machine is different enough to merit the patent, I think it's fair so someone doesn't come along and copy my machine.

Now people are patenting ideas. So when Amazon patents "one click shopping" and I write open source software with different libraries to complete the same task, they claim patent infringement, which is nonsense. They should be able to patent a particular method, which is really just their code, not an end result. Even then it's redundant, as it's already protected by copyright.

Large corporations do not want patent law sane, because it's their multi-million dollar playground, and they can crush innovation and startups by simply filing a few law suits. The first nation with decent infrastructure to fix their system or abandon it will lead the way into the 21st century.

tl;dr: The ability to patent end results instead of particular methods is why the current patent system is a failure.

Standards need IP assurance

[TheSync]

(speaking for myself), I feel that the biggest problem is when a collection of intellectual property is bundled into a standard (such as H.264), but despite the valiant efforts of standards development organizations to require their participants to document their participant's intellectual property in a standard, you can't really ensure that "submarine patents" by non-participants are not accidently infringed upon, and frankly even the SDOs don't have any real police power over participants that either on purpose or accidently don't disclose IP.

I feel that some organization (perhaps ANSI, or just the PTO) should be responsible for designating the most valuable standards (such as "national standards") for "mandatory IP declaration". This would start a process where within a set time (1-2 years), all intellectual property owners would have to "put up or shut up", i.e. declare that their intellectual property is covered by the national standard. If they don't declare during the limited time period, they forfeit their right to sue for infringement on implementation of the national standard.

Real businesses do not generally have a problem licensing known intellectual property protected by patent. It is the unknown that is the bigger risk, and makes adoptions of new standards slower (I've personally seen this at a previous position).

I say there arent ENOUGH software patents

[metalmaster]

People need to start patenting dumb software ideas; not just the great ones.

How many $obnoxious_sound or $picOfTheDay apps do we really need? Some dumb schmuck needs to claim those ideas and put a limit on entrants to that cesspool.

To patent or not to patent

[woboyle]

Generally, I am against software patents, however I do believe that there are some software innovations that should be patentable. My feeling is that they should be held to much higher standards of innovation, non-obviousness, and utility than current software patents are. I say this as the holder of a software patent related to adaptive systems that allows compiled applications to alter the structure and behavior of application classes at run time without writing code or recompiling the application source code. I honestly believe that the innovations I invented (and are in use by most semiconductor and similar high-tech manufacturing enterprises to enable them to tailor off-the-shelf manufacturing software to their particular enterprise without writing and integrating new code) meet the same standards required of physical devices and such. That said, most of the software patents I have read about, or studied in detail, are totally bogus and should have been laughed out of the Patent and Trademark Office. Period of time for software patents? 10 years max in my opinion. 20 years is absurd. As stated elsewhere in these comments, software innovations occur too rapidly to lock up some concept for that much time.

Contact your Senator!

[Jim Hall]

I am already working for "software" patent reform. I live in Minnesota, and this morning I visited with Senator Franken's office, discussing "software" patent reform. Sen. Franken is also the Chair of the Senate Subcommittee on Privacy, Technology and the Law, so this is a topic he's paying attention to.

The America Invents Act is already in progress, so the reality is that Congress won't have the appetite to pick up the patent topic again in the next few years. So I'm afraid that the window has closed to get a law to enact "software" patent reform. Temporarily, anyway. This is unfortunate, since AIA doesn't really address any issues related to "software" patents or patent trolls.

However, it is possible to have a procedural change made with the USPTO. So I'm working that angle now.

To do that, we need to build popular support, enough voices that the Subcommittee will hear. As cheesy as it sounds, that means we need your help to contact your Senator's office and voice your opinion. Seriously, call their office, write them a letter (handwritten carries the most impact, but typed will do), or visit them in their office. Ask your Senator to push for "software" patent reform. Have some examples of "software" patents handy, and feel free to make a suggestion for how to fix the system. I think the citizen review method is a workable option. I've been discussing this topic with Sen. Franken's office for a while now, so if you don't have a particular suggestion to offer, ask your Senator to see what Sen. Franken is up to. I'm not kidding, they do listen to that. You can reference Katie Topinka, in Sen. Franken's Minnesota office, as the staffer who's closest to this. I'm working with her on this (and hopefully she won't mind me mentioning her name.)

If you live in Minnesota, New York, Rhode Island, Connecticut, Oklahoma, Utah, or South Carolina, note that your Senator is already on the Subcommittee for Privacy, Technology and the Law. This is a topic they will listen to.

The best advice I have heard in politics is that it's your politician's job to listen to you and to take that back to Washington to get it done. I'm lucky that my Senator actually does this. That's why if you don't feel your Senator will listen to you on "software" patent reform, you need to send them to Sen. Franken, because he will work on this. And Senators do talk to one another, even across the aisle.