Peer to Patent
July 16, 2005 11:19 AM   Subscribe

Peer to Patent (PtoP): A Modest Proposal This modest proposal harnesses social reputation and collaborative filtering technology to create a peer review system of scientific experts ruling on innovation. [via beSpacific]
posted by mlis (11 comments total)
 
Interesting, but complete and utter nonsense.

1. Instead of letting one examiner hunt for prior art, the proposal says, let hundreds of experts do the work. Well, the problem with that is that most of the stuff that is patented is not so accessible as the crustless sandwich and those other things you see on the news, but pretty arcane stuff. Often, the only people who really know enough about the technical background are members of a handful of companies who are competing with one another, and thus an honest assessment cannot be expected. Also, companies with thousands of employees (think Microsoft or IBM) could easily crush start-ups or single inventors.

2. Initially, a wide range of input from experts will be sought by the patent office looking for relevant “prior art,”

Right, like those experts don't have anything better to do with their time.

3. In a way, peer review is already happening. Many jurisdictions, like Europe or Japan (but not the US), have an opposition system in place, where an examined patent is laid open for public inspection. Competitors who think that the patent is not worthy then have several months to inspect the patent and lodge an opposition. The patent is then sent back to a panel that hears both views and makes a decision.
Most patents aren't opposed, because it is obvious that they try to protect technology that is irrelevant in practice, for example. But where a monopoly threatens the competition, peer review is taking place.


The patent system is far from perfect; one problem is that opposing/fighting patents is expensive, which gives an edge to the company with the most money. But the proposed social software nonsense offers no improvement (and is completely unimplementable).
posted by sour cream at 11:46 AM on July 16, 2005


What a load of horse manure.
posted by caddis at 1:26 PM on July 16, 2005


Heh. Be spacific. Good one.
posted by mleonard at 2:20 PM on July 16, 2005


What sour cream says. This guy sets out reforming the patent system without knowing how it really works in the US and elsewhere and reinvents the wheel in the process (see sourcream's objection 3).

Objections number 1 and 2 are also quite significant. The problems with external experts is that:

a) They'll generally have a big interest in the patent not being granted.
b) For some cutting edge stuff, the number of qualified experts will be in the dozens, not hundreds, or even down into single digits.
c) Reading and understanding a patent is often both difficult and time-intensive. It takes patent offices thousands of examiners working full-time. Multiplying this effort a thousand-fold, as it is suggested here, would cripple industry. If those experts could be bothered, of course...
d) Moreover, to understand the real scope of a patent it isn't just necessary to understand the technology, but also the very special legalese in which patents are written. An expert in the technology could well accept a patent thinking it has a much more restrictive meaning than what the patent attorneys have actually managed to sneak into it. (When I see how many "experts" still interprete patents by their titles, instead of reading the granted patent claims, I despair).

Not that I don't see some interesting points. IMHO, one of the biggest issues of patent systems everywhere is that of, even knowing the prior art, separating the obvious from the inventive. To do this, patent systems and courts use tests, based on previous decisions of courts, that can be contrived and artificial in the real world. Some reliance on a real "jury of peers" could help there.
posted by Skeptic at 10:42 AM on July 17, 2005


A better proposal: Patents do not exist to protect creativity, ideas, pure research, or even applied research. Patents only exist to (1) encourage people to "bring products to market" by (2) protecting the capital involved involved in building a factories, get FDA approval, etc. We can draw a few conclusions from this:

(1) Patents should only apply to a product which can be put on the market. So your patent should actually say how your going to make it.

(2) Patents should ideally only apply to products can not be profitable without them. So your patent should say how much its going to cost to make it, and where your going to get that money if you get the patent.

Now patent approval process is trivial: Your patent application is published, and you wait say one year. In this year, anyone is free to contest the patent OR bring the product to market first. If your beaten to the punch, no one ever gets the patent. If no one does it without the patent, then you've proved it needs a patent, and you get your patent.

BTW, nothing like FDA approval existed 200 years ago, so it requires a more subtle compromise, but the above should work when safety testing is not the major issue.
posted by jeffburdges at 4:57 AM on July 18, 2005


A even simpler solution: First, eliminate personal income tax, replacing it with a corporate income tax, which should be mostly equivalent (but also help prevent monopolies). Next, change patents so that they don't restrict who can make the product, but instead they just grant the first guy to make it a tax break on resulting revenue (for say 10ish years). Now companies can just fight with the IRS over how much of their product is covered under the patent, but everyone else gets to use the idea.
posted by jeffburdges at 5:02 AM on July 18, 2005


jeffburdges, quite aside from whether those are good ideas generally (I'll admit, they're at least kind of interesting), are you saying you actually think that's a simple solution? Consider the disruption in shifting the tax basis, alone: There's no way in hell that companies wouldn't be shifting their new tax burden to their emplyees, and the mechanics of that shift would be fantastically expensive. (Ever been part of a PeopleSoft rollout? My former employer spennt something in the range of a hundred million, and still didn't have a usable system at the end of the day.)

And that's just one small administrative component.

Put another way: Whenever you simplify, you have to account for the cost of porting the data and mapping the business processes.
posted by lodurr at 7:52 AM on July 18, 2005


[self /]: ... fantastically expensive.

On reflection, it could be an even bigger boost to the tech sector than Y2K. I say, let's go for it.
posted by lodurr at 7:58 AM on July 18, 2005


jeffburdges Patents do not exist to protect creativity, ideas, pure research, or even applied research. Patents only exist to (1) encourage people to "bring products to market" by (2) protecting the capital involved involved in building a factories, get FDA approval, etc.

Well, that is not true, as a quick look at, for instance, Art. 1, Section 8, Clause 8 of the US Constitution shows:

To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries;

There are similar statements of intent in other countries' patent legislation. The purpose of the patent system is to reward inventors who disclose their inventions. No more, and no less, and whoever says otherwise is being utterly disingenous.

This is also quite important, because of simple division of labour: some people are good at inventing things, whereas other people are good at bringing them to market. The patent system, with its licensing agreements, is actually not bad at bringing them together and giving incentives to both. Unlike your "first to market" ideas who would reward only the latter, much to the detriment of actual invention.

This said:

(1) Patents should only apply to a product which can be put on the market. So your patent should actually say how your going to make it.

This is actually an already existing requirement in every patent system: inventions and how to realise them should be fully disclosed. So, nothing new there. The trouble is that actually ensuring this requirement is fulfilled is far from easy.
posted by Skeptic at 3:29 PM on July 18, 2005


No, first to market would not "reward" the guy who beat you to get your product on the market, it would simply judge your product to be trivial if someone would do so.

The relevant words in your quote are "science" and "inventors," but these two notions have relatively little to do with one another today. The restriction to "products" makes it clear that inventor, not science, is the correct word. Inventors are still be protected under my proposal, just less so. Anyway, patents simply aren't paying for research, even drug companies just get the federal gov. to pay for their research.

Still, now that you mention it, no reason to disclose the manufacturing details for a whole year. Divide the patent into two parts, descriptive and constructive, and hold back the constructive part until late in the year. If the implementation is obvious from the description, it clearly does not deserve a patent.

As for enforcing that patents cover an actual product, not its not that hard really. There is just a lot of money spent on abusing the notion of product.
posted by jeffburdges at 2:03 AM on July 19, 2005


No, first to market would not "reward" the guy who beat you to get your product on the market, it would simply judge your product to be trivial if someone would do so.

For what reason? That something is easy to copy does not mean it is trivial!

The relevant words in your quote are "science" and "inventors,"

Wrong again. If one word is relevant it is "Arts", in one sense of the word quite common at the time the US Constitution is written, and still present in the words "artisan" and "artifact" as well as in the patent-speak expressions "prior art" and "state of the art": technology.

but these two notions have relatively little to do with one another today.

It would be nice if you had some factual support for such fancy assertions.

The restriction to "products" makes it clear that inventor, not science, is the correct word.

Clear to you, perhaps. I fail to see any connection.

Inventors are still be protected under my proposal, just less so.

I don't know how.

Anyway, patents simply aren't paying for research, even drug companies just get the federal gov. to pay for their research.

More unsupported rubbish. Even this extremely anti-Big Pharma report admits that big American pharmaceutical companies were investing 12% of their income in R&D in 2000.

Still, now that you mention it, no reason to disclose the manufacturing details for a whole year. Divide the patent into two parts, descriptive and constructive, and hold back the constructive part until late in the year. If the implementation is obvious from the description, it clearly does not deserve a patent.

Because you say so? At least in mechanics "constructive" and "descriptive" are very much the same. But even if by "descriptive" you meant a statement of what the invention is supposed to achieve, the problem it solves, rather than what the invention is, this would mean very little. Many far from trivial inventions have been made almost simultaneously by several inventors. In fact, rather than indication that those inventions were trivial, it often means that they were important, since a lot of people had an interest in the matter.

As for enforcing that patents cover an actual product, not its not that hard really. There is just a lot of money spent on abusing the notion of product.

I'm afraid that you have completely misunderstood the matter. I did not say that enforcing that patents cover an actual product is difficult, what I said is that it is difficult to enforce that the invention and its implementation are fully disclosed in the patent.

You see, you have a completely misguided fixation on the concept of "product". An invention is not necessarily a product. In chemistry, metalworking, and many other fields, a new method of manufacturing a known product can be just as important an invention as any new product. Think of aluminium, for instance: even if it is one of the most common elements in the Earth's crust, until near the end of the XIX century, producing it was so expensive that it was used for the tableware of emperors. Then a new aluminium production method was invented that turned it into the cheap commodity it is today. Wasn't that a very real, very useful invention?
That is why it has always been possible to patent both products and processes.
posted by Skeptic at 1:34 PM on July 19, 2005


« Older scrap it   |   Coming Apocalypse? Newer »


This thread has been archived and is closed to new comments