Take the 2-minute tour ×
Programmers Stack Exchange is a question and answer site for professional programmers interested in conceptual questions about software development. It's 100% free, no registration required.

In my application I want to use a mixture of several algorithms described in some recently published scientific papers. But I have some legal questions.

Q1 If an algorithm was published in a scientific paper, and the authors were affiliated with academia not some private company, can it be protected by a patent at the same time?

I read somewhere that it is not possible to obtain a patent for something that was described in details in a publicly available source.

Q2 If I use some algorithm described in a paper do I have to give credit in "legal" or "about" sections in my program or is it just voluntary and would be simply nice of me?

It is not that I don't want to give credit. It is rather that I don't want to reveal what the program is based on internally and would like to hold copy-cats off for as long as possible.

share|improve this question
add comment

6 Answers

up vote 2 down vote accepted

Some academic work is patented, some isn't. There are some situations in which research conducted on a Federal grant can't be patented, I think, and not all work that can legally be patented is.

In the US, you generally have a year to file a patent application after publication, and an application might have been submitted before publication, so all you can say from a publication is that exactly what it describes can't be patented twenty-one years later, so any publication before May 1980 should be fair game as I write this.

You really do need to talk to a lawyer to get the legal aspects right. The lawyer can give you advice on how to deal with possible patents. Don't rely on any advice I give you here.

Giving credit to the author is a good thing for the author, and is polite. In addition, citing a scientific paper can lend some credibility to your work. In any case, cite it in source code as a reference.

share|improve this answer
    
Thanks for your input. Especially thanks for the bit about work under federal grants. But can you clarify these: "In the US, you generally have a year to file a patent application after publication" and "exactly what it describes can't be patented twenty-one years later". It's a bit confusing. –  bor May 18 '11 at 0:09
    
(Remember that IANAL.) As I understand it, in the US a patent is twenty years. You can apply for a patent up to a year after publishing the invention, as long as you can prove it's your invention. (The US is on a first-to-invent system rather than the first-to-file most countries use.) Therefore, if a paper was published, it counts as prior art for anything not filed in a year, and so twenty-one years later nothing can be patented. (Unless, maybe, it's twenty years from granting the patent rather than filing. See a lawyer for reliable answers.) –  David Thornley May 18 '11 at 3:39
    
Ok, thanks now I understand. –  bor May 18 '11 at 4:11
    
I believe, although I'm not an expert, that in the US it's 20 years from grant, and some companies deliberately spin out the approval process for that reason. –  Peter Taylor May 18 '11 at 5:52
    
+1 for the "check with the lawyer" advice. –  Jay Elston Jun 14 '11 at 18:45
add comment

Most academicians will hop to the rooftop with joy when they meet somebody who wants to use their algorithm. A citation of their paper will be very nice (and helpful for their careers).

share|improve this answer
    
Clearly agree for the citation. See it as an "intellectual payment" ;-). –  Agemen May 17 '11 at 7:23
    
I would :). I also agree about citation. –  bor May 17 '11 at 11:21
add comment

Wikipedia has some good info on this.

Generally speaking, if I am understanding the legalese correctly, you can't patent an algorithm, just the practical application of that algorithm.

The easy way to know, is to just send an e-mail to the authors and ask. I've done this when considering using material from a book.

share|improve this answer
2  
+1 always ask the author - they should know (and they may tell you new stuff that no-one else knows...) –  Gary Rowe May 17 '11 at 8:23
    
In the US, algorithms can be patented. That's one reason PNG images appeared on the net: GIF production was under a patent (I think LZW compression), and the patent holders started making moves to collect. It was apparently impossible to create a GIF without infringing on the patent. It isn't technically legal to patent an algorithm, but it's possible to get patents that will ensure that any use of the algorithm is an infringement. If this confuses you, that's understandable. If this is important to you, talk to a lawyer. –  David Thornley May 17 '11 at 14:21
    
@David Thornley, my understanding is that the patent applies to the method in which LZ78 is applied. LZW (as well as LZSS and LZMA) is based on the algorithm, and it (they) are patented. The algorithm (ie. replacing repeated characters with some kind of reference to their replacement) is not. Again, it's a pretty complicated situation which is a bit over my head tbh. It appears to me that the patent infringement surrounds people having used the implementation details from the IEEE article that was written on LZW in their own projects not on using LZ77 or LZ78. Your insight here is appreciated. –  Steve Evers May 17 '11 at 14:50
add comment

q1: Academia protects their patents fiercely. The authors may not care, but their employers care.

q2: ask.

share|improve this answer
    
Thanks for answers. I agree that in case of Q2 it's best to ask the authors. In case of Q1 I can't agree. Universities do not automatically apply for patents for every bit of novelty developed by their employees. –  bor May 17 '11 at 11:09
1  
@bor, the question you asked wasn't "Is all innovation from academia protected by patents?" but "Can innovation from academia be protected by patents?" @cidermonkey is giving you the correct answer (yes, it can be and often is) and some additional warning information on top. –  Peter Taylor May 17 '11 at 12:29
    
Yes, as Peter clarified, it is true that not all unis will do this; but one should never assume that just because it's an academic paper, it is not protected by patents. Also, i'll just leave this here. –  cidermonkey May 17 '11 at 17:49
    
@Peter Taylor - Actually the question was more if publishing and patenting can be mixed :). As others hinted, it can, but is not always possible. I think that universities as employers can apply for a patent, authors of the innovation also can. The question is if the road to patenting is still open after the research was described in detail in a published paper? David Thornley gave as a hint that there can be some patenting issue when the research was publicly funded. The fact that the patent holder will fight for his/her rights (the bigger the fiercer) is rather obvious, don't you think? –  bor May 18 '11 at 0:23
add comment

IANAL

One generally cannot copyright an idea, only the material expression of an idea. In other words, algorithms cannot be copyrighted but implementations can be.

That said, an algorithm may be patentable, but publishing alone is not (nearly) enough to create a patent. The best thing to do is to ask a patent or IP lawyer. The second best thing to do, as SnOrfus said, is to contact the owners of the idea and ask them. The latter might be more expedient given the relative costs but it is not a foolproof substitute for obtaining legal counsel. Of course, obtaining legal counsel and doing The Right Thing is not a foolproof protection against litigation either, so contacting the author is prudent in any event. So: contact the author. If you have any qualms, also talk to a lawyer.

If the algorithm is indeed patented, you will have to do whatever the license you obtain tells you to do. Otherwise, you needn't do anything.

share|improve this answer
1  
Thanks. The question was more if publishing can close the road to patenting. As far as I know if the patent wasn't obtained before publication, the method cannot be patented anymore after the publication. But you're right anyway. The best way to go is to ask some IP lawyer / the author(s). –  bor May 17 '11 at 11:20
    
There's also a certain practice some companies follow: do NOT, under any circumstances, check for a patent. Apparently, in the US, infringing a patent you know about triples the penalties or something, so if you think you might be infringing on a patent, and you think you'll probably get away with it, and you don't think you can feasibly avoid infringing all patents you can find, it may well be worthwhile never to look. –  David Thornley May 17 '11 at 14:18
    
@David Even if true (which is not hard to believe), I would certainly not recommend such a strategy without consulting a lawyer. –  Rein Henrichs May 17 '11 at 16:08
add comment

Anywhere in the World with notable exception of US:

  • algorithms cannot be patented and only literal code is protected by copyright. Thus it's perfectly legal to implement these algorithms.

US, US occupied territories, people traveling to/via US:

  • algorithms also do not fall under copyright, but can be patented, thus you have to contact author.

Legal issue aside, I really think from moral point of view you ought to acknowledge the author of the algorithm in your app's credits.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.