Slashdot Mirror
Software Patent Sanity on the Way?
Ars Technica is reporting that the traditionally silent US Patent and Trademark Office (USPTO) may be starting to turn things around. It seems that in recent action the USPTO has started to make it much easier to invalidate software patents with some saying that the abolition of such patents may be in the distant future. "Duffy cites four recent cases that illustrate the Patent Office's growing hostility to the patenting of software and other abstract concepts. While the USPTO hasn't formally called for the abolition of software patents, the positions it took in these cases do suggest a growing skepticism. In the first two cases, decided last fall, the United States Court of Appeals for the Federal Circuit (which has jurisdiction over patent appeals) upheld patent rejections by the USPTO. They were not software patent cases, as such. In In Re Nuijten, the court considered a patent related to an algorithm for adding a watermark to a digital media file. The Federal Circuit did not invalidate the claims relating to the watermarking algorithm itself; everyone seemed to agree that the algorithm was patentable. Rather, the decision focused on whether a digital signal could be the subject of a patent claim. The court concluded that it could not. A victory for common sense, perhaps, but hardly a rejection of software patents."
A lot of the article is talking about another article that was on slashdot recently.
http://yro.slashdot.org/article.pl?sid=08/07/24/1458215
repeat from last week?
It does appear that consistent negative publicity helps. Too bad that it has taken over a [i]decade[/i] before changes appear.
It also seems that this long delay is going to lead to a rather messy situation... although possibly less so than it already was :-)
see a Text Widget
I can hardly believe that the US government would make a sensible decision that also happens to be in the interest of its citizens.
.... it is abstact matter with a definable physics supporting its use.
To bad there is not stock to invest in on this overall debate as the outcome is certain. Software patents will become a thing of the past. An embarrassing thing at that.
The thing that arstechnica is missing is that Duffy is a troll. He is representing amicus in the Bilski case, and is raising the possibility that the USPTO is adopting a position that will invalidate most business process / computer based patents as a FUD attack against any attempt to limit the scope of patentability in this field.
From the article:
Innocent infringement : defendant was not aware of any copyright infringement, and upon information and belief some or all of the copies which she downloaded did not bear copyright notice.
This looks very weird - when people rip CDs and DVDs, they rarely (if ever) attach any copyright notice to the resulting mp3 and avi files... Would it mean, that because the copyright notice has been removed (it was on the CD case for sure, or the load screen of the DVD), then you don't know you are infringing? As much as I applaud the rest of the complaints, this is just silly. On the internet it is mostly: "everything is copyrighted except if it's explicitly noted", not the other way around...
On the other hand, if it gets accepted, then everyone is pretty innocent from this point on... Would be fun. :)
I've always believed that patents should include not just the idea being patented, but also details on how to recreate such an idea (ie. the prototype).
For physical objects, this means schematics. For drugs and such this means formulas. For software it means source code.
See how many companies will be willing (or in the case of patent trolls, ABLE) to patent software when they have to pony up a working implementation as part of the patent application (and thus public record).
I am also against 'secret' or 'partially secret' patents, how is someone supposed to know they are infringing on a patent if they can't get all the details on a patent?
I've asked before and I've never gotten a good answer. How can you patent the act of using something in exactly the manner in which it was designed to be used? A computer is designed to execute an arbitrary series of pre-defined instructions. That's it's only function. Software is just a list of such instructions. How is that patentable? It's not a new invention. It's not an extension of the original device. It's like patenting the act of driving a nail with a hammer, or letting fresh air into a room by opening a window.
Say a particular calculator is patented, and I patent the act of entering 2+2 on it. Then someone else comes by and patents the act of entering (3+7)/2. Hey, it's an innovative new application of an existing device!
I have found there are just two ways to go.
It all comes down to livin' fast or dyin' slow. -REK, Jr.
This is too easy to stop and will probably suffer an ignominous death.
If it ever gained a little steam outside the right-thinking-patent-repair-club, it's dead in less than 30 seconds.
Fox News and many other runs a blurb something like... "Sources inside the White House have stated that in difficult economic times like this, it's a terrible idea to introduce more regulations that will surely lead to fewer jobs..." See how easy that is?
Sadly, this is the state of discourse in American politics.
BTW, more patent regulations support the terrorists, illegal immigrants and tooth decay.
http://www.maxineudall.com/2010/02/should-economists-be-sued-for-malpractice.html
More than embarrasing. Costly.
Like the saying, "Make hay while the sun shines" the fallacy that equates license to property will burn many more fools before we are done. Money in the bank is always decreasing in value. Using that money keeps it alive. Many profited in the domain name business, if they were wise enough to buy and sell them in a short lived market. The coming ICANN changes will soon massively devalue once treasured domains and those foolish enough to be left holding them, but not using them, will lose out.
The same is true of bogus patents. Fine if you were able to catch the wave of USPOs mistake that sparked the software patent war in the first place, and leverage ivalid patents, but disaster for those caught holding them in the belief that they are tangible property.
When the correction finally comes (and it will) it will wipe billions off the value of some companies. It couldn't happen to more deserving people of course. And perhaps that's the main reason it hasn't happened already. Software patents always have been, by widely recognised standards including those of UPSO itself, invalid. It's only the propagation of an error that has allowed some to profit meanwhile. The value of a market resulting from an error should not be justification to allow that error to persist.
Let's say someone finds a new way to cut logs that let you build log cabins almost as easily but many times more sturdily and with better isolation. This would surely be patentable.
Let's say someone invents a new file system, that lets you access files almost as quickly but with many times the protection against data corruption. Why would this not be patentable?
My impression is that the case against software patents is really a fight by proxy against patents in general, recognising the growing role that software plays to make anything happen.
That's an odd thing to agree upon, because algorithms are not patentable (search for `algorithm')
But methods are. I forget where I read this, but the difference was explained something like this --
Bob: So, algorithms are not patentable and methods are. ... algorithms are not patentable, but methods are ...
Lawyer: Right.
Bob: But what's the difference between an algorithm and a method? Aren't they pretty much the same thing?
Lawyer: Listen carefully
Bob: ?
http://en.wikipedia.org/wiki/Submarine_patent.
Submarine patent is an informal term for a patent first published and granted long after the initial application was filed. In analogy to a submarine, its presence is unknown to the public; it stays under water, i.e., unpublished, for long periods, then emerges, i.e., granted and published, and surprises the relevant market. This practice was possible previously under the United States patent law, and is now not practical with present patent filings since the U.S. signed the TRIPS agreement of the WTO: since 1995, patent terms (20 years in the U.S.) are measured from the original filing or priority date, and not the date of issuance. A few potential submarine patents may result from pre-1995 filings that have yet to be granted and may remain unpublished until issuance. Submarine patents are considered by many as a procedural lache (a delay in enforcing one's rights, which may cause the rights to be lost).
Then DNA should not be either. Hopefully that logically follows from this, and gene sequence patents are complete bravo sierra.
The real issue, which most people avoid addressing, is that there is no practical distinction between software patents, chemical process patents, or machinery patents as a necessary consequence of basic theory. The reason this has become an issue at all is because there is increasingly little distinction in practice as well. Consequently, any dividing line is going to be arbitrary and capricious. Note that there is a similar emerging problem with copyright law, which is also premised on a false model of the universe that is starting to become obvious in practice. Yet few people are suggesting we solve this problem by rectifying the law with reality, instead opting to promote an alternative fantasy model of the nature of the universe that will ultimately break when it intersects with reality.
As every computer geek should know, there is no theoretical distinction between the machine, the program, and the data. At one time there was a practical distinction, but those lines have been blurring for many decades now. Any solution that pretends like these are theoretically distinct classes of thing solves nothing, as the cause of this problem was pretending a theoretical distinction exists where none does in the first place.
The problem I see with software patents is that people are patenting the wrong ends of their ideas (they're putting their makeup on their asses in other words).
Take for example Amazon's one-click checkout. The idea of a one-click checkout should not be patentable. Anyone number of people should be able to accept a single click to check-out, what should be patented is the system behind the checkout. The mechanism for tying in the user's login, prioritizing recently used shipping addresses and payment methods, etc... You need to patent the process or the invention, not an ethereal idea.
If I invent Widget A that performs task A, and am awarded patent "Widget A for performing Task A", and someone realized widget A will also perform task B without any modifications, they can't patent "Widget A for peforming task B", because I still own the patent for Widget A, which is all that really matters. I own the exclusive rights for Widget A, no one else can reproduce Widget A regardless of what they want use it for. Amazon didn't invent one-clicking (didn't Microsoft patent that recently?), so they can't say no one else can use one-clicking for checking-out.
It's similar to the patent that the adult-entertainment (read: porno) industry has been fighting for years. Some company patented the idea that videos could be downloaded from the internet. Problem is, since that company did not create the internet nor the http protocol nor the first web-browser, they didn't actually create anything that had to do with the content their patent covered. Improving upon an invention means changing the invention, not mentioning something else the invention could be used for. That's the problem with software patents. People are patenting what existing technology can do, but if they don't own the existing technology, they can't tell other people they can't use it for other things.
Eggs
Milk
Bread
Cat Litter
Soda
Odd. I was under the impression that the USPTO had some degree of control over the USPTO.
I just read Slashdot for the articles.
The first two cases cited are from 2006, hardly recent.. Also not relevant.
"i stand on the edge of destruction" -shai hulud
If algorithms/Computer programs can be proven to be correct or incorrect using mathematics, why are they not simply an extension of mathematics? They are almost entirely interchangeable to the point of in many cases being substitutable languages for one another-- f(x)= a+b, int f (int a, int b) { return a+b; } Thus if mathematical algorithms are not patentable and it can be shown that the "laws" of programming and logic are essentially the same as the laws of mathematical computation and proof, shouldn't that be enough to say that if Mathematical algorithms are not patentable, then mathematical algorithms expressed as machine instructions (programs) are not patentable?
Okay...so if I design a new game, say, Duke Nukem Forever or something, it should be free for whoever wants to play it? Is that what people are arguing at some places here?
Yeah, you'd think huh? But if their decisions are challenged, they escalate to the Board of Patent Appeals and Interferences, and then the United States Court of Appeals for the Federal Circuit.
...'tis easier to blame than to improve.
Thanks for clearing that up somewhat. Still, it sounds a lot better than the 'wait for Congress' picture the AC painted.
I just read Slashdot for the articles.
I think everyone on this board should read:
http://en.wikipedia.org/wiki/Diamond_v._Diehr
"an otherwise patentable invention did not become unpatentable simply because a computer was involved."
I agree with scaling back of software patents with certain reservations. I don't think you should be able to patent abstract concepts such as formulas or even general purpose application software, but you should to make patent claims on certain kinds of software so long as the scope is narrow and there are specific hardware and software specifications.
Software that nonsuperflously extends the basic functionality of a particular device beyond what it was designed to do should be patentable. The best example of this is custom device drivers and but other novel software extensions might also apply.
Furthermore internal network architectures should be patentable. If you have a network that you own and operate, and you have a protocol that you use for that network, you should be able to patent your protocols so that third parties can't operate on your network by reverse engineering your protocols.
Comment removed based on user account deletion
The example of the widget's patentability ignores that not all configurations are "novel" and "non-obvoious".
I like the standard that the Supreme Court put forth in the KSR case. In a nutshell, simply combining already known components and design elements and getting the expected result is not patentable. If, however, the result is not something anyone "skilled in the art" would expect, then it qualifies.
So, say we design a chemical plant to produce gasoline from coal (which has been done many times before), but happen to run across a tweak to the materials in the pipes that causes the reactions to occur faster than theory predicts, we have a patentable configuration. The addition of a previously unknown catalyst is the patentable idea, not the already-known process.
Unless the code to be patented does something unexpectedly beneficial, it falls into the same category of "obvious". Just solving a new question with a combination of already-known steps doesn't cut it here.
I've said it before and I'll say it again, Software patents should be Copyrights, not patents. Patents belong on physical items. It would be like patenting the method of turning a page in a book.
If someone can figure out a better way to code the same thing you are doing without using your code, more power to them. That's innovation.
Every time I start to have faith in humanity, I ruin it by driving to work between 7 and 8 am.
So what? The development costs for software can be pretty huge. You still need folks to figure out which problems really need solving, and folks to figure out how to best solve those problems; that's gonna cost you. And suppose you've paid that cost, and are working hard to sell your solution. Well, guess what, some bigger fish can come along and build the same thing for cheaper (since they can copy the design that cost you so much to make), and beat you.
Really, the problem with software patents is that they're being awarded at the wrong level of granularity, and for things that are too obvious. For example, a patent for a simple compression algorithm is something that's too fine-grained; while a patent for something like one-click purchases is too obvious. A patent for a novel approach at using software to manage a specific kind of supply chain problems by increasing the efficiency of one kind of factor? Now that sounds better.
Are you adequate?
It *is* interesting.
i wish i could stop
"Yet few people are suggesting we solve this problem by rectifying the law with reality"
OK, so I actually read the PDF. You are indeed correct in this case. The decision here hinged on how to interpret the word "manufacture" in the context of patent law, and to do that the court relied on a definition used in a previous case involving the American Fruit Growers. So, this court has decided that a digital signal is not manufactured based on a completely different case involving fruit. (They've also declared that a photonic or electrical signal is not matter, something Mr Einstein might have something to say about, but I digress.)
The fundamental problem is the law works like a flood-fill algorithm. The law looks at narrow definitions and specific cases and then extends them to other cases regardless of whether they really should be extended in this way.
Strangely, I find myself agreeing with the dissenting judge here. Digital transmissions _are_ manufactured by the machines described by this patent (notably, the machines themselves are protected by other claims not in dispute here). Drawing a line which says the transmissions themselves are not "items of manufacture" to me implies that other results of manufacture such as drugs and genes should not be patentable either; only the means to produce them should be. Consistency please!
If being anti-social is a disease how are we not socialists?
Software is a not totally abstract like math; and has a physical form as bits on your hard disk. Software is an easy-to-reproduce machine that controls a more general purpose machine: the computer.
How is that any different than coming up with a new fantasy book with Orcs, Elves and Goblins? If it's more entertaining than the Lord of the Rings for one reason or another, why deny it's existence because Tolkien owned a patent on a fantasy book dealing with an alternate universe filled with magic, elves, and walking trees.
Like it or not, software programs use a language to perform a task (entertain, productivity, etc.) just as books use language to perform a task (entertain, educate, etc.)
Every time I start to have faith in humanity, I ruin it by driving to work between 7 and 8 am.
I've just patented a method for evaluating the credibility of a statement relative to some predefined threshold level, based on direct visual evidence.
At the bottom of the
Congratulations on not understand how the U.S. patent system works. Here's a cookie.