Patent Office Program To Speed Computer Tech

coondoggie writes "Looking to address critics, the US Patent and Trademark Office this week is starting a program to speed up and improve the review of computer hardware and software technologies. The agency is set to launch a peer-review pilot project that will give technical experts in computer technology, for the first time, the opportunity to submit technical reports relevant to the claims of a published patent application before an examiner reviews it. The idea is to get as much knowledge about a particular claim in front of an examiner as quickly as possible so they can make a decision faster, the agency said. IBM, Microsoft, General Electric, Hewlett-Packard, CA, and Red Hat have already agreed to review some software patent applications for the one-year community review project. Intel, Sun, Oracle, Yahoo, and others are also part of the project. The pilot is a joint initiative with the Community Patent Review Project, organized by the New York Law School's Institute for Information and Policy.

Comments

What a great idea!

[hasbeard]

I just wish they would make it retroactive to all the other patents currently awarded.

Re:What a great idea!

[theantipop]

You can always request reexamination of an issued patent.

Foxes and henhouses

[Watson Ladd]

Compeditors have more to gain from a patent portfolio+cross licensing agreement then they do from invalidated patents. Unless we have public review or honest people reviewing this won't work.

Re:Foxes and henhouses

[theantipop]

Sign up to be a peer reviewer. I can't seem to find what kind of requirements exist to be able to participate in this, but it couldn't hurt to register if you feel this is important.

Re:Foxes and henhouses

[eggnoglatte]

What on earth are you talking about? Patent applications are already published on the USPTO website. The whole idea behind the patenting process is that the full method needs to be disclosed at a level suficient for reproduction. If you want a trade secret then don't patent waht you are doing!

Suggested Improvement

[Zashi]

Realize that software is not a patentable innovation.

The use of patents has seriously gotten ridiculous and has made me lose faith in the US Patent Office.

Re:Suggested Improvement

[mr_death]

Realize that software is not a patentable innovation.

At least in the US, the courts have ruled differently. Imagine the absurdity of Ron Rivest being granted a patent on the hardware version of RSA, but not the software version. Both are the same truly innovative system, yet there are those that argue that the software version is somehow less worthy of protection.

Re:Suggested Improvement

[Daniel Dvorkin]

Imagine the absurdity of Ron Rivest being granted a patent on the hardware version of RSA, but not the software version. Both are the same truly innovative system, yet there are those that argue that the software version is somehow less worthy of protection.

Because one is a mathematical expression, and the other is a physical device. Why is this difficult to understand?

Re:Suggested Improvement

[mr_death]

Because one is a mathematical expression, and the other is a physical device. Why is this difficult to understand?

I must disagree. One is (I'm guessing the preferred hardware embodiment) an asic, where the RSA functionality is enabled by the function blocks, their connections and topologies; one is general purpose hardware where the RSA functionality is enabled by the software. Both devices give the same functionality; both are enabled by the directions of a smart person. Why is only one deserving of a patent? Or, to phrase it another way, why is the world free to copy one but not the other?

Re:Suggested Improvement

[adolf]

OK, I'm imagining it: A world where software implementations of RSA were never covered by a patent.

And what I see is a world, circa 1996, in which early adoption of SSL was not hindered a patent on RSA. In this world, telnet became deprecated much earlier on, free browsers included strong authentication and encryption, and it wasn't illegal to run Apache with mod-ssl in the US.

It doesn't sound like such a bad place.

Now, of course, what really happened is that RSA was only successfully patented in the US. The rest of the world was free to do whatever they felt like, as outlined above.

Yay us!

Re:Suggested Improvement

[vux984]

Why is only one deserving of a patent?

IMO neither is deserving of a patent. The hardware implementation is hardly innovative.

Or, to phrase it another way, why is the world free to copy one but not the other?

Its better to work with examples that make sense, like the difference between a working hard disc, and someones obersvation that hey you could use magnets to change state of something. Its not that the latter wasn't innovative thinking when it was thought up, but its not a patentable *invention*. There are countless different ways an idea might be 'implemented' in an invention. But if you simply patent the abstract idea for the process itself, then it covers every possible implementation.

And when the 'idea' itself is simply a mathematical equation, patenting leads to near absurdities. RSA encryption is really simple to implement; anyone who is given the algorithm, or even just the underlying mathematics and rough explanation of technique can do it. I'm not saying RSA wasn't an innovative idea, merely that it didn't lead to or require any innovative inventions to make real.

The RSA idea itself was the hard part. Unfortunately, the RSA idea is really no different than the Pythagorean theorem, or the forumula for computing the volume of an oval cylinder, its pure math. Do we really want to live in a world where the first person to solve an equation 'owns' the solution. And once solved, no one else may find the area of a triangle or compute the volume of a cylinder without licensing fees, even if the problem is easy enough for a high schooler to solve and merely no one had been asked to solve it?

Should the equation: sum(1..n) = (n^2-n)/2 be something someone can own, just because it requires some creativity to prove it?

As ideas not inventions become patentable innovation and creativity are stopped in their tracks because truly innovative inventions are limited in scope -- so you can identify them, license them, or take a different approach when you prefer to step around them. But abstract ideas? rough processes or techniques?... its impossible not to borrow from them, reuse them, derive from them, even completely innocently.

Claims of misuse?

[TwistedEvo]

How long until we start seeing reports of rejected patents, that are later submitted by the big companies involved in the peer review?

How about...

[Consul]

"from the foxes-guarding-the-henhouse dept."

That's basically what this is.

Should some study be done first

[pembo13]

as to if software patents are good at all? I would like to believe that there is enough money involved that finding out whether or not any effort should be spent on them would be a necessary thing. I myself aren't convinced that software patents are useful, but I don't claim to be an expert - some evidence would be nice however.

Re:Should some study be done first

[Kyojin]

Agreed. The patent system was intended to promote innovation by ensuring that individuals and companies would be able to recoup the costs involved in inventing a new product by giving the individual or company a monopoly on that product for a set period of time. In the pharmaceutical industry, billions of dollars are spent on new drugs, most of which never work, let alone go to market. Unless the pharmaceutical companies were sure they could recoup these billions of dollars, there would be no incentive to invent any new drugs.

Software is different, and the patent office has made some attempt to recognise this: You cannot patent an algorithm, but it gets disguised as "a system and method" and the patent is awarded. The costs of "inventing" the things that software companies have patented in the past is next to non-existent. One-Click shopping? I bet that cost a few billion dollars.

If it can be truly argued that these companies spent a large sum of money on software innovation, (note NOT programming, every company does that), then perhaps a patent could be awarded. But these sums of money are likely to be tiny. Perhaps the patent could be awarded for a shorter period of time, perhaps 3 years? Even Debian releases new products at least that often. Or 5 years so that Microsoft can milk us for longer?

Finally a lot of what gets passed of as patentable should come under copyright law. A user-interface is not an innovative means of performing an action. It is a cloudy picture painted over what you're actually doing. Indeed, by publishing a product with a particular user interface you automatically own copyright on that user interface and can sue those who copy you. There is no need to bog down the patent system with something that is already protected and is not even innovative in the first place.

Re:Should some study be done first

[MillionthMonkey]

Unless the pharmaceutical companies were sure they could recoup these billions of dollars, there would be no incentive to invent any new drugs.

I don't even like your Patents don't really serve the public interest here either, because we don't really need pharmaceutical companies in the first place. One sign that our kleptocracy has completely warped our thinking is this strange assumption everyone makes: if drug companies don't make lots of profits inventing new drugs, nobody will have any incentive to invent drugs.

We really don't need pharmaceutical companies. The public wants access to a wide range of cheap effective medicines. So we have a natural incentive to invent new drugs because we keep getting sick and dying. There are plenty of ways to solve the problem. A straightforward one would be to create public drug discovery laboratories, fund university labs, and pay for scientists to find the drugs. That's a "tax and spend" solution. We decided on a solution where we replace our natural incentive for better drugs with Pfizer's incentive to get rich selling them to us.

That works to the extent your desire for better drugs remains compatible with the perogatives of a for-profit corporation. Sometimes it isn't. A company makes more money by developing treatments as opposed to cures. It saves money by making copycats of drugs already shown to be profitable, like penis pills. They concentrate their efforts on diseases with the widest markets, and don't do much research into rare diseases. And of course they spend a lot of time looking into what they should do if they want to pull even more money out of your pocket. My wife and I are still young but we each have our own chronic neurological problem. Just the copays on these prescriptions are exploding. Ours are running about $150-200/month. And the trail of patents and monopoly rights left behind by this process is undesirable in and of itself, even if getting them did provide the company's incentive. For one thing, the patents rise into the atmosphere and do not expire for years and years. The air becomes clogged with patents and they accumulate into a dark cloud that casts shadows and disincentives upon drug research below- no matter who is doing it. So our current path isn't sustainable.

There is plenty of incentive to invest in new drugs as long as people are sick and dying. Even if a private company isn't interested, there are enough people who do that research, and sufficient public interest in getting it done, to ensure that it will get done, even if nobody is getting rich running commercials for me-too penis pills. Only patents could screw it up.

Re:Submarine Patents

[mr_death]

First-to-file sounds great at first blush, but it ends up screwing the small, innovative company. Imagine a discussion between Startup LLC and Microsoft. While discussing a partnership agreement, Startup mentions some tech they're working on. One of MS's patent lawyers is in the meeting, writes up what Startup talked about, and submits a patent app. Startup LLC is now screwed.

Give me first-to-invent any day of the week.

Am I Reading This Right?

[aldheorte]

FTA

"IBM, Microsoft, General Electric, Hewlett-Packard, CA and Red Hat have already agreed to review some software patent applications for the one-year community review project."

Wait... so large companies with lots of existing patents have volunteered to review new patents in the field to try and help the examiner dismiss them? Was not the patent system set up in part to encourage small inventors and entrepreneurs? Could this be an even more obvious conflict of interest?

"Technical experts in the computer arts registering with the CPRP website will review and submit information for up to 250 published patent applications, with no mare than 15 patens being accepted from one applicant/company at a time, the USPPTO said."

250? Drop in the bucket? Only 15 at a time from one company sounds like convenient plausible deniability for organizations that file hundreds per year.

"Consent will be obtained from all applicants whose applications are volunteered and selected for this pilot... Some applicants today can wait up to four years for a first response on software applications. The idea with the pilot is to shorten that wait considerably."

So you can either go to the end of the line or get to run the gauntlet of the entrenched companies trying to help dismiss your patent?

Re:Submarine Patents

[roseanne]

First to file doesn't mean you can steal ideas. You still have to show independent invention. If you're a startup discussing any plans with MS, you really should get some papers signed, if only to leave a paper trail to show that MS was aware of your tech, thereby damaging their ability to claim independent invention.

Re:Submarine Patents

[edwdig]

What meeting between a startup and Microsoft doesn't result in the startup getting screwed?

You're looking at this from the wrong angle.

[mrchaotica]

The question is not "why should the software be less patentable than the hardware?" but rather, "why should the hardware be more patentable than the software?" They're both just implementations of a mathematical algorithm; neither should be patentable!

I'll do a "study!"

[mrchaotica]

Studies for a bit...

Nope, software is still just math, so it's not patentable. And it's already covered by copyright, so it's doubly not patentable!

What more evidence do you need?

No it doesn't

[Flying pig]

Under first to file in the UK you submit a low-cost provisional application the moment you have the idea. Now you can go and discuss it with VCs etc. and you have established your priority. Only if it then looks like a worthwhile invention, or when you have the funding, do you have to pay the full application costs. What's more, if you decide not to file, the provisional application then lapses to prior art so nobody can come along afterwards and patent it.

The US system has now changed, but under the old system Startup LLC had to keep detailed lab notebooks and probably have them witnessed by an attorney in the US, while in the UK they just had to write up the idea and submit a provisional application. Which is better for a small company?

The US system originated when the US had terrible communications and was designed to deal with the small inventor who had been making something for years in Outer Fencepost, Wyoming, and then Bad Company, NY came along with the railroad and copied his idea. Nowadays, this is not an issue.

Computer Program?

[tezza]

if("Microsoft".equals(patent.client.name) ) {
      approve(); return;
}
else if ("Google".equals(patent.client.name)) {
      approve(); return;
}
else if ("IBM".equals(patent.client.name)) {
      approve(); return;
}
else {
      inspect(); return;
}