Ordinary Skill In The Art

ClarkEvans writes: "Jeffrey D. Ullman, professor at Stanford University and famous contributor to the excellent Dragon Book, writes about software patents in his paper, Ordinary Skill in the Art. He has some very serious analysis here; I hope Congress reads up." It's intended for computer scientists rather than Congress, and it looks like he has some good ideas.

This is actually quite fascinating

It seems like it would be quite difficult to obtain a patent on any kind of software. Mr. Ullman describes three guidelines for software to be patentable:

• The idea must be "novel"
• The idea must be "nonobvious"
• The idea must be useful

The one that seems particularly difficult here is the nonobvious part. Mr. Ullman says that "novel" simply means it does not appear in previous literature. But for software to be nonobvious....

It seems that there are so many ways to achieve the same results with software, that it'd be very difficult to come up with something that someone else couldn't figure out to do in a completely different way.

The neat thing about this, though, is that someone who has truly stumbled on something fantasticly useful, unique, and clever is rewarded. RSA for example....

Re:Patents

Virtually everything you just mentioned refers to trademarks, not patents - at least not in any juristiction I'm aware of. Patents are rights confered upon the inventor of some invention that allows them to regulate the production of items using that invention, including collect royalties or refuse to allow anyone to produce it in the first place. Once you have a patent, you have those rights for a fixed length of time and nothing can result in those rights being removed from you.

Trademarks, however, are pretty much granted for life but have to be enforced or they will be taken away, which is why companies like Filofax, Biro, etc, tend to be snippy with any magazine that refers to a personal organiser, pen, etc using their trademark genericaly.

Obviously that "patent" lawyer was more confusing than you think...
--

LEGO blocks and patents

I think of software patents as things built out of LEGO blocks. We all have the same blocks, and the same potentially infinite number of ways to assemble them.

Is it possible to think of a truly novel way to fit them together?

I made something never made before!
Well, of course you did. Like the author's example of the programming assignment -- everyone will do it a little differently.

I found new ways to put blocks together!
Like, by gluing the flat sides together. Is that useful? Probably. Our lawyers will have to look into it.

I used a mold and some plastic and invented a new block!
You did? Congratulations! It probably deserves a patent, let's check the literature to be sure. . .

Metaphors R us,

-Omar

Open source as a demonstration of prior art

However, documents such as unpublished software are excluded as a proof that an idea is not novel, because the software itself is not a public document.

There have been discussions here on Slashdot before about the obvious idea that open source clearly qualifies as publication of the ideas embodied in it. It is possible that open source software will save the entire software industry from becoming mired in disaster of its own making through its own patents. While I respect intellectual property law protections for those ideas that clearly meet all three criteria, I think we have all seen examples of many that don't. As an example, would Emacs' info mode qualify as prior art in defense against BT's hyperlink patent? I honestly don't know enough of the details of that patent to say, nor am I a patent attorney.

It is clear however that patents are being granted that are falling short of the obviousness test. Unfortunately, when viewed as individual cases, it is difficult to determine where to draw the line. However, if more source code were published, there would be no question that many patent applications failed the novelty test. The bottom line is that if an open source project uses ideas from a patent in code that was released before the patent application, then either it doesn't infringe because the patent wasn't novel, or it doesn't infringe because it doesn't fall within the scope of the patent.

Re:In defense of drug companies

The R&D that created every new "miracle drug" was funded by the investors and bond holders of private companies.

And the basic reasearch that this applied research was based upon was most likely funded by grants from the NSF, NIH or any of several other TLAs in the US government. And if it wasn't funded by the government, lots of non-profit organizations (the American Heart Association, American Lung Association, American Cancer Society, etc.) fork over grants to university rearchers for doing work on the organizations organ/disease of choice. I certainly don't think that pharm companies should be replaced by the Feds, but it would be nice if people would stop acting like they worked in a vacuum and generated their profits purely de novo, instead of using public domain data to base their IP on.

Do you suppose that the government would use our tax dollars with anywhere near the efficiency that publicly held corporations would? How would the research priorities be set? Would we postpone research on cures for more widespread ailments so that we could focus on "celebrity diseaeses" whose cures would further someone's political career?

It's funny that you offer this up as a pro-pharm arguement, after your sample 'wonder drugs' - Claritin, Flonase and Viagra - cure such life-threatening issues as hay fever and impotence, not cancer or heart disease. I have prescriptions for Claritin and Flonase (and Aerobid-M and Albuterol and Accolate), and my quality of life has improved since I started taking them. But I'd rather have more and better anti-cancer meds for my mom and grandparents so they'd still have the uteruses, kidney and section of colon that they were born with and I'd live with my runny nose and sinus headaches, okay?

Of course, if I needed that Viagra prescription...

Re:In defense of drug companies

And the basic reasearch that this applied research was based upon was most likely funded by grants from the NSF, NIH or any of several other TLAs in the US government.

Basic research in no way gives you a drug that can be admistered to the general population with any degree of assurance of safety or effectiveness. All it does is increase the knowledge of the mechanisms in place. Commercialization of that knowledge to the point where you can take it in the form of a pill is a 10-15 years long process that may well cost a BILLION dollars. No private organization would undertake this effort without assurance of being able to recoup their investment.

The work of the NIH is of course crucial to the process, because it provides the seed. However it is not sufficient. If you look at the real numbers the percentage of the funds that the NIH contributes to the overall process of developing a new drug is no more than 1%.

Ullman may be an expert in CS and patent trials. .

BUT:

He is a bit weak on business.

Let's take his example of Amazon. His notion that selling books on the internet ought to be protectable as IP is flawed from the business, legal, and patent perspective. This is a field where the free market rules.

Let's say I get the idea of selling dictionaries from a boat. I believe this would be a novel business model. Should I be allowed exlusive legal rights to this idea?

No. The idea of selling from boat is not novel. The extentsion to selling only dictionaries is in no way inovative, even IF it is unique. Anybody could have tried it, and if succesful there is no way to restrict anyone else from taking up my * business model.*

Selling books on the internet is no more a novel and protectable idea than selling anything else on the internet, and selling ANYTHING in a public venue so obvious to anyone who has ever sold anything.

Should the first person to sell books from a fixed location of business, i.e. a storefront, have been granted exclusive rights to do so? The idea is logically, and legally, ubsurd.

How about selling books from a pushcart, or a bicycle, or an airplane, or the inside of a trenchcoat, etc., etc., etc.?

ALL such solutions are obvious and available to anyone. It may take GUTS to be the first to try any of these out and see if they work, but none of them are in any way novel in the terms of the *idea.*

Mr. Ullman seems to wish to EXPAND the idea of patents to cover IDEAS, when they are specifically intended to cover *devices.* It is unfortunate that the idea of "device" has been expanded to include mathmatical algorithms. This is the REAL problem with the system as it stands. FURTHER expanding the system into the realm of purely abstract concepts would be a horrible mistake.

This is the realm of the free market, where the * customer* decides what is valid and what isn't.

If selling dictionaries from a boat proves to be a failure * because of lack of sufficient customer support to provide profit* than that business model will be dropped. Perhaps pushcarts will work.

If it works, it will be copied, and SHOULD be. Without having to pay a "licence fee" to the first person to have tried it out.

What would the world be like now if Sears and Roebuck had been able to obtain a patent on 'remote delvery of goods'?

Now returning specifically to Amazon for a moment, what they have REALLY attempted to do is patent the IDEA of saying "OK" to a customer who says, " Give me one of those and two of those and put it on my account." As we all know this idea is patently obvious and has been used in EVERY forum available for selling since the begining of each of those mediums. It fits Ullman's catagory of those things that are SO obvious they arn't recorded in the literature, in this case because its actually implimentation * predates * literature itself. It has been common usage known to preschool children for * millenia.*

Ullman wants to see a system where single solution patent cannot cover the CLASS of solutions to the problem. He ALSO wants to see a system whereby the QUALITY of the solution determines patentability. This the solution of an academic thinking in the terms of *academic* validity. It is poor BUSINESS solution, and patents are business problem.

In fact, the first solution, making only unique solutions patentable not the class, makes the second redundant * because the solutions will be subject to the powers of the free market* rather than a doctoral of journal review board.

In this sense the market *IS* the review board.

So, let us say that Amazon can only patent their specific SOLUTION to the obvious problem of one click shopping. Ok, someone comes up with a DIFFERENT solution that may, in fact, be academically INFERIOR to Amazon's, but. . . they can patent it, use it, licence it, and the MARKET will decide whether Amazon's solution is superior enough to pay a licence fee to use it, pay a lessor licence fee to the holder of the rights to the "inferior" technology that accomplishes the same thing, or to simply spend the time, effort and money to develop their OWN novel solution to the problem.

I think with a bit of thought you will see that his two ideas are in a BUSINESS sense actually mutually incompatible, and incompatable with the very philosophical basis of patents themselves, which is to promote the MULTIPLICITY of ideas available for use to the public and to allow the consumer the choice of which idea is most suitable. The lack of a licence fee, in the MARKET, is, in fact, one of the parameters of suitability!

Anyone in the open source community ought to understand that last point.

The granting of class protection to single solution to an obvious problem would be equivilant to saying that the patent on the combine harvester granted IP rights to ALL methods of harvesting, patently, if you'll forgive the pun, ridiculous. Yet this is exactly what juries are doing for algorithms.

The Wright Bros. were not granted a patent on the airplane. That was deemed both obvious, and with a wealth of prior art. What they WERE granted a patent on was three axis control * by means of warping the control surfaces.*

THAT was novel, useful, and non obvious.

So what happened? Because the patent did not cover the CLASS of possible control methods inovation was actually STIMULATED by the patent as people sought a way to accomplish control by means NOT coverd by the patent, and thus the superior aileron was born. THIS is the way patents were supposed to work.

If wing warping were in fact superior it would have won in the market * anyway * because people would have prefered to use the superior system even though it meant paying a licence fee. The aileron was both superior, and fee free, making it the clear winner * in the market.* It would have worked just as well the other way around. If the aileron had a licence fee attached and wing warping didn't the aileron would STILL have prevailed. Can you imagine a P-51 Mustang using wing warping?

Again, anyone in the open source community should understand this idea and see the clear parallels.

A pity patent clerks, and most especially juries, do not.

Re:Patent requirements

This goes largely on a state by state basis, subject to complience with the federal constitution, and would require my posting code from all 50 states to answer.

However, we operate, with the exception of Lousiana, under the British common law system. This means that *case law,* i.e. precedent, has much of the force of actual CODE, even though so actual code does not, in fact, exist.

If you have been paying any attention to our congress critters, and the popular media reporting their inane spoutings lately, you'll have run across complaints about " judicial activism,* i.e. the 'problem' of judges writing law from the bench and usurping the power of the legislative branch of government.

The fact is, that the very basis of the common law system and philosophy is that * judges are empowered to do this.* Yes, they 'interpret' the law, but in any specific case their interpretation IS the law, unless, and until, overturned by a higher court.

If the ruling is subsequently ratfied by * the consensus of other judges* ruling likewise it, in fact, is law in practice, without the existence of code.

ANY lawyer ought to know this, it is a rookie mistake to go into court and cite LAW, and get shot down by a more experienced opponant who cites CASE.

So, to answer your question I must cite not only federal code, and code from each of the 50 states, but I must also argue and cite PRECEDENT in both federal courts and the courts of all 50 states.

You ask me to do an awful lot. I'm afraid I can't comply in a Slashdot post. It would take me at least a small book to do so, and many very LARGE books have been written on the subject and can be found in any law school library if you would like to do some research on your own. Start your search under the subject heading " Jury Selection."

Now that fact of the matter is that there is relatively little CODE about the selection and conduct of juries, but there is MUCH case going back centuries.

I'll point out a glaring example though, in Colorado knowledge of the *rights of a juror* is grounds for both prejudicial dismissal of a juror during selection, and declaration of a mistrial at any subsequent point.

So I suppose you are correct is some technical sense here. There is no law that says he CAN'T serve on a jury. The law says the *trial is invalid* if he serves on a jury.

A valid legal point which could actually have impact in actual case, but IN PRACTICE it has the *effect* of making it illegal for someone with the 'specialized knowledge' of the rights of a juror to sit on a jury!

The BULK of American law is of this sort, that without either specific code OR precedent, but with the practical EFFECT of law.

In most states the combined power of code, precedent and *practical* effect make it illegal for a person of specialized knowledge in a case to sit on a jury.

Last I knew all states that didn't have actual code forbiding the sitting of such a person on a jury had code and/or precedent requiring the juror to be instructed that he could not *use* his speciallized knowledge in reaching his decision, and the use of such knowlegde was grounds for dismisal and/or misstrial.

Anything which is grounds for a mistrial is, IN PRACTICE, illegal. Lawyers will actually use those words. " We can't do that. It's illegal. They'll get a mistrial."

Again, TECHNICALLY you are perhaps correct. Grounds for a mistrial is not EXACTLY the same thing as its being illegal.

ANYTHING where a judge can say " you can't do that" is effectively law in his court. Any such statement that the Supremes would validate * IS LAW!*

The correct answer to the law professors question, " What is the law in this case?" is NOT a citation of code, but " Whatever the judge and/or jury say it is."

My favorite classic example of this is the jury ruling, " Not quilty your Honor, but he has to give back the mule he stole from the plaintif."

And THAT was * the LAW!* Code be damned.

And THAT was the way the founding fathers INTENDED the law to work.

Ullman is off-base on patents

Ullman writes: "I'd love to see a requirement for the demonstration, along with every patent granted, of the superiority of the method."
Why? If the new method is inferior, that's the applicant's problem. No public purpose would be served by such a restriction. It may take years, and further development, to find out that a new idea is a good one. Xerography is a classic example. Chester Carlson's original 1939 process produced very fuzzy images, and it was many years before xerography beat photography for copying. Inventions relate to commercializable concepts. You don't know whether a concept is commercially superior until you turn it into a product and sell it. That's not a test the patent office can, or should, apply.

Ullman also writes: "An idea is nonobvious if it would not be discovered by one of ``ordinary skill in the art'' when the idea was needed." And one proves non-obviousness to the patent office by showing that the idea was needed but someone else didn't discover it. It's less about what someone might think of than what other people in the field actually did think of. If you're working in an area where the problem is recognized but the available solutions are bad, and you have a better solution, then that new solution is non-obvious. Much is obvious in hindsight. But that's not how patents are granted. Nor should they be.

Obvious patents are a problem only in newly opened areas where there is little prior art. They aren't currently a problem in, for example, gears and linkages, which was a hot area in patents around 1890. Each new technology goes through a period when there are patents on the basic ideas. Automobiles, telephones, electric lights, and radio all went through this period, accompanied by litigation battles of much greater significance than anything we've seen in the computer software area. Eventually, the patents time out, and everything in the new field is in the public domain. That's the way it's supposed to work.

I could say more, but that deals with Ullman's main arguments.

Re:"Intellectual Property"

From what I understand, intellectual property is usually undisclosed information,
Then you understand nothing. Go to google and do a search for "intellectual property faq." Read
it. Don't post again until you have.

Once you've done that, go find the guy who moderated you up as "INSIGHTFUL" and force him to eat a hard copy of the FAQ.