A software researcher (from an American university) published a new method/algorithm and then an implementation of it, licensed under the MIT license. I wrote to him to ask if he had filed a patent or if he intended to do so in the future and he told me that neither he nor the other authors intend to patent the method and that I don't have to worry about it (but I do it anyway).

My doubts are:

  1. if I publish an application that uses this method (and/or part of the code), if one day the author patents the method (or sells the method to some company who then patents it), can I simply get out of trouble by removing the publication of the application or is there some form of retroactive infringement (does it make sense to speak of infringement if no patent existed when I published the software)? If yes, how much time do I have to unpublish my software before I can be sued for infringement (I mean, I can discover a bit later that a patent was filed!)

  2. how do I know if a patent is filed, since it may not have the same name of the publication? That is, how to find it?

Thanks in advance.


2 Answers 2


If they have not filed before the publication, there will be no ability to later file in most of the world. In the U.S. there is a sort-of one year grace period to file after publication.

You will not have a way to definitively know if there is an application pending until it publishes or issues. Normally an application publishes and is open for the public to see 18 months after the first priority filing. That applies to the rest of the world, but in the U.S. it is possible to opt-out of publication. In that case no one can see he application until the day it issues as a granted patent, if ever.

If you are making, selling, offering for sale, using or importing an infringing process or product when a patent issues you could be sued for patent infringement. I do not see how this can be seen a "retroactive".

As a practical matter, if you fold immediately upon the patent issuing it is unlikely you will end up with any liability.

You can set a search in more than one search facility to trigger if a patent is issued to one of the known inventors or use with other criteria but this is not foolproof.


Of course your product could infringe any number of patents unrelated and unbeknownst to the party who published the software. That party not patenting has nothing to do with what others may have filed before they published.

  • 2
    In 35 USC you will find the language "A person shall be entitled to a patent unless . . ." None of the statutory conditions below that are anything like "appear to have dedicated to the public". I am quite confident the MIT license will not stop them from geting a U.S. patent. Enforceablity is outside my expertise and something like that might be a possible equitable defense, but I would be surprised to find that out. Aug 19, 2021 at 0:58
  • 1
    An aside - A few decades ago I ran a technology company that had licensed one generation of our technology to the, then, 2nd largest computer company in the world. As we started on our next generation product they recommended a consultant to us who might have some ideas we could use to improve our design. He exposed us to a very clever idea that we decided to use. Yes, that company subsequently got a patent on that clever idea. I got some advice, held my breath and hoped there would be some equitable defense if they came after us. Fortunately I never found out if there was such a defense. Aug 19, 2021 at 1:11
  • 1
    Actually under the pre-AIA patent law we had 35 USC 102(c) "he has abandoned the invention" as one of the "unlesses". The MIT license might have been seen as abandoning the invention although it actually was intended for the situation of "gave up trying to make it work". In any case that is not part of the post-AIA law. Aug 19, 2021 at 1:25
  • 1
    An inventor can publish an invention and still later patent in the U.S. as long as it is within a year. No other person else can be an inventor of the exact invention discussed in the publication as of the day after the publication. Aug 19, 2021 at 17:39
  • 1
    Patents can be owned by companies but the true inventor must be named and listed on the face of the patent. There are separate search fields for inventor and assignee. Aug 19, 2021 at 18:08

does it make sense to speak of infringement if no patent existed when I published the software?

Not really. The right to enforce the patent arises only starting when it is granted.

If yes, how much time do I have to unpublish my software before I can be sued for infringement (I mean, I can discover a bit later that a patent was filed!)

While there is no good faith or ignorance excuse to direct patent infringement, which can be committed innocently, a patent owner is require to mark a product invention with notice of the patent or give actual notice to an infringer to recover money damages from the infringer (subject to some complicated rules that apply when an infringer is copying another knowing infringer without knowing that the person they are copying is infringing, for example if the author's publication was actually an unacknowledged infringement of someone else's pre-existing patent of the same idea).

So, usually, the opening salvo of a patent infringement case is a cease and desist letter, threatening to bring an action for injunctive relief and to collect money damages for infringements taking place after the cease and desist letter is received.

If you had no prior notice that the patent had been issued, and you cease to use or sell the directly infringing invention once you receive a cease and desist letter from the patent holder or other actual notice of the existence of the patent, while you have technically engaged in direct infringement, there is no meaningful penalty that will flow from that conduct.

As long as you take no affirmative acts to infringe the patent once you learn that it has been issued, you should be good. Actions you took before the patent was issued were, by definition, not infringing.

The inventor's communication to you would be powerful evidence that you not only did not have actual knowledge, but that any potential to get actual knowledge was actively concealed from you, which would probably make the case against awarding money damages against you prior to you obtaining actual knowledge of the patent a strong one.

In addition to direct infringement, indirect or contributory patent infringements also exist.

Contributory patent infringement is selling something non-infringing that is specially made knowing that it will be used in order to commit direct patent infringement. Since this has a knowledge element, one cannot be an innocent contributory patent infringer.

Indirect infringement is basically the patent law equivalent to conspiracy or solicitation in criminal law, when the indirect infringer solicits an infringement by a direct infringer or conspires with a direct infringer, and carried joint and several liability for the direct infringement (whether or not the direct infringer itself is sued). But, if no one involved knows of the patent's existence and ceases to use it immediately upon being told, again, there are no meaningful consequences for the indirect infringement, even though technical indirect infringement occurred.

Further considerations

As noted in another answer, the U.S. patent would have to be applied for within a year of the publication or the public disclosure would make the idea not patentable. The public disclosure would also prevent all patent liability in other countries.

Also, the statement that was made to you sometime after the publication, would only have to be honored for less than a year after it was made to be effective at protecting anyone from a patent being issued.

The statement made to you may also estop the authors of the software from claiming money damages from you at least until you learn that they changed their minds and patented the idea.

If the idea had already been patented by someone else, and was then enforced without notice to you since the initial infringer had notice of the patent, you might even have a claim for fraud against the author who told you that it wouldn't be patented (concealing his knowledge of the patent) for your damages until you could take action to mitigate them (e.g. in the case where they stole the idea and you were directly infringing without notice for that reason in the period before you received a cease and desist notice or legal complaint for patent infringement).

Bottom Line

So long as you observe the limitations of the MIT license, you should feel comfortable that you are not likely to be exposed to meaningful patent lawsuit liability from the group of researchers you contacted without notice.

  • 1
    Thanks for pointing out that a third party may very well have or get a patent that covers whatever product the OP creates. Contrary to your example it could be a third party completely unknown to the American researcher. That does mean your bottom line could benefit from the words "from that American researcher" inserted Aug 19, 2021 at 1:49
  • 1
    The first publication of this method (by the same authors) dates back to 2018 which essentially introduces 2 innovations. Subsequently, in 2021, another major innovation was published by them and added to the overall method. On both occasions, the code was released under the MIT license. The publication is well known in the industry and impressed with the quality of the results obtained. I think the latest innovation is quite unique... for the previous, who knows? But I wonder, if a patent exists, shouldn't the authors have been warned by who else has patented it in these 3 years?
    – AGPX
    Aug 19, 2021 at 17:57
  • 1
    If you are concerned you can have a freedom to operate search professionally done. Aug 19, 2021 at 18:59

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .