With software patents, I'm afraid that we are repeating history. Before WWI, various patent battles had crippled the infant aeronautics industry. Countries like France were far ahead of the United States by the beginning of the war and we had to play catch-up.

Patents are supposed to provide an incentive for inventors by protecting their inventions giving them a 17 year exclusive right to produce a product.

The term is limited to provide a counter balance so that innovations can rapidly be duplicated when the term expires. This results is wider spread access to the technology by limiting the term of the "monopoly" control of the patent holder. Given the rapid speed of innovation in computer hardware and software, 17 years is almost an infinity.

Patents are only supposed to be issued for nontrivial, original, novel implementations, and are supposed to be issued to the inventor of the product.

Software patents are issued for simple ideas like one click purchasing. It has become a "I thought of it" first protection for obvious solutions. There are IP firms that scour the web looking for ideas to steal and patent as their own. They then turn around and threaten to sue the original authors of the idea, for patent infringement unless they pay for protection. The patent system has become a vehicle for extortion by IP Law Firms operating as the mob. Software patents have been issued that are trivial, derivative, or have "priori art". For example, Microsoft being awarded a patent for the fat32 filesystem which is based on the CP/M filesystem. Even if the principles of priori art and others are still supposed to apply, the Patent Department has proven to be totally incapable of applying its own principles when it comes to accepting software patents.

Patents are issued for working products, and not on "ideas for a product".

Programming is a field of mathematics. Software is an mathematical expression of an abstract idea. Mathematics is not patentable. A program is an abstraction. By law, it is not possible to patent algorithms. The very first software patent was issued for the RSA algorithm. Not a propitious start.

Patents are supposed to be accessible to individual inventers.

When my dad made a wood carving bench, he thought about patenting it. The $20K that a legal patent search would cost would need that he would need to produce and sell benches for 150 years just to pay off the patent search, so he dropped the idea. Independent software authors have no access to the patent system. This alone could eliminate the shareware industry. Software patents today resemble the MADD doctrine. Only large institutions can afford patents. They collect large caches of patents so that in the event they get sued for patent infringement they can retaliate. The mere threat of a software infringement lawsuit presents a small software company with the choice of bankruptcy or capitulation, even when the claim is bogus. I've never heard of an IP lawyer going to prison for fraud for trying to extort money from someone even when they loose the case. But the victim faces bankruptcy even if they challenge and win the case. At the very least, development and support are halted putting the company at a disadvantage.

Completely agree, although I must ask, aren't you kind of preaching to the choir? Who is going to come out on an open-source message board and support software patents? I sure won't. :-) Anyhow, preach on.

Rants and other things not directly related to Linux (asking questions) on a technical level belong in General. This has been moved.

MS financed trolls. However they won't agree, but they deserve a little preaching. ( Especially for their connection of the SCO lawsuit. I find it interesting that even with the $50 million dollars that MS invested to kick off the lawsuit, as acknowledged by the BayStar managing partner two years ago, that Bill Gates makes the claim of never ever even hearing of BayStar. ) Before BayStar bailed out of their investment, they had demanded of SCO that they abandon Unix and concentrate totally on becomming an IP firm. Was this Microsoft's game plan all along? Was the main target to sucker SCO into destroying Unix, giving Microsoft a commercial server monopoly? Would this even be legal? )

"Things that make you go Hmmm"

My mistake.
I'm sorry, I thought that I had posted it in General. Honest.
I must of done a firefox tab booboo and switched forums without realizing it. Or maybe it's ADD related.

No biggie, that's what I figured. I only mainly posted my response for all the newbies here to the forum.

I just remembered something ironic. Unix was created as an in-house operating system to write patent applications.

Ok, I guess I'm an MS-financed troll. I have a computer engineering degree, I'm in law school with plans to get into computer-related IP, and I support software patents in some form. Note, however, I said "some form" -- not necessarily the current form.

First, I would like to ask where the original information comes from. Some of the information seems to directly contradict the information taught in my patents course. But I'll continue the response anyway...

Of the points mentioned, I would agree with this one the most, but it is inaccurate. What I agree with is the speed of innovation. I would support a reduced patent term, but I'm not sure what a reasonable term would be--I haven't wrestled with the concepts behind it. There are already exceptions to the term limit: drug companies for instance. Because drugs must meet FDA approval, Congress grants extra time to compensate for the approval process. Therefore there is no reason to believe that a reduction in patent term couldn't be carved out for software/computer related technology while progress continues at the current rate.

However, the 17 year quote is inaccurate. An inventor (or his/her assign) receives a patent for 20 years, but the "clock" starts running at the time of application for a patent. The application process is a back-and-forth process with the patent office almost always rejecting a first application because claims in the application are too broad and conflict with prior art. An amended application is submitted to address the rejections. The process is repeated until the USPTO is satisfied that the claims are sufficiently restricted (and issues the patent) or the inventor decides it is no longer worth the time and money. After a handful of back-and-forth exchanges with the USPTO, the USPTO begins to charge continuation fees because this application continues to consume time dealing with (re)analysis of claims. Regardless, as with most governmental institutions, the USPTO is not exactly speedy. The process is almost guaranteed to take a few years. The patent then has an effective term (on average) of about 17 years.

Amazon.com's one-click patent sure is popular with the anti-patent crowd, but do you really know what it patents? It's not a patent covering a business model allowing a consumer to order an item with one mouse click. The innovation entitled to the patent was the underlying handling of information. Amazon.com changed the way it handled orders by making it possible to store sensitive customer information (credit card numbers, expiration date, security code, etc.) on their servers and thereby prevent the information from being transmitted from your computer, over the Internet, to their server for every purchase you make. That may not be considered very innovative today, but consider it back when the patent was applied for, and imagine the security effect it now has. Are there underhanded individuals trying to monitor major online retailer traffic to leech sensitive information? Of course there are. Isn't that change in business practice to protect consumer information worth something? I think so.

I'm not sure about the FAT32 filesystem patent, but I'm certain there is something similar behind it. Perhaps a reorganization of the data for more efficient access time, a slightly different encoding of the data to reduce filesystem structure size, or something like that. I'll go see what I can find and post a link.

I don't agree with that. Patents protect ideas--not products. Copyright protects actual creations (painting, sculpture, music, documents). Patents protect the underlying idea that allows a product or process to perform some new, useful task.

However, the quoted text is accurate to an extent. Pure mathematics (and natural phenomenon for that matter) cannot be patented. However, I disagree with the classification of an algorithm as pure math. The RSA algorithm relies on principles of mathematics - it is not mathematics in and of itself. There are any number of ways to implement RSA encryption -- some faster, more efficient, and smaller than others. And that is exactly what patents are meant to foster: the discovery of new methods/tools and improvement of existing methods/tools. If the RSA algorithm were true, pure mathematics, there would be only one way to implement it.

Patents are accessible to inventors. Indeed, an inventor can go to D.C. and do all the searching desired. Or, if traveling to the capital is not their cup of tea, the USPTO makes the patent repository available online: USPTO Patent Search Page. In addition to allowing searches on granted patents, an inventor can search pending patent applications.

Patents can be filed by the inventors themselves, and the only cost is the USPTO application fees. However, some inventors will probably feel they've gotten "in over their head" when they see the inevitable first rejection response. Some choose to fight through it and others find an attorney to assist them. However, there is no requirement that an attorney draft the application or claims. Though, it's usually a good idea to get some help.

There are a couple of other things I'd touch on, but this is already a monster of a reply. So I'll wrap it up quickly.

If you, and anyone else reading this, don't like the way patents are handled, then by all means, let your representatives know about it. I know I've seen a handful of "calls to action" to initiate patent reform. I've even seen a newspaper article where Microsoft was voicing patent reform. Don't sit idly and watch things unfold. Get involved. That's part of the reason I went to law school. I plan to voice my opinion.

EDIT:
Using the search page linked to above, I found this FAT patent for Microsoft: Patent 7174420
I don't know if this is exactly the one referred to earlier, but the premise of this patent is the ability to make a transaction/write-safe filesystem. It boils down to having two (or more) copies of a file allocation table with only one table being "active" for file access. After a write occurs, the non-active table is updated with the new file data, the active table is "switched" to the newly updated table, and the old table is updated after becoming inactive. So that begs the question: did the CP/M filesystem referred to alter the filesystem tables in-place? If that is the case, the change proposed in the patent seems like it's useful/novel in regard to data protection.

The patent goes on to state other claims, but I haven't fought through the language. Partly because I have reading for class tomorrow, and partly because I don't know if it's the right patent.

I got the patent term length wrong. I hope you realize that being 20 years instead of 17 makes matters even worse for software patents due to the period of obsolescence being less then the period of the patent.

I should have used the terms "working design" or prototype instead of product. A working prototype is a requirement isn't it? (Otherwise patents could also cover non-working solutions, preventing any solution.)
But I still hold to the idea that software is something that can be copyrighted but is an abstraction rather than an implementation. Don't get me started on copyrights! There, the time period a work is covered gets extended every 20 years, for another 20 years. I heard a speech by a member of Supreme Court railing against the latest extensions. His idea was to make the term of a copyright reflect the time period a published work can obtain around 90% of the revenue. So music might have a shorter term then a photograph, but both would have terms closer to the original. It has already reached the point where by the time the copyright expires, the media storing it has degraded to the point where the work is destroyed and is lost forever. Motion pictures covered by copyright are approaching the point where the film will turn to dust before the copyright expires.

There is a case of a Mathematician whose hobby was large prime numbers. I'm not sure if he had to withdraw two prime numbers from a paper, because they were a violation of patents, or the DMCA. But this was a case where a mathematical paper had to be withdrawn.

The patent prevented exactly that, "the discovery of new methods/tools and improvement of existing methods/tools". You couldn't come up with an improvement. I'll grant that this patent didn't cover a trivial or obvious solution, but if you look at the details of the MS attempt to patent smilies for example, I'm sure you will find countless examples of trivial solutions. The patents being approved of lately have gotten so general that it may be possible to get a patent on all encryption.
Besides, take calculating square roots as an example. There is more than one method. Newton's method, the divide and conquer method based on expanding (A+B)^2, and a method that Russian Farmers would use to calculate them in their heads.
This may be a problem with the legal system rather than the patent system. If a lawyer handles it you are betting the farm on success.

Business models in use even before computers were invented are being translated into software are being patented. Implementing something on a computer that is done on paper should be regarded as trivial. I remember an AP article in the Fargo Forum on this exact problem.
Do you mean that they were able to patent caching and indexing. There is a patent application on using pointers to index dynamic array structures. ( An old lxer article. I did try to find it.)

I'm just wondering if you meant this last part as a joke!

Well, the 20 year patent term doesn't guarantee protection for 20 years. It still has to go through the approval process which could actually give the patent owner an effective patent term less than 17 years just as easily as more than. Either way, like I said, I'm more than prepared to support a reduced term for computer-related patents. So we both agree that this is something that should reasonably be reevaluated.

The requirement of a working prototype, I believe, is unique to the United States system. Even so, it's beginning to lose prominence/importance. The "enabling" aspect of the patent is taking it's place as the most important aspect... perhaps tied with the importance of the patent's claims. Actually, don't hold me to that. I might be getting the working prototype confused with the "best mode" requirement. I'll have to go refresh my memory.

And I agree completely with you. I really dislike the idea of copyright extension. It shouldn't even be considered; we're trying to foster and enrich the material in the public domain. Indefinite renewal of protection runs counter to that principle.

I'm skeptical about that. I'd need to read an article or something about that. I'd be willing to bet that it was more of a publicity stunt by a person to draw attention to what they felt was unfair about the patent/DMCA. I know for a fact you can't patent a number... whether prime or not.

This gets at an issue that bugs me in the debate. Having a patent does not prevent anyone from improving on the idea. The purpose of the patent is this: the government exchanges the protection of the idea for public access to it. That gets back to the "enabling" requirement I mentioned earlier: the patent must be described in such a way as to allow a person skilled in the art to employ the idea without undue experimentation. In other words: tell the public, with clarity, how to get the benefit you want protection for... don't "hide the ball." This is supposed to foster innovation by allowing other to jump ahead of where they might be otherwise. They can then work on, and improve the process or idea and patent the improvement. Yes, individual #2 can patent an improvement on individual #1's original idea. Here's the thing: individual #2 can't produce a product that uses his patent because it would violate individual #1's patent. But the reverse is also true: individual #1 cannot produce a product that uses individual #2's idea without violating individual #2's patent. Some people think "great, deadlock-what happened to innovation?" What usually happens is cross-licensing. Individual #1 grants a license to individual #2 and vice versa. Then both individuals can produce the improved version freely.

Going back to the rate of improvement of the computer industry I would argue it's imperative to enter into cross-licensing. Companies need to constantly improve their products to keep customers coming back. If they don't cross-license, they not only delay releasing an improvement (allowing a competitor the opportunity to steal market share) but they also lose the opportunity to continue improvement/refinement down individual #2's path.

Well, the other side to that contention is that it's not trivial to translate it to computer. If the business model relied on paper, isn't there something to be said about a gain in efficiency by converting it into a soft-copy, allowing unlimited reproduction in a fraction of the time? Whereas the older paper-based model relied on the actual transfer of physical material, included the cost of more paper to make copies for distribution, and the time and effort of an employee to make the copies. In addition, how many ways can the business model be implemented with a computer? Just like the RSA algorithm, there's more than one way to accomplish the same task, and some will be more efficient than others.

I'm honestly not that interested in the business-model side of IP. So I'm not familiar with what factors are involved. They're also a relatively recent addition to the patent system. It used to be that you couldn't patent a business model. I'm not sure which side is "right."

In a sense, yes. But this "caching" has definite, articulable side effects. The information cached in your computer isn't subject to third-party analysis... third-parties that can alter/create legal, financial obligations. They are similar problems, but they exist in entirely different environments, and that gets back to the innovation. We want to reward people for innovating, even existing ideas, to be better suited to a given purpose/role. There is a need for compression. Just because #1 creates the algorithm shouldn't deprive #2 of patent for an improvement on the compression ratio or similarly deprive #3 of a patent for an improvement in the algorithm's speed. Different environments have different needs and require different solutions.

Only partly
Actually, there recently was (or may still be pending) proposed amendments to the patenting procedure. So the concept is on the lawmakers' minds. There's no better time to voice an opinion than when they're dealing with the issue. I believe one of the changes is a change from "first to invent" to "first to file." The current system allows an inventor to invalidate a patent by proving he/she came up with the idea before the person awarded the patent (first to invent). Most of the rest of the world uses the first to file system where the person that files an application first is entitled to the patent regardless of whomever thought of the idea first--a "too bad so sad" approach. I personally prefer our approach, but there are legitimate arguments for the other side.

As a law student, you probably know something I don't.
I *really* don't understand what you mean by "in some form" (only the time aspect as you mentioned above ... ?)

I would fully respect Intellectually Property if it didn't include patents.

All I mean by "in some form" is some combination of benefits/restrictions out there for software/computer-related patents, whether I can articulate them or not, that would earn my approval. Though, I have to admit, I think the system is workable as it is now. In my opinion, it could use some tweaking, but it's workable.

In other words, patents are not black and white. Congress has the power to carve out exceptions or to change the law regarding patents as it sees fit. Modifying the patent term for computer-related patents is one possibility. Maybe computer-related patents only deserve 10 years of protection from the application date. Maybe 5.

However, I'm very much against tying the patent term to any sort of economic "viability" model/speculation. I've heard it suggested in places that the patent term should last only so long as necessary for the inventor to recoup R&D costs. That undercuts the whole purpose of the patent process. The incentive to disclose the invention to the public is profit. By taking that away, there's no incentive, and the inventor may very well choose to keep the invention secret (for an indefinite competitive advantage) and possibly "take it to the grave" -- denying the public valuable information forever.

Compulsory licensing may be another: force inventors to cross-license when one inventor improves on the invention of another - such that the improvement warrants a patent of its own.

Another form of compulsory licensing: if a company/individual makes a product that utilizes a patented concept, but can't afford the requested licensing fee, impose a percentage "cut" for the inventor. Say 95% of the profits that go toward the licensing cost until the license cost is satisfied. Then lather, rinse, repeat when the license comes up for renewal. If the company/individual can never fully pay the licensing cost, they will never get full profit from the product until the patent expires. In theory, if the company can't make enough money on its product to pay for the licensing, it's likely the product isn't a valuable enough improvement to begin with. And if 95% seems high, compare that with the inability to produce the product at all (were it not for this kind of concept).

There can be any number of possibilities. I'm sure someone out there has a very creative solution. It might be worth patenting

Dark_Helmet, you (or others) might not care, but anyway, I want to thank you for the factual and intelligent way you expose your point of view.

Although I don't agree with you, and although I am in the "opposite group", I see that your reasoning is valid. It is a point of view, yours.

Well, honnestly, thank you. I've read so much "lobbying speach" on that subject, that I find it refreshing to read your words.

Actually, you made me wonder: isn't the misunderstanding between the two opposite point of views due to some differences between countries? That is to say, is the patent law the same everywhere?

For example: in France, I'm sure (I've been taught that) that patents do NOT apply, are not allowed to apply, to ideas and commercial processes. In France, a patent MUST apply to an industrial process. This does not concur to what you said, yet you are probably right too, since you were taught that (in the USA). This would imply that patents law are different in different countries. Is that so?

Yves.

That is the intent, but for most Software patents, the idea will be obsolete long before the patent expires, so the public doesn't benefit.

theYinYeti:
Being from France you may have heard of the Paris Convention. Here is a related article on wiki you may find interesting.

http://en.wikipedia.org/wiki/Patent_cooperation_treaty

As I understand it, in Europe, the EC hasn't decided whether to allow software patents at all. I'm sure you are familiar with the debate. Most of the patents files in the US are from countries outside of the U.S. It may be the case that if the US Patent Office accepts a patent, at least for industrial patents, that the U.S. standards are accepted, or at least hold more weight. The devil is probably in the details of the treaties and the laws of each particular country.


I think I got the venting out of my system. It relieved a lot of pent up frustration. So my head is far less likely to start spinning around and explode!

A French company recently won an mp3 patent case against Microsoft for 1.5 billion dollars. How they could have lost the case when they had purchased rights to use it when it was owned before that company bought Lucent Technologies, I don't understand. This is one of those rare occasions when I'm on Microsoft's side.

I have also heard of the following argument against software patents; It is possible for rich software companies to start threatening with lawsuits against hobbyist programmers, if the hobbyist's free alternative to the commercial software is perceived to be a threat to the profits. The independent programmer, lacking the financial means to risk losing a lawsuit, is then forced to kill the project. Even if he has not infringed any patents at all, it doesn't matter. The risk of financial ruin due to legal fees, other costs and a negative verdict (as courts do not always rule fairly) is enough. It can destroy your life. Is that fair? Or is the argument false?