views:

635

answers:

13

Now that it looks like software patents are going to be severely limited, does anyone have a good argument for keeping them. It seems like copyright law serves software fine and patents just add overhead to what should be an almost frictionless process. Are there any examples of software that wouldn't have been written if not for patents?

+8  A: 

Yes. Though there have been some really stupid patents issued, but that does not invalidate the entire system. The alternative to the patent system is secrecy and stagnation. Think "Order of the FreeMasons" but for software developers - where PKZIP is a jealously-guarded secret of the Order of the ZLIB, and mp3 files don't exist because Fraunhofer could never have recovered their R&D money for inventing them.

The article linked above is primarily about 'business method patents', which are an absolute farce; other articles linked herein indicate that the blowback from these stupid-fest patents will finally put the lid back on the requirements that a software patent actually be for something novel and non-obvious.

granted, many if not most software patents are categorically silly, but that is more due to the defects of the examiners rather than a flaw in the system itself

a patent for 'dual synchronized scrolling list displays' is ridiculous, any first-year CS student could have come up with that solution in less than a six-pack - this fails the "nonobvious" test

a patent for pkzip compression, though, not so obvious maybe

caveat: i have a software patent pending, but of course i think mine is not just novel and nonobvious but also ingeniously clever and useful ;-)

Some cynics assume that anyone in favor of patents must stand to make money from them, but I was in favor of patents long before i ever though i might have one, because the alternative to the patent system is secrecy and stagnation.

Remember, in return for patent protection the inventor must publish the invention so the world can learn from it. Without this incentive, most really useful inventions would remain closely-guarded secrets. The world is free to learn from the invention - which is the really important kind of freedom that advances civilization.

Also, patent protection is for a very limited period of time. The alternative - before patents - was "trade secrets", i.e. jealously guarded secrets about How Things Worked that made progress crawl at a snail's pace, instead of advancing at Net speed.

the original question asked "Are there any examples of software that wouldn't have been written if not for patents?" This is not really the right question, not just because it falls into the unrepeatable-experiments category, but also since without patent protection there would be a lot of software that no one ever heard about.

Steven A. Lowe
I'm sure you can't give details on your patent, but how much is it costing you? I looked into one more than a decade ago, and the various lawyers estimated $10,000 - $50,000, which was $9,999 - $49,999 more than I had on hand.
MusiGenesis
@[MusiGenesis]: we've spent about $25K on the patents for the USA and Europe, which is less expensive than most. Part of the process requires publishing so see http://www.freshpatents.com/Digital-differential-watermark-and-method-dt20070920ptan20070217649.php if you're interested in the details.
Steven A. Lowe
@[MusiGenesis]: and please note that coming up with the idea behind this novel method took a few minutes' conversation, but developing the prototype to prove this novel method took three years out of my life! Ideas are cheap, implementation is expensive ;-)
Steven A. Lowe
Does the patent on software require that your release the full source code for the software, such that when the patent expires, people can actually reproduce your invention?
Kibbee
@[Kibbee]: no. The patent covers the method, not the software. In my case, we will most likely develop a hardware implementation of the software for efficiency.
Steven A. Lowe
@Steven: very interesting stuff, especially if the watermarking survives lossy compression - I really didn't think that was possible. Good luck with it - it's definitely a great example of a legitimate reason for a patent.
MusiGenesis
@Steven: I also agree completely with the "ideas are cheap" sentiment.
MusiGenesis
@MusiGenesis: thanks! it's nice to get validation from a peer!
Steven A. Lowe
Could you explain what is innovative about it? It looks trivial to me.
Stephan Eggermont
@[Stephan Eggermont]: the criteria is "novel", not "innovative" ;-) It's "novel" in that no one else ever thought to do it this way (out of the thousands of related patents). As for it being "trivial", perhaps it is...once you've figured out how to do it - which took 3 years out of my life.
Steven A. Lowe
@[Stephan Eggermont]: The *concept* is "trivial" - coming up with it took about 5 minutes (after a month of research into how mp3s work) but getting it to work took 3 years. If I was an expert in DSP (which I'm not) it might have taken less time - but the "experts" didn't think to do it this way ;-)
Steven A. Lowe
I think that is exactly the problem: DSP experts have their own 'language' in which I think this is not novel (sorry, English is a second language). A patent lawyer is unlikely to be able to recognize that. And I am fully aware that 'trivial' ideas take significant time to implement.
Stephan Eggermont
@[Stephan Eggermont]: The DSP experts I have discussed this with - and the SMEs our patent lawyer uses - disagree. Have you read the patent application, or any of the hundreds of other patents in this field? Many novel things are 'obvious' once someone shows them to you. ;-)
Steven A. Lowe
Your comment about 'dual synchronized scrolling list displays' made me remember one of the first apps I worked on as a pro. My coworker had implemented a 'fake' multi-column listbox by using a fixed-width font and padding out all the 'columns' with extra spaces. I decided this was an awful way to do things, so I replaced it with a whole set of synchronized list boxes, one for each 'column'. In my defense, this *was* VB3.
MusiGenesis
+19  A: 

No. Copyright laws are enough to keep people from stealing/reusing your source code without permission. Anything beyond that makes the world less free. I don't feel like the ideas/process behind software should be protected.

Patents may encourage research and breakthrough in the private sector, but they limit how technology is used, since one company or person 'owns' the concept. The idea may be well known, but others can't use it without permission, thus they are less free. Patents discourage innovation and breakthrough in open source software. The owner of a patent is the only one 'free' to do what he or she wants. Everyone else (the world) is not free to do so.

postfuturist
I think the problem with software patents is that it was used to patent the most simple constructs, concepts and algorithms (from things like URLs and links and sorting algorithms and Ribbon UIs). There has to be laws that prevent the patenting of very basic concepts like these.
Jon Limjap
@[Jon Limjap]: I agree these silly patents contributed greatly to the negative image of software patents and by extension all patents, but there are already laws for this - the patent law! Sadly the USPTO did not employ the brightest bulbs as patent examiners...
Steven A. Lowe
and note that one could make the same suprious argument as above for copyrights, i.e. "copyrights make the world less free"
Steven A. Lowe
I must disagree and downvote. Saying that patents make the other less free is a tautology - that's what patents are for. The question is whether this is useful or not. Obviously, due to patent misuse (as noted in comments above), the disadvantages are currently quite big.
Roman Plášil
+4  A: 

Patents were invented for physical creations, copyright is for textiles. I think the problem is that in the days of early computers the distinction between hardware vs software functionality was blurred so they could get away with it and the tradition has unfortunately stuck.

Yes Fish...
actually, the novel aspect of many patents is the algorithm embodied in the mechanism being patented, and not the mechanical device itself
Steven A. Lowe
Finally someone who understands
WolfmanDragon
And that is wrong, as algorithms are not, and should not be patentable
Stephan Eggermont
@Stephan Eggermont: they are not supposed to be patentable, but they are. Search as patent database for "System and method". (The "system" part is just to satisfy the law, nothing more.)
Zifre
Not in civilized countries :)
Stephan Eggermont
+2  A: 

This is an excellent essay (as usual from Paul Graham): Are Software Patents Evil?

MichaelGG
a most excellent essay, thanks for the link!
Steven A. Lowe
+3  A: 

The problem with not protecting software from being picked apart for people to use as they please is that some of the people (not all) that are behind the great software that we have will stop making it altogether. While open source is great and there are plenty of people who like to create great software either for the pure joy or just because it will be useful to others, a lot of people do it because it's how they make money.

If someone makes something that's new and fresh, why shouldn't they be granted exclusive rights to it for a little while? If they don't have the capital or marketing power of a big company, they'd be hard pressed to make any money off of their work and if there's no protection, a bigger company could just come along and pick it off.

Companies already skirt copyrights and patents as it is, it won't get any better for the little guy if they are suddenly given free reign to snatch up every great concept before the creator is able to benefit from it.

That said, the current system we have (as others above have said) is quite ridiculous and needs a tremendous overhaul.

dhulk
well said. "No bucks, no Buck Rogers"
Steven A. Lowe
+1  A: 

Viewpoints on this issue tend to be based on how much a person stands to gain or lose.

One site that tends towards not liking software patents - and patents in general is techdirt.

And for some possible benefits that software patents might have: John Carroll: The benefits of patents

cofiem
or how much they understand why the patent process was established in the first place. i was in favor of patents long before i ever thought i might get one, because the alternative to patents is secrecy and stagnation
Steven A. Lowe
+3  A: 

I think the biggest problem with patents (and copyright) in a field that moves as fast as computer software is that they simply last too long, to the point of being counterproductive. In an industry where 5 years is an eternity, it seems silly that someone shouldn't be able to build on your technology for 20 years or whatever the limit is after you created it. Software patents might be much more palatable if they lasted for a much shorter time period (maybe 1-3 years), to give the initial creator a chance to make a few extra bucks without stifling others that want to build on that innovation too much.

dsimcha
1-3 years may seem like an eternity on the internet, but it is about how long it takes to get a start-up business off the ground.
Steven A. Lowe
Fair enough. That's just a number I pulled out of my @$$. At any rate, it should be substantially less than 20 years.
dsimcha
it used to be 17 years, and was extended to 20. Twenty years is good, i think - that's essentially one generation.
Steven A. Lowe
20 years is way too much for software. In 20 years that software method is *beyond obsolete*; I vote for 5-7 years.
iconiK
+9  A: 

The original expressed intent of patent law was to create a mechanism whereby individuals would be encouraged to publish details of their inventions and innovations, knowing that they could still profit from the invention through patent enforcement. Without patent protection, inventors could only protect their inventions by keeping the details secret. So the basic rationale of patents is that they provide a good to society as a whole.

I cannot think of one single example in the software world where a software patent actually served this purpose (i.e. the [possibly anticipated] awarding of the patent allowed the original inventor to freely reveal the details of some innovation). Instead, software patents are used to legally badger people who either invented the process independently or are doing something incredibly obvious (like one-click shopping) which violates the critical "non-obvious" requirement of a patent award.

So I do not think there is any value whatsoever in software patents. There could be, if the people awarding the patents were actually knowledgeable about software. They appear not to be.

MusiGenesis
most software patents are indeed worthless and without merit, but not all of them - the patent mechanism is still a good idea, but it does sort of assume qualified patent examiners to make it work!
Steven A. Lowe
Do you know if the patent office is hiring? That would be a pretty fun job. Maybe.
MusiGenesis
@[MusiGenesis]: I don't know, but they should be!
Steven A. Lowe
I don't think qualified patent examiners are a solution. The problem is that patents put the burden to check for patents for every single stupid thing one does to the programmer, and to find workarounds if there is a patent. That is, if patent is not really complicated as, maybe, the MP3 patent(s), so that involuntarily infringing the patent is almost impossible. But being complicated is not an requirement for patents.
hstoerr
+4  A: 

I don't know if I feel that software patents in and of themselves are wrong, but I do feel that our current patent system is beyond broken. For instance, IBM filed 10 patents a day last year. Now IBM is a great and innovative company and everything. But I don't buy that they really invented 10 things a day that needed to be patented last year.

In my opinion, this is a system that inherently benefits the IBMs of the world and not the smaller development shops. To be fair though, companies like IBM tend to regard patents more as a defense against patent trolls (who I feel are the real bad guys in our system) than any offensive measure.

Jason Baker
IBM's policy is to patent everything they invent - as you mentioned, as a defense against patent trolls. With about 400,000 employees it only takes 1 in 100 to have an idea on a daily basis, and there's your 10 patents per day. ;-)
Steven A. Lowe
A: 

I am reviving an old thread as I see, and yet a small question. Is there a written patent somewhere on the concept of Sub Version Control (SVN\ SVC) or Source Control?

Don't use an answer to ask a new question. Ask a new question instead. At the top right ios a button to create a new question.
Mnementh
+1  A: 

To answer the question, didn't the PNG image format get developed in direct response to the GIF patent and the much publicised intention of the rights holder (Unisys) to recover royalty payments from the likes of Compuserve?

andora
Correct. It was Unisys patenting the LZW algorithm, which is used as the data compression algorithm in GIF formatted images, that spurred the creation of PNG. More information about the algorithm: http://en.wikipedia.org/wiki/Lempel%E2%80%93Ziv%E2%80%93Welch
joseph.ferris
Moreover, that's an example of the patent system working correctly: it stimulated innovation that advanced the state-of-the-art.
Donal Fellows
+2  A: 

The patent difference between software and other technologies:

I think software patents should not be allowed. The big difference between software and other technologies is that it consists of extremely many ideas, and almost none of these has any significant development cost that needs to be recovered.

I can see the necessity of patents in industries where you need to invest millions of dollars to create new ideas, which could then easily copied by other companies. or instance if medical drugs could not get patents, nobody might go through the immensely expensive processes of finding new ones.

In software, however, it is a rare exception that generating a single innovative idea here is very expensive, so patents are not necessary to ensure innovation. Its value lies in the well executed composition of elements, which is not patentable, anyway - the appropriate way to protect software investments are copyright laws.

So the primary business case for software patents is to go after competitors, which is not good for competition and a serious latent threat to innovative small businesses who do not have the ressources to check for every small detail whether someone snatched a patent one needs to work around.

This comment on serverside about the new zealand legislation banning software patents has also a very good argumentation.

hstoerr
+1  A: 

In theory patent provide a incentive for inventors. In practice it doesn't. For example the Berkley patent survey of inventors found that they provide a very weak incentive.

There are also plenty of studies showing that the patent system holds back innovation. (I'm a new user, so I can't post more than one link, but google "Yet Another Study Finds Patents Do Not Encourage Innovation")

wassname