I read an algorithm that is patented, I was inspired in one part of it and applied something similar (not exact) to my algorithm (e.g one step they applied out of their 10 steps algorithm I used in my algorithm not exactly identical but similar notion). Will there be any issue?

Can I simply say, this step was inspired from [ ]

  • 7
    As a side note: AFAIK in Europe algorithms can not be patented, thus patented algorithms from other countries can even commercially be used as far as they are not sold in the said patented countries. Example: Scale Invariant Feature Transform (SIFT). – Ander Biguri May 21 '15 at 11:24

If your question is whether you can cite the patent in a research paper, then yes, absolutely: patents are fully citable sources, and patents don't protect publication of related ideas.

If, on the other hand, you want to use your related algorithm commercially, you'd better get legal advice on that...

| improve this answer | |
  • 8
    The very word "patent" has the archaic meaning of "free to be seen." The point of the patent system was to enable information about new inventions to be readily available for other people to see, while protecting the original inventor's investment of ideas and effort. This means the original inventor can make a profit without having to keep the idea a secret, while everyone else can see how the invention works and use that information to make further advances. – Buzz May 21 '15 at 13:07

If you have a look at

Can an algorithm be patented? (Programmers.SX)

You'll notice that:

  • In many countries, algorithms, and even software, are simply not patentable.
  • Even when software is patentable, it is not the case any abstract algorithm is patentable. See this answer

So even in a commercial context you might still be safe. Also, person X might be inspired by a patent to publish an academic paper, then person Y could implement the paper by X, and that reduces liability even further if it's legit. I would think.

| improve this answer | |
  • Can you clarify "not any abstract algorithm is patentable" - does this mean "no abstract algorithm is patentable"? – Alec Teal May 21 '15 at 10:06
  • 1
    @AlecTeal: Some may be patentable, some are not. I can't clarify because I do not know the details. Luckily where I work it is Not My Job (TM) to worry about whether anything is patentable. – einpoklum May 21 '15 at 10:20
  • 1
    It seems I am one further, I'm in the EU where it is (as you noted) true that software and algorithms are not patentable. So... there! :P – Alec Teal May 21 '15 at 10:22
  • 1
    Lucky for you, you don't live in the land of the 'free' so you don't have to be one of the brave... – einpoklum May 21 '15 at 10:28
  • 1
    @AlecTeal Note that in EU software is not patentable but it's protected by copyright as works of art. Even if the code you see doesn't have a copyright notice and it's publicly accessible the default is that the author retains all the copyright not that it is in the public domain! Only an explicit open source license can allow you to use the code. – Bakuriu May 21 '15 at 19:41

Assuming you have correctly cited the patent, in many countries there is an exemption from patent infringement if the potentially infringing work is academic and/or non-commercial. So, in this case, even if your algorithm does potentially infringe on the patent, you don't need to worry about unless you ever try to commercialise it.

| improve this answer | |
  • I think you should s/many/almost all/ actually. – einpoklum May 21 '15 at 10:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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