Inane Patents

Is it just me, or can you patent just any old thing? Maybe I just don't understand the legalese, but it just seems weird that you can "own" these:

US5295256: Automatic storage of persistent objects in a relational schema. (owned by Racal-Datacom)
US5819281: Notification of aspect value change in object-oriented programming (owned by EDS)
US5983227: Dynamic page generator (Yahoo)
US6025810: Hyper-Light-Speed antenna (also accelerates plant growth)
US05443036: Exercising a cat with a laser pointer (note that it took two people to think of this )
TheIsNotPatentApplication

You can look these up on http://www.delphion.com --AnthonyLander

How about AnnoyingPatents?? LZW compression comes to mind. Unisys quite cheerfully let the patent hang for awhile after they bought <insert name of corporation here>, while in the meantime, LZW became the most popular application of any patent in the world. From GIFs to PDFs to everything in between. Oh, now that everyone is using it, let's start charging for it! Ahhhhhhhh.... BaitAndSwitch. -- SunirShah

The patent was awarded to Sperry in 1985; in 1986 Sperry and Burroughs merged to form Unisys.

Maybe, but I think most people not opposed to patents altogether can at least agree that LZW was something that a patent should legitimately have been granted for. -- DanielKnapp

Yes, but when they failed to defend their patent they should have lost it.

I guess what really bothers me (and hence prompted this page) is that certain things seem to just belong rightfully in the public domain. Binary Search or hashing, for example. Could someone have patented them?

What about things that are arguably discovered rather than invented? Genes come to mind. Why do I get to "own" something for being the first person to see or describe it? Especially if it's been there longer than I have. What about mathematical or physical properties of the universe? Can someone own them too?

Can someone straighten me out on this? --AnthonyLander

Genes are technically not patented. However, for any individual gene patent, every known method of isolating that particular gene is patented, so anyone who does any work with that gene is almost certainly going to be subject to the patent.

I will take a shot at this, but know that I am not a lawyer, and the law in this area is in flux. Note also that I am speaking about the US--I don't know much about patent law in other countries.

In practice, there are three main things that an idea has to be to make it patentable:

Not part of the prior art -- that is, new in some way.
Inobvious to one skilled in the relevant field(s).
Not purely theoretical--you aren't supposed to be able to patent Fermat's Last Theorem.

All of these are open to interpretation, and over the past decade or two several things have changed.

A new theory for getting patents for more theoretical concepts has been accepted. The idea basically is that you convert your theory into a process for doing something, and then patent the process--for instance, if you figure out the formula for the surface of a right circular cylinder, you can't patent that, but turn it into an algorithm controlling a device for dispensing the correct amount of frosting on a layer cake, and it likely is patentable. Prior to the acceptance of this theory, the RSA patents on public-key cryptography (to take an algorithm I consider both novel and inobvious) would not have been issued because they are strictly mathematical. So since elliptical-curve theory apparently has cryptographic uses, maybe you can patent FermatsLastTheorem...
Software has become vitally important in a wide number of fields, so people want to protect their algorithms/methods. The problem is that traditionally, software practitioners neither could (because the above theory didn't yet exist) nor cared to patent their algorithms, although they may well have published them. Now they can, and because patents are seen as proof that a company has proprietary intellectual property, which translates into cash via venture capitalists and IPO's, there is a tremendous impulse to patent anything possible. It also gives a company the opportunity, even if a patent is groundless, to legally harass competitors. It costs money to have a patent overturned, even if it is extremely dubious.
Patent examiners typically don't have a lot of experience in this area, and aren't set up to search the prior art--they are much better set up to search for earlier patents, but because there are relatively few old software patents, this doesn't accomplish as much as it would in other fields. It also seems that sometimes they look narrowly at prior art--so that sometimes you see a patent for doing something in the context of the web. If the patent examiner regards the web context as different from previous computational contexts, then it is more likely they will find no prior art.

As a result, it seems that many patents are granted for things that either existed before or are obvious to experienced software people One of my favorites was a patent for displaying a screen cursor by XORing it with the screen image--certainly a good idea, but pretty obvious to anyone who knows what an XOR is. The one-click shopping Amazon patent is a recent absurdity. (for a report on a discussion about this patent and its implications, you can sorry, used to be able to look at http://www.oreilly.com/ask_tim/bezos_0300.html)

Personally, I hope the Patent Office becomes more selective in what patents it grants, but I have no idea whether that is likely to happen.

-- MatthewWilbert

That's a very good explanation. Unfortunately, I still have that sinking feeling in the pit of my stomach. Ugh. --AnthonyLander

The granting of patents is supposed to generate a greater good, by encouraging invention. As part of the deal, patents expire.

In order to be a valid patent, the patent must be instructive. It must provide sufficient information for a practitioner to instantiate the invention.

Thus at the end of the term of the patent, we all get to play. Now you can argue that the current term is too long. But I think that there is some merit here.

See http://www.delphion.com/details?&pn=US04100569__ . This is a patent for the Ultimatte compositing process. It was issued in 1978, and hence has expired. Now this happens to have originated as an electronic process, but it is certainly all in software today.

As the Ultimatte patents expire one by one, the industry at large can build better and better compositors. Ultimatte keeps its edge by continuing to innovate. And in 20 years, we'll all be able to do what Ultimatte does today. Not only will we be permitted to do what Ultimatte does today, we can take our instruction from the patents.

So where is the line? What constitutes an InanePatent?, as opposed to a SoftwarePatentBecauseEverythingIsSoftwareToday??

-- GarthDickie

You know, when I started going through the inane patents on this page, I had the urge to scream. Loudly. But I didn't want to cause alarm in my neighbors, so I looked around for something to scream into. I found a suitable jug and expressed my rage.

Unfortunately, I discovered that I violated the following patent:

http://www.delphion.com/details?&pn=USD0360249__

-- JohnPassaniti

(It may or may not be worth pointing out that that's a design patent, so it only covers things that look just like it, not the general idea of, say, screaming into jugs. So you probably didn't really violate it! *8) -- DavidChess)

While looking at the "device to scream into" patent, I noticed the reference to this:

http://www.delphion.com/details?pn10=USD0337071

which seems very apropos to the topic of this page. --KrisJohnson

This is 'a friend told me that someone they knew said that...' type of thing so I can't vouch for it's factual content but... Microsoft apparently patented 'Maximize' and 'Minimize'. Anybody able to confirm or refute this? --LanceWalton

You cannot patent a word. Trademark maybe, but you don't see any TM or (R) next to the words Maximize and Minimize in Microsoft documentation. --KrisJohnson

What about Microsoft trademarking 'Windows'? It's particularly disgusting when a company figures out a way to gain control of a generic term. -- MitchellModel

How about "Sun", "Apple", "Oracle", .... Microsoft hasn't done anything that thousands of other companies haven't already done. And it's not like people are no longer allowed to use the word "windows"; you just can't use it as the name of a software product.

I feel that patents on genes are fundamentally wrong. (And that we need to find some other social/legal mechanism to encourage and protect such work.)

"If it's in my body, then I own it."

...is what I say. -- JeffGrigg

Although I've been working for a biotech company, I've always found it truly bizarre that genes could be patented -- are we going to have to pay licensing fees for the proteins our bodies generate, such as insulin, hemoglobin, and seratonin? However, the public gets fed a very superficial version of the biotech patenting issue. In fact, although I've been discussing it for years, I missed for a long time a very important point that was fairly clearly reiterated in a recent federal ruling (see http://frwebgate.access.gpo.gov/cgi-bin/getdoc.cgi?dbname=2001_register&docid=01-322-filed.pdf, p. 1093, which addresses some of the common objections to gene patenting): the DNA sequences that are being patented don't exist in your body, at least not in the form in which they are patented. In their natural form a gene's DNA sequence is broken up into several coding regions (exons) separated by non-coding regions (introns). To produce a protein, the cell machinery first transcribes a gene's exons into RNA, then translates the RNA to assemble the protein. (This is an extremely brief summary of a more complicated process!) Biotech companies normally work with and patent "complementary DNA" (cDNA), which is the DNA-code equivalent of the gene's RNA -- the gene's DNA without the introns, a form which doesn't occur naturally in cells. So although this is a pretty fine distinction they are actually patenting a synthetic product -- a DNA molecule representing information extracted from the actual DNA of the cell -- that does not occur in that form in nature, though the equivalent information does in the actual DNA and the product of that information does exist in the body in the form of proteins built from its instructions. Moreover, patent applications are required (as the ruling cited above emphasizes) to specify the utility of the DNA (or DNA fragment -- it doesn't have to be the entire sequence for a protein); you can't just submit ATCG strings and hope to gain a patent, although this was tried a lot in the early days of genomics. -- MitchellModel

ISTM that the cDNA sequence is an obvious application of a discovery, since anyone in the field who knew that a particular RNA sequence was interesting would know, in principle, how to determine the cDNA. IMO that should make it unpatentable.

Patents expire one day, so maybe patenting genes found now is not that bad. However, a common technique is to refresh old patents by building new ones on it, or by splitting a patent into sub patents. This is a practice that I personally think should be controlled better. Further, maybe more organizations or a General Attorney should take an interest in the revocation of dubious patents.

Can I patent my unit as a "gene delivery mechanism"? --PhlIp

You can if your use of your unit is novel and not obvious to a skilled practioner of gene delivery.

Here is an article opposed the the current level of scrutiny given to software patents: http://www.nytimes.com/library/magazine/home/20000312mag-patents.html (this is the NY Times, so free registration is required if you haven't already done it, or grab a login from http://bugmenot.com).

The March 29th edition of /. referenced the following article: http://public.wsj.com/news/hmc/sb954286078412266261.htm whose title is U.S. Will Give Web Patents - More Scrutiny Under New Plan

CategoryLegal CategoryStupid