Trend: More Software Patents

The number of software patents issued this year increased to 22500, while a new type of hardware predator has arisen: companies that increase their revenue by patent royalties. Indeed some companies focus on making patents on products they will never make, just to get IP-revenue. Neither is extortion out of the question ("Come here laddy, prove to me your not violating any of my 100 patents !"). Software engineers don't like it, and in one company had to be threatened to play the lawyers' game. Even though few companies currently use software patents offensively, at a $20,000 cost per patent, they could hurt Free Software badly, should they follow the hardware trend and do so.

Comments

One Possible Solution

[Greyfox]

How about forming a nonprofit group whose sole purpose is to go through the patent portfolios of assorted companies that are filing for these nonsensical patents? This group would research prior art and anything else that could get each patent thrown out in court and post their findings on a web site that would serve as an information clearinghouse. Kind of like a discordian IBM patent database.

It should be pretty easy to at least partially automate this process -- IBM is generously supplying the patents in electronic form and it should be possible to build a program that cross references the patent for relevant web pages. Just have your debunker check out a patent to work on and he gets a nice page of info already researched by the computer. Hopefully a group like this would also encourage the PTO to improve their own process of granting patents.

Just a thought...

Change the rules

[ajs]

Just create a $10,000 fine for "spurious patents", and define spurious just
as vaguely as the USPTO defines valid software patents. This will put
the breaks on most companies (if, for example 1/3 of Microsoft's software
patents were found to be invalid, that would probably be enough to
cancel out profit from Office for a year!)

The other way to go, that a lot of people cite is just limiting the life
of a software patent to 3 years. I'd sign on to this, IF there were
a reasonable way to speed up the approval process.

One way to speed it up might be to open it up. Let everyone see every
submitted patent that has been past the first "does it fit the
required format, and did they pay up" pass. Do it over the Web, and
allow a Slashdot-like feedback. Only patents that don't get a bozo-alert
from the masses get sent on to the next stage (internal USPTO technical
review). Thus the next 200 people to submit "talking online" and
"window-shopping online" will be bounced in less time than it takes
to Slashdot a personal web-server into the ground.

Fight back

[Chris Johnson]

Cellular is a fairly unimportant little cellular automata program.
Staccato is a reminder program, which might have some ideas relating to intelligent input parsing somebody is trying to patent. It sets things up so entries in the info file are very easily done, with the 'API' extremely easy to master. I'm sure someone would try to patent that.
Sitebot is a particular method of keeping data files as plain text with a couple of easily added headers, and 'compiling' that into a website which can then be uploaded. If anybody means to patent a narrowly defined method for writing plaintext and having it read and turned into a web site with the same structure as the plaintext files and folders, Sitebot is prior art.
ROTSOS (Return of the Son of Spacewar) is the best yet, being a radically different approach to game engine design. It offers literally the ability to produce game 'maps' equivalent to data files billions of gigs in size, in fact the ability to have billions of worlds each with 'maps' (not all of which will be distinct, but for all practical purposes...). It requires that game map creation be an exploratory process rather than a creative process, a major innovation in map design IMHO as the person who came up with it after reading lots of stuff on AI and artificial life. Took some years to work out, and naturally I've had to produce flashy demos (mostly movies, others to come) to illustrate what's being done here.

What do all these software products have in common, from the trivial to the actually innovative?
They are _all_ Free Software under the GNU GPL. That is including ROTSOS, and I have every expectation that somebody else with ship a GPLed game before I can get one together. I understand that and approve of it. I also understand that I'm going to stay poor and won't get diddly from all this.

Then why on earth am I doing it?
Because I'm just another soldier in a different sort of war. This patent stuff is deadly serious, but it's not fought with guns (unless they are patentable ;) ). It's fought with ideas, and sacrifices have to be made. If intellectual property is not to be become a ball and chain, if people are to retain the ability to work with their minds and retain control over their own ideas and clever inventions, it seems the only safe haven anymore is the hardcore free (libre) software side- the determination to produce ideas and add them to the shared hoard. If these ideas are in use and known to be in existence, matching ideas cannot be patented. If the ideas are unsung and never seen by anybody, then lawyers will have a relatively easy time raising doubt that a matching idea came first. Publicity is the friend of free software, the handmaiden of 'prior art'.

So.... STEAL MY SOFTWARE!!!! That's right- go grovel through its ugly depths for any ideas that might make your open source project take off! Grab anything you want! Be grateful or not, say nice things about what neat ideas I have or not- the only requirement is that it stay GPL. Take all the credit for stuff that I came up with, while releasing it as GPL! Because as long as somebody gets publicity for a GPLed idea or algorithm or program, as long as that idea is obviously prior art and not ripe for a patent, that means I get to keep using it. And if the ideas languish in obscurity, it's all the more likely that some clown will patent some broad notion, hire better lawyers and enjoin me from ever using the idea that was mine in the first place. I'm not kidding. Wittingly or not, this is war now.

Write GPLed software (that being the most hardcore of the licenses)! Get glaring publicity! Anybody who can, _please_ make sure as many ideas (broad or specific) are within the camp of 'free software, prior art' as possible. Because it's a real problem, a serious danger, and these people trying to fight it by staking out defensive patents are only compounding the problem.

Time to choose sides!

The jug with two handles

[jd]

Ever read that children's story? Well, it applies here, too.

The large software companies kep Free Software down, by pointing out that there's nobody to sue if something goes wrong.

This puts them in a bit of a bind. If they sue over patent infringement, then (win or lose), they let the genie out the bottle. They show to their customers that, indeed, free software -can- be sued, and is therefore a liable source of software - something THEIR lawyers insist on.

Do that, and they lose customers to OSS alternatives. BIG time. And there's no way in hell they are going to take that chance.

On the other hand, if they don't sue, they risk wasting their cash, they risk losing their patents and they risk losing their image. Can you imagine what the press would make of it? "MegaCorp, Inc, handed their multi-trillion dollar patent file to Dweebs, Inc, a bunch of long-haired computer nerds who's only claim to fame is being any good at programming."

The computer companies would get -SLAUGHTERED- by the press, AND their shareholders, if they simply bow down.

So what can they do? Nothing. There is no answer, at least, not in the direction they're going. Whatever their response, if an OSS group infringe their patents, they are going to be cut to pieces. Their only hope is that we don't.

I don't know...

[um... Lucas]

I think I'm pretty much for patents in most cases. The goal of most companies is to make money and keep a competitive edge. That means spending lots of money on R&D, and you don't want to make a discovery, get a product to market and then find that your largest competitor copying your work. That, i think, would cause the death of innovation in so many industries.

I think that there should be different classifications on patents, if that would be at all feasible. For instance, computer-related patents should only be allowed to be enforceable for 5-7 years (the 3 year life of any given computer, plus a couple years just to be generous)... That'd be enough, I think that a company could come up with something trully ingenious and make their money from it, but they'ed also have to work their buts off on their next product rather than going "Okay, now we've got 20 years of revenues from licensees"...

Patents on drugs are more difficult. We need drugs to live comfortably (and i'm not even talking about the recreational ones)... Should a company be allowed to patent it's discoveries? Yes. Should they be allowed to recoup their investments? Yes. If they couldn't, then they'ed all make the next diet drug, because those are a lot surer to bring back profits for the shareholders...

Perhaps the gov't should contract with the Pharmaceutical companies to develop drugs they feel they need developed. They could say, we need a new AIDS drug, start your bidding. Then the gov't would pay all associated costs for 5 or 7 years to develop the drug, and give the company 15% of the income derived from the drug.

That way, companies would still get their due, and consumers wouldn't get nearly as screwed as can be the case these days (ahem... UNISYS)

Interesting implications for closed-source, though

[konstant]

IANAL, but this article seems to be implying that the burden of proof rests upon the defendant in patent-violation cases. If this is the case, then it creates a curious set of conditions in relation to the openness of the product code base. Software that is open, and thus can be inspected freely, is far easier to defend against charges such as these than is the case with closed software. If the the code is kept proprietary, then the only means by which a developer can prove innocence is to allow a court-supervised inspection of the code. This causes all sorts of legal troubles, because typically the suing corporation has a vested interest in learning how the proprietary code works.

You may remember Microsoft's tiptoeing with the Caldera case (I think it was Caldera). They had to explain at length to the judge that opening their Windows code as a defense exhibit would endanger their business model. The code could only be examined after much legal wrangling and numerous NDAs had been signed. No doubt this is pretty costly - legal work doesn't come cheap.

By contrast, since the code of open-source projects is, by definition, open, we might expect fewer spurious suits of this kind levied against FSF/OSS products. Why press your luck suing somebody when you know quite plainly they have not violated your patent and that it would cost them nothing (apart from lawyer overhead) to demonstrate that fact. It's a financially losing proposition. The incentive to settle and let the patent pirate laugh its way to the bank is far smaller.

By the same token though, if you do open your source, you'd better be damn sure you really aren't violating any proprietary code.


-konstant