Hacker News Reader

> If software patents are done away with, I think we just throw the pendulum in the opposite direction and now everyone with money and strong development teams steamroll the competition.

Historically, the alternative to patents is trade secrets. For example, you have Coca-Cola's closely-protected formula. In the example you gave, it's just as likely that Skype would instead protect its technology as a trade secret, and reverse-engineering it would fall under current laws.

It might be possible to consider the rapid development cycle in the software industry and simply shorten the period of protection granted by a patent on software to something like 2 years, which seems long looking forward but is pretty short in retrospect. Companies like Lodsys would be forced to spend a lot more money acquiring new patents on a regular basis -- enough extra to continue their litigious cancer on the industry -- but wouldn't manage to substantially halt development or production in the software industry.

Even then, I think I'd still prefer to see patents abolished altogether for non-mechanical and non-chemical inventions.

I think the 2 year software patent might be the way to go. That's a pretty compelling argument to say that things move much faster in software so the timespans should just be shortened significantly. You might even be able to convince lawmakers with that.

Agreed -- I also think 2 years in software is would be considered sufficient time to have a technological advantage over your competition. Maybe even 3 years.

But I think it's currently 11 or something like that isn't it?

In 11 years I'll probably be half robot... that seems like a long time.

It's 17 years.

The 2-year software patent wouldn't fix other problems, such as the impossibility of determining novelty or nonobviousness in a field as varied and active as software. It's not wise to give the patent office an impossible job, when case law is largely based on presumption of validity.

It is worse than 17 years, actually. It is 17 years from date of grant, not date of filing. The USPTO takes years to review a patent, so your effective monopoly period is 17 years + USPTO review time.

In Europe and other jurisdictions the protection period is 20 years from date of filing (rather than starting at the date of grant). There's patent reform legislation currently being considered that would make this improvement to US patent law (and, yes, you understand correctly that the typical software patent review takes more than 3 years, often much longer. That makes the terrible job they do all the more inexcusable, even though companies do game the process with things like claim amendments). In any case, you're absolutely right that the 2-year software patent doesn't solve the novelty and non-obviousness issues.

I'd add that a 2-year software patent would have to be 2 years from date of filing. Anything starting at date of grant would be far too long. Unfortunately, a 2-year from filing software patent would mean that you'd need to speed up USPTO review by an order of magnitude and make it higher-quality. If anyone can come up with a way of doing that that is better than a robot just stamping "No" on every application, I'd be pretty impressed.

I've never understood why trade secrets need any special treatment as "intellectual property" in our legal system. Aren't trespassing and contract laws (non-compete, NDA, etc.) sufficient?

It comes into play when someone under an NDA violates it. It seems reasonable that a company shouldn't be able to use something that required violating an NDA to find out.

It's very hard to come up with genuine instances of software patents protecting small developers from larger competitors. The little guy usually can't afford the money and time to patent things, much less enforce them. But even disregarding that, assuming your patent isn't overly broad (in which case it should definitely not be valid), your competitors can just work around it and then steamroll you with their money and strong development teams.

Essentially, a competitor that could steamroll you without a patent would be able to steamroll you with a patent.

The one case I can think of where it would be reasonable for a patent to protect you is if a company offers to buy your product, looks at the code, and then has its developers create a knock-off based on your code. And even then copyright law is probably sufficient (analogous situations arise in other industries where copyright is the only law).

Stac Electronics is an example of Microsoft doing exactly what you described; it was patent - not copyright - law that allowed Stac to collect damages.

It was under patent law that they happened to file this suit, but similar copyright suits have been filed (this is why movie studios don't accept spec scripts and a prominent game developer deleted his Twitter account). And it's still pretty rare.

When chc mentioned "The little guy usually can't afford the money and time to patent things, much less enforce them", I assumed (s)he wasn't talking about little guys like Stac Electronics, a publicly traded company for whom a $9.7 million sales quarter was evidence of declining business.

http://articles.latimes.com/1994-06-22/business/fi-7159_1_pa...

As an indie E.U. based developer, I must confess that everything that I'm reading here about patent trolls is scaring me off. It will probably keep me from even thinking about building a product for the U.S. market for a while.

IANAL but I have to imagine that one simple rule would bring back the validity of patents as an innovation protector:

If you have a patent, you MUST have a product or be working on a product using the technology in question or you cannot litigate others using your idea.

Patents are required to protect the truly innovative individuals and groups from huge corporations who can, as you said, just steamroll them and steal ideas.

I'm afraid it will get a lot worse before we finally get this fixed, though.

That rule is easily circumvented. Just write a 10 line python script that references the patent. Release it as MyPatentApp version 1.0 and we're back to litigation square 1.

I believe that in such a quickly changing engineering landscape as software, best way to fix patent troll problem is to make software patents expire in 2-3 years.

Hrrmmm... good point.

So all that would do is require patent trolls to keep around a small development team to mockup wrappers to whatever IP they buy up to show they wrote it?

I suppose the due-diligence requires to disprove shenanigans wouldn't scale and could easily be gamed by reams and reams of "source code" that actually did nothing described in the patent but was sufficiently complex so as to confuse.

This makes me think your approach to expiring them after a short period of time is the right approach to start with.

I've heard an (apocryphal) story about a bubble-gum brand that protected an otherwise unused trade name by manufacturing and selling one stick of gum under that brand each year.

I completely agree that software patents should expire in 2-3 years. Even a 2 year advantage is an eternity in the software market.

Agreed, shorter time spans are the way to go:

I'd propose an even quicker expiration if the patent is sold/transferred, to curtail hoarding.

In this case the patent is related to PacketVideo's business.

If you have a patent, you MUST have a product or be working on a product using the technology in question or you cannot litigate others using your idea.

That doesn't help in the case where non-innovative patents are used in an anti-competitive manner. Such as the patents Apple is asserting against HTC.

Sure, it may not help in THAT case but it would help with people hoarding patents for the sole purpose of suing others. There isn't going to be a single solution that solves the patent problem. It will be a combination of rules like 1) you must have a product on the market using the patent 2) no business method patents 3) software patents are limited to 2-3 years etc etc

The problem with "must have product in the market" is the same as with the idea of fixing the domain name system by adding a requirement of "use it or lose it".

Who gets to decide what "having a product in the market" means?

It's impossible to generalize this to an objective set of rules. Thus we'd soon be back on square 1 (read: in the court room).

Not really. Firstly because not every invention can so easily be produced. What if I patent something to do with space travel? What if it's something that isn't simply a software paradigm? And if it is just something that can be done in software, why not create some semi-working version to prove it's in production? Or are you going to add that it has to sell? Or be appealing? What if you say you have a version for sale but it's a million dollars a unit?

That law is so simple I'm sure there are even more ways to get around it than there are real inventors who will be shafted by it just because they don't own the facilities to produce it enough to satisfy you.

I don't think this rule is sufficient. Beside the other points people already raised, you still have the issue of people patenting obvious stuff with prior art (like the recent Apple vs HTC case, for example).

I am really curious if an event like this will eventually promote change in the software patent laws

The answer, I have come to realize, is no. It will always be no. Ridiculous examples do not drive change. The only thing that will drive change is a direct attack on John Doe (aka patenting the mRNA polymerase gene and demanding licensing fees from everyone) or collaborative effort that has a direct effect on John Doe- for example, a "strike" by a large number of significant companies, or a "dark knight" who uses patents to completely shut down a very popular company like Apple and thus incite the people.

I am a pessimist, I know, but honestly if you've paid attention these past few years, you'll have noticed that obscene patent stories just keep coming.

If Apple, Microsoft and Oracle managed to drive Android out of the US, that might be enough, but I don't expect that to happen. Even if they won all their lawsuits on the broadest possible terms, some other large patent holder (e.g. Sony, IBM, Nokia, etc.) would have a powerful incentive to pick it up, invoke MAD in the other direction and milk the US Android exclusive for all it was worth.

>I am really curious if an event like this will eventually promote change in the software patent laws.

Me too. I'm thinking that the Apple/Lodsys one has the best chance of making a change, as for Apple so much is at stake.

I don't know how it is over in the US, but here in the UK the whole issue isn't really high profile, even with stakes as high as they are. (IE. HTC not being able to sell mobiles if the patent decision is upheld in December.)

One thing is for sure, I for one would not put any app in a US app store, which is a shame for me and my potential customers.

as a contrast, trademarks are handled in what most people would consider a 'commonsense' way, where it is required that the trademark owner actively use the mark for some period of time and also police that others aren't using or abusing the mark in any way.[1]

Obviously, Trademarks and Patents are two different beasts, but the mechanism for enforcing and protecting trademarks i think is usually along the lines that most people consider fair, which is that the trademark is used to protect an interest for as long as that interest is active (and semi-vigilantly policing its mark) and is using that trademark to protect its own notoriety.

Obviously, trademarks and patents are two different beasts. But (and this is why i'm repeating myself) my point is that you don't have to get rid of these sorts of laws. Trademarks are equally nebulous, but their model seems to me a huge success in the world of intellectual property management whereas (software) patents seem to be a stickier situation altogether.

just my two cents, because again, i don't think we should necessarily get rid of software patents, but i think that thinking of them using this old framework has been problematic in a way that trademarks haven't been (despite similarly ridiculous, yet defensible, use cases, i.e. trademarking 'tiffany blue').

[1]: http://en.wikipedia.org/wiki/Trademark#Maintaining_rights

I don't think patents protect the little guy as much as you think. Yes, in theory it sounds nice and perfect, but I don't think that's how it works in reality.

Look at Zynga for example, how they've ripped off many other smaller developers, and they haven't really "paid" for it. The other commenters are right. Small developers can't afford a big lawsuit either way.

But I think copying each other leads to higher quality products, even if it's done by the big companies. How small companies can stay ahead is with "disruptive innovation". The way disruptive innovation works in general is that the big companies don't want to copy you, because it means dramatic changes to their company and the way they do business.

It also usually means lower revenue (per unit) which they don't like because they think it can't sustain their much higher cost structures. Adopting the new disruptive innovation that is cheaper also means cutting into their own profits from their more expensive products.

Take Intel for example. It's not that they can't build ARM chips or don't realize the potential of ARM chips. It's that they really don't want to do it. They've tried it before and it presented them with a huge conflict of interest and it gave them revenues they didn't think were appropriate for a company like theirs. So they'd rather continue to try building an x86 chip to compete with the ARM chips, which is Atom, even though they keep failing coming up with a good alternative. I don't think Atom will ever be able to keep pace with ARM chips in all 3 of these factors: performance, energy efficiency and price, or by the time they do it, it will be too late.

I was talking to a coworker about software patents, and he said they were terrible. Then I asked about non-software patents, and he thought they were absolutely necessary.

To be honest I don't see the difference. For any given complicated piece of hardware, you can build an equally complicated piece of software. For every simple software patent out there, there is an equally ridiculous non-software patent (e.g., foot-pedal washing sinks)

I think whatever your stance on the issue of patents is, you shouldn't treat software as a special case.

- everyone with money

There's your problem right there.