Apache license exists as a permissive open-source license as opposed to MIT/BSD license with the perceived benefit that it also protects authors from patent violations. Although I think I read somewhere the rationale behind it was more to prevent exploitation from patent trolls.

Did Apache come into existence due to a specific reason? Was there a particular case that warranted the evolution of MIT and thus creation of the Apache license? and given this, is it still safe to use MIT/BSD these days, or should something more aligned with Apache be used (assuming I want to release with a permissive, do anything style license)?

Although Apache protects the author from patent violation, what would happen if there was indeed a patent violation?

I'm guessing while the author cannot be 'sued', a cease and desist or equivalent could be issued preventing the inclusion of the violation in future distribution, however I really don't know?

1 Answer 1


The relevant clause from Apache2 is s3:

Subject to the terms and conditions of this License, each Contributor hereby grants to You a ... patent license ... If You institute patent litigation ... then any patent licenses granted to You under this License for that Work shall terminate...

For a work by a single author, so licensed, that clause doesn't protect the author: it protects the users. It tells people who receive the work under that licence that they can use it, free of any fear that the author will turn round and assert an infringement claim against them related to any patents the author may hold. It is true that the licence granted terminates if the user sues the author for violating the user's patent, but that doesn't protect the author from that lawsuit, it only means that that user cannot continue to enjoy his/her licence to the author's patents whilst so suing.

What Apache2 isn't is some kind of special patent sauce that you can sprinkle on a piece of software you just wrote to protect you against later claims that you have infringed anyone else's patents. It doesn't, and cannot, do that.

As such, then, I think the answer to your question is yes, the MIT/BSD/etc. licences do still have a place, depending on the desires of the author. They are simpler than Apache, and if you as author don't have any patents in your name (as, let's face it, most of us don't) you have nothing to licence under Apache s3, so they are functionally equivalent.

  • 4
    This Apache clause does discourage patent enforcement against the project or its users, and therefore provides mild protections against authors as well. Let there be entities A, B which hold patents PA, PB respectively, which are used by a software in question. If A is a contributor to the software, and B is a user, and B enforces PB against the project, then B would no longer be able to use the software due to lacking a license to PA. Doesn't matter for random projects by individuals who don't hold patents, but is hugely relevant for infrastructure-level projects like Kubernetes.
    – amon
    Aug 26, 2021 at 9:01
  • 7
    @amon I'm not saying that the extent to which this would discourage the filing of a lawsuit is of necessity zero, but I do think that it is subjective and hard to quantify, and therefore unsafe for the author to rely on. I also note that the loss of patent licence is limited to "that work", which I interpret in context as "anything received from the defendant under Apache2 that uses the patent over which you are filing suit" - so I find it difficult to imagine an infringement claim that would render unusable something as large (and, I agree, important) as, eg, the whole of kubernetes.
    – MadHatter
    Aug 26, 2021 at 9:09
  • Thank you for these clarifications.
    – lfgtm
    Aug 26, 2021 at 10:36
  • 4
    IMHO this really starts to become relevant in contexts where one or more contributors are large corporations with tons of patents. At that point, you start getting into the realm of patent MAD - nobody can sue any of the contributors, for fear of getting sued in retaliation. The problem, however, is that this does nothing to stop non-practicing entities, who can't get sued for infringement because they don't actually do anything other than threaten people with lawsuits.
    – Kevin
    Aug 26, 2021 at 18:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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