Asymmetrical Information

I'm dubious. In the computer/telecom field, it's very clear that patents are totally destructive, and my opinion is that the entire system should be abolished.

At the very least, the original definition of a patent that pretty much required a _material_ device or process as its subject matter should not have been directly expanded to software.

I also work in the computer field, but don't think patents should be abolished. The problem with technology patents is that there is a general lack of knowledge in the USPTO about what is and is not prior art, which allows people to submit dubious patents on technologies that should not be patentable. Sorry for going off topic, but had to respond to the first commenter.

Patents are just a bit too short for pharma (25 years would be defendeable). A better alternative, of course, would be to simplify FDA requirements and keep it at 20 years. Food industry is similar (FDA approval needed).

They are way too long for almost anything else. In computers/telecom, 2 years would be best.

The notion that the whole patent system is "destructive" and needs to be abolished is...shall we say...unconventional. A teesy-weensy bit of rhetorical support might be called for.

I don't see anything wrong with software patents in and of themselves, but it does annoy me that the disclosure required does not include the source code.

That said, I wonder if one improvement to the pharma patent system would be to grant patents to the first entity that finds a use for a compound rather than just creates it. As it stands today, an academic chemist could create a compound that, many years later, turns out to be a miracle cure for breast cancer. The fact that the chemist created it just because he thought it looked nifty or involved some clever reaction won't stop the PTO from denying the patent to whomever does the medical research that makes it useful.

There are such things as "use patents", and the topic occasionally comes up in the drug industry. If there's no way around it, one company might end up claiming a new utility for another company's compound. Then they're at an impasse - one company owns the chemical matter, and one owns what can be done with it. Generally they come to terms at that point.

But this is rare. Usually we can find a way to break the first chemical matter patent, coming up with something that will have the same use but is free and clear for ownership.

Jane,

I entirely agree: patents have about the right lifespan to accomplish the goal of stimulating innovation: long enough to reward innovators, but not so long as to constitute a first-past-the-bar-perpetual-annuity.

Copyright is obviously far too long: 10 years would be more than enough time for authors to capture the majority of benefits, and short enough to allow a lot of creative re-use.

I disagree with some commentors about software patents: it's far from clear to me that "In the computer/telecom field, it's very clear that patents are totally destructive". Sure, that's the consensus on slashdot, but I've never seen anyone try to make the case in a coherent argument.

I'm a strictly amateur economist, so I guess that was my cue.

It seems to me that lengthy patents/copyright are like huge CEO salaries. The point is not so much to get more effort out of current inventors/CEOs -- which, as Lowe points out, goes against human nature; we are all inclined to rest on our laurels -- as to get younger inventors and middle-level managers to try harder, in order to get on the gravy train.

Looked at that way, there is nothing particularly irrational about copyright that extends beyond the life of the author. It arguably increases the incentive for the rising generation even though the author himself is beyond the reach of any incentive.

Of course, to recognize that copyright beyond the author's lifetime does have benefits is not to show that those benefits are worth the cost. I'm inclined to think that patents are about right, that copyright should be longer-lived than patent protection, but that current copyright is much too long-lived. For example, I find it distressing that many historical photographs and writings from the Second World War are still under copyright. So, on different grounds, I find myself pretty much agreeing with you.

I'm not disputing that extending benefits beyond the author's life could have benefits . . . authors care about their children. But the idea that retroactively extending past copyrights, as well as current ones, for a period that would require authors and corporate managers to care an inordinate amount about great-grandchildren, and successors, who may not even be born when the author/manager dies, is ludicrous.

Contra anon (9:08), I'm in the telecom field and I think patents are a good idea. I also concur with Nick that the prior art review process is horrible (as I have friends in the patent office, I'm implicitly criticizing my friends here) as is the ability to review technical claims.

All of the patent horror stories I know of (e.g., Blackberry) all trace back to a lousy review process.

I am open to the idea that patents in telecom should be of shorter duration and less expansive in scope (that's largely a function of the review process). I also think the IEEE standardization process has become quite the mess with companies working to direct communications standards to adopt algorithms that require the use of their patents.

I think it depends to some extent on the patent. I would support much longer patent lives on new antibiotics. Antibiotics lose effectiveness the more they are used, so there is social benefit in limiting their use. In that case, the reduced production that comes from a patent monopoly is actually a good thing.

You make a good point that patents on pharmaceuticals are effectively shorter than other patents because part of their patent life is wasted in the FDA approval process. It might be a good idea to begin patent lives from the date when the drug is first legally permitted in the market. Drug companies have plenty of incentive to get their drugs to market early anyway. The current system provides a little too much incentive to cut corners and creates more uncertainty about how long the effective patent life will be.

1. I've heard a market based patent approval process suggested. They put up a pending patent and if you can demonstrate prior art within, say, a 3 month period you get paid x dollars.

2. Regarding antibiotics, could anyone suggest a good IP protection model for bacteriophage therapy which wouldn't slow innovation? After all, they're easily stolen. They're naturally occuring, so couldn't be patented under the current law (unless altered re: chakravarty). But we need to support their development to deal with the current antibiotic crisis.

3. With copyright extensions; I think the current is way too long of course. At the very least require constant renewal of copyright, with a high price attached. While I don't think this is the best idea, it would represent a political compromise with the large, politically powerful IP holders like Disney. For everyone else, if you didn't have an up to date copyright then your work would be considered public domain.

Pharma patents seem just about right to me, particularly since they lose a huge amount of their run time to FDA trials

Patent term adjustment now allows companies to extend their patents up to 5 years to compensate for time lost in the regulatory process. See: http://www.fda.gov/cder/about/smallbiz/patent_term.htm

Rob: What is being patented is not the source code, but the (theoretically novel) idea it implements, which is why the source code is not required. (Indeed, there are a nigh-infinte number of ways to implement, if subtle differences and different lanaguages count as "different" ways...)

For that matter, when it's an innovative interface, it's an odd thing to ask for "the" source code, no? (In an event driven model, for instance, we'd end up with a number of event handlers and a long description of how they're meant to interact. Probably easier to do WITHOUT the source, honestly.) Is it the specific code that implements that interface in a specific API? Would pseudocode suffice?

I don't, as a programmer, see the point in requiring source code, so long as the description of the patented technique/interface/algorithm is sufficiently clear to describe it. At most, pseudocode would be enough; while I haven't looked at the content of any software patents (not being interested in what they look like), I wouldn't be surprised if that was actually done.

(Much like patents for physical devices don't need to specify every material used or every other specification needed to manufacture, but simply indicate how the thing works schematically.)

Having worked extensively with an IP consulting firm that performs reverse engineering, patent review, IP marketing, etc., I also take the position that patents -- notably software patents -- are ultimately destructive to the high tech industry. The most basic problem of all is that the high tech scene is a polar opposite to the drug industry: the patent often far outlives the serviceable life of the invention.

Likewise, as others noted, there is a poor review process in the patent office. Software and business process patents make this problem worse. IMO software is suitably covered by copyright; a patent just hands the king a gun. Meanwhile business process patents are commonly applied to processes so broad and obvious in the art (e.g. 'one-click ordering') that they don't even pass the laugh test, but somehow they do pass the patent office review. To prove otherwise, competing parties then have to pay court costs.

Companies and savvy individuals, in turn, have been exploiting these weaknesses and making collosal IP land grabs, then converting that land into gold via the fiat of the courts. And the way some of these patent claims can be twisted to at least make a "mine, not yours" run in the courts (whether successful or not) resemble playground arguments writ large.

Lawyers and well-placed consultants are making a killing off this trend, but it isn't helping spur innovation. And good luck reforming the patent office directly; unless we intend to offer reviewers a $200k/y salary and require them to undergo extensive 'state of the industry' training every 12-18 months, it will perennially be understaffed, in majority by underincentivized underachievers.

Sigivald,

The point of patents is the grant of a temporary monopoly in exchange for sufficient disclosure to enable someone who is "faimilar with the art" who comes along after the patent expires to build your product based on your description.

Now, I'm a grossly incompetent programmer so it may be that software patents already do that for people who actually know what they're doing, in which case I have no complaint at all. True, you need not specify every material in your new engine design. But if your new engine depends critically on some particular material or specification not obvious to engine builders, you have to say so. When I have looked at software patents, what I see in a is roughtly the equivalent of a mechanical patent which says "I have built a black box which does something useful" followed by a description of the buttons and levers on the outside of the box. No peek inside. It would not be tolerated for mechanical patents.

(Or maybe it would, if the buttons ("interface") were the point. But if the software is doing some thing useful--taking fourier transforms, or monitoring a factory for bottlenecks--then just saying "I claim factory monitoring" ought not to be enough without explaining in detail how.)

As for "the" source code, the current rule is that the best known implementation of any given device or process must be disclosed. Whatever reasonable method allows others to copy you is good enough.

Again, perhaps real programmers can make sense of the patents and its just me who can't.

"I think it depends to some extent on the patent. I would support much longer patent lives on new antibiotics. Antibiotics lose effectiveness the more they are used, so there is social benefit in limiting their use."

There's a social benefit in getting new ones to market as quickly as possible to replace the old ones that lose their effectiveness.

"Lawyers and well-placed consultants are making a killing off this trend, but it isn't helping spur innovation. And good luck reforming the patent office directly; unless we intend to offer reviewers a $200k/y salary and require them to undergo extensive 'state of the industry' training every 12-18 months, it will perennially be understaffed, in majority by underincentivized underachievers."

Actually, that's not a bad idea. Given the untold trillions of dollars' difference that patent reviewers can make to the future technological advancement and wealth of our country, I'm all for ponying up $200K a piece to get talented ones.

I actually have a cheap technical fix for this! Call it "Progessive Patent Extension Fees". Allow patents to run as is for some small number of years, say seven. Further allow patents to be extended on a year-by-year basis, upon the payment of a extension fee, where the fee doubles for each year. If adding an 8th year to a patent costs $1000, adding a 20th year to the patent would cost ~$4M, and a 30th year a whopping $4B. All parameters are of course configurable. If you don't like how those numbers ramp, just change the exponent. There's a lot of advantages to this:

• It gives preferential treatment to valuable inventions, while clearing out the deadbrush of valueless ones quicker.
• It penalizes patent squatting, and probably makes squatting on a large portfolio of dubious patents economically unviable. (That alone would help out the computer/telecom industry a lot. It's not so much the well known patents that are hindering innovation, it's the plethora of landmine patents lurking in wait.)
• It captures back some of the excess rents that any patent system is prone to, but selectively, based on the value of the invention. The patent holder would always have the option to extend or not extend, after all.
• It provides a "back-pressure" whereby simply letting the patent lapse in the face of competition or litigation would make progressively more sense as time went on.

What Dave said. Except I'd make it cheaper for copyright and give authors lifetime control of their characters and universes, but not their ideas.
Say, 1$ for the first year? That cleans out a lot of copyright for nonfiction. Since it already costs a lot for patents to get issued, it's not so much of a problem. Twenty years is plenty long enough. With an extension for government holding it off the market.

Dave,

In one sense we already have progressive patent maintenance fees. At issue, the fee is
$1400 + $300 as a publication fee. At 3.5 years, the fee is $900; at 7.5 years the fee is $2,300; and at 11.5 years the fee is $3,800. (Fees are half that for small entities.) There are tons of patents out there which expire at the end of the 11.5 year mark for failure to pay maintenance fees.

Rob, a patent is a limited government monopoly in exchange for full disclosure of the invention. As long as "one of ordinary skill in the art" can build the invention, no additional disclosure is required. For software, a flow chart or pseudocode meets this requirement--simply give the flowchart to any competent programmer and workable code can be written. Source code is normally provided only if the protection being sought is copyright, but it's much easier to get around a copyright protecting source code than it is to get around the idea behind a flowchart.

Patent terms seem to be about right except for pharmaceuticals. The five year additional term doesn't adequately cover the length of the regulatory process. A patent is supposed to grant a monopoly for 20 years starting from the filing date (or earliest priority date). Three to five years to get a patent granted and you are left with 15-17 years of patent life. With pharmas, it seems that a product is on the market for only 5-8 years before the patent expires and hordes of generics take over, many of which do not work as well on many people for various idiosyncratic reasons.

With ordinary manufactured products, production and distribution can be ramped up in a couple of years, giving the patent holder a good 15 years of marketing value before the patent expires. By that time, however, other companies will have gotten into the act, spurred on by the disclosure of technology in the published patent (or published patent application, which usually occurs 18 months after the filing date) to make improvements in the basic technology that weren't thought of by the original inventor. Cross-licensing becomes the order of the day, as the latecomers can't build the product without infringing on the original patent, while the original patent owner can't take advantage of the improvements patented by the latecomers. Competition is spurred, which is always healthy for the economy.

I can't compare aids drugs to a software program. And the world really won't abide by any law regarding them. So I guess you're just talking about the U.S.

Rex mentions that ". . .hordes of generics take over, many of which do not work as well on many people for various idiosyncratic reasons."

I have to disagree here. Although I work in the drug industry, and thus derive my livelihood from patented drugs, I have no problem with generics, and take them myself when I have a prescription. The generic companies are quite competent, rather alarmingly so in some cases.

I can imagine situations where problems might occur, but in reality I don't think that the individual responses to generics vary more than the responses to the original drugs.

Derek,

That might be so, but there are several drugs where the knock-off brands and generics don't work as well on me as the originals. Sudaphed comes to mind, as does the difference between Bayer aspirin and off-brands. Clarinex works better for me than Claritin, even though Clarinex, as I understand it, is simply a downstream metabolite of Claritin. For all I know, the difference between brand names and knock-offs or generics might even be related to the non-active ingredients and binders used. All I know is that there is a difference to my body.

I am not too troubled by the patent system but the copyright system, where wholesale extensions have been lobbied, is a mess. There is a song from a 1950's movie which I would like to arrange. The composer is dead. The song is out of print. It has been a nightmare to find the current owner of the copyright and then get permission to arrange and then get a library which has the actual sheet music to release it.

Read this---

Israeli Warplanes Target Suspected Hezbollah Bunker in Beirut – [may have the leader of the terrorists]

I think we would benefit from extending drug patents considerably. Other companies would have a stronger incentive to develop a superior alternative drug instead of just selling the generic once it goes out of patent.

Also, unlike in the arts, the value of drugs lies in their effectiveness, not in their aesthetic or intellectual appeal. Even if you protected a drug for 100 years, few people would still be using it 20 years later, because more effective alternatives would be available.

As a Yankee fan, you should be skeptical of anything said a Red Sox pitcher turned Dodger.

How about the following system for pharmaceutical patents:

1) Patent terms do not start until FDA approval.
2) Company asserts a development cost as part of FDA approval
(I'm even willing to allow this to include a risk factor to account for the drugs that don't make it).
3) Specify a fair return factor, like say 4.
4) Any company can then buy into the patent by paying the return factor times the development cost divided by the existing number of companies who have bought into the patent plus 1. The buy in fee being split between the existing companies bought into the patent.
5) Patent expires after some term or other (say 17 years).

So for example company A develops a wonder drug foocasil, with a development cost of $1 billion dollars. So the return factor times 4 is $4 billion dollars. It turns out foocasil is a good seller, so company B decides to get in on the act by paying company A $2 billion dollars (essentially buying a half interest in the patent). A couple of years later, company C decides to buy in by paying $1.33 billion dollars ( half to company A, half to company B). A while after that company D buys in by paying $1 billion dollars (a third to each of A,B,C) etc. Thus if you have a particularly successful drug, you essentially start developing a generic market early, while still protecting the original developers ability to profit.

This system is particularly useful for dealing with 'humanitarian drug needs'. Need aids drugs for Africa? OK, sure, let humanitarian groups simply buy into the system and produce them. Basically it provides an equitable way to bring new players to the table before patent expiration.

Sigivald, etc.: How about a software patent where the "inventor" never wrote a program to implement their idea, never attempted to market it, but simply waited until someone else came up with a similar idea independently, did the work to implement it, and put it on the market. Should the person who never did any work on the idea beyond filling out a patent application get to hold up a company that put 50,000 man-hours into it? It's happened... It's happened with hardware patents too, but far less often, because it's much harder to get a patent on hardware approved without a comprehensible description of how it works.

Two centuries ago, submarine patents like these weren't a possibility; the patent office required a working model of the invention, which separated the dreamers from true inventors. That requirement was dropped, because for one thing they ran out of storage space for working models. However, with software, storage of a working model is no problem - a source code CD-R would be the least bulky item in the patent file.

Bringing back the requirement of a working model for software patents, in the form of compilable source code, would both eliminate those who abuse the system by patenting ideas they have no intention of developing and marketing, and would in the long run promote the development of more and better software as the patents ran out and the source code in storage became part of a public library of free code. I wouldn't require a working model with the patent application, but I would require it within 3 years after the application was made. Nor would I require the patent office to compile the code and check that it ran as described; instead, require the inventor to certify that it will compile and run, so someone who challenges the patent and shows that the original source code doesn't work could sue the inventor for fraud.

Markm,

The "software" patents I write are not for software per se, but for a process or method of accomplishing something. It's ususally implemented in software, but could also be implemented in hardware or wetware or any combination thereof. These patents have been valid long before it was generally acknowledged that "software" patents could be obtained.

Onething that hasn't been mentioned yet is the idea of mandatory licensing as some other countries use. In that scheme, infringers have the right to a mandatory license, but of course, they still have to pay royalties.

No one has yet questioned the Need for Patents and/or Copyrights?

Curious. There are discussions that argue those points at www.mises.org

Beware, what even isn't Questioned.

I'd strongly recommend that those who are certain that patents are 'about right' check out David Levine and Michele Boldrin's "Against Intellectual Monopoly"; draft available at http://levine.sscnet.ucla.edu/general/intellectual/against.htm
... the anti-IP position isn't just for cranks or Communists, far from it.

Comments are Closed.