I've read the Apache 2.0 license (A2L), in particular its clause regarding patent licensing (see below). I don't quite understand what's the use of it.

If Alice lets Bob use a piece of software under the terms of a permissive license without that clause - doesn't that already imply Bob has the right to " make, use, offer to sell, sell, import, and otherwise transfer the Work"? (I left out "have made", let's ignore retroactive considerations here.) That is covered by the license to "reproduce, prepare Derivative Works of, publicly display, publicly perform, sublicense, and distribute the Work and such Derivative Works in Source or Object form."

If the former is not covered by the latter, than - why not? And what could happen to Bob with the non-A2L license as opposed to the A2L?

PS - If it matters, suppose Alice and Bob do not live in the US.

The relevant clause:

3. Grant of Patent License. Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable (except as stated in this section) patent license to make, have made, use, offer to sell, sell, import, and otherwise transfer the Work, where such license applies only to those patent claims licensable by such Contributor that are necessarily infringed by their Contribution(s) alone or by combination of their Contribution(s) with the Work to which such Contribution(s) was submitted. If You institute patent litigation against any entity (including a cross-claim or counterclaim in a lawsuit) alleging that the Work or a Contribution incorporated within the Work constitutes direct or contributory patent infringement, then any patent licenses granted to You under this License for that Work shall terminate as of the date such litigation is filed.


1 Answer 1


Explicit is better than implicit. It is right that some have argued that the MIT and BSD licenses contain an implicit patent grant. However:

  • These arguments have typically focussed on the US, leaving aside other legal systems that might be less open to finding far-reaching implied terms.
  • AFAIK an implied patent grant is at this point a completely untested legal theory. In the future, the interpretation of whether the license terms include a patent grant will be for courts to decide.
  • This leaves substantial uncertainty for both creators and users of open source software. Such FUD hurts the entire open source ecosystem.
  • An implied patent grant may be revokable or limited. Without explicit license terms, this is simply not clear. Again, a court would have to decide what the implied terms are.
  • As a worst case, an open source software could be a vehicle for a submarine patent that is only enforced after the software has found widespread use.
  • The absence of clear terms gives power to bad actors that abuse their patents and thrive on the surrounding FUD.

For many open source projects, the presence or absence of a patent grant is absolutely irrelevant because no author holds any patents.

However, there are also a lot of open source projects that have institutional contributors from academia or industry, which do have a patent portfolio. The issue of patent licenses is much more urgent for those. The Apache Software Foundation has many such contributors in its projects, and therefore included suitable terms in its flagship license.

The Apache 2 License includes two sets of terms regarding patents:

  • Contributors provide a patent grant to any of their patents that would be infringed by their contributions. Therefore, no bad actor can inject code and later claim that this infringes their patents.
  • If anyone claims that the software infringes their patents, their patent license from (other) contributors ceases. This prevents bad actors from simultaneously enjoying the use of the software while trying to deny others the use. Furthermore, this offers some amount of legal shielding for contributors against claims by unrelated entities.

This explicit handling of patents makes the Apache License very attractive for the implementation of standards that require patent licenses. For example, consider a software that implements some patent-encumbered compression standard. Even if the patents are available under FRAND terms, this effectively excludes open source implementations. Arguably, an open source implementation might be fine but downstream users would be responsible for acquiring licenses (e.g. consider how Android patent licensing works: the hardware manufactorers are typically on the hook for acquiring patent licenses for the devices they sell to consumers).

However, if the patent holders come together to develop an open source reference implementation under the terms of the Apache 2 License, then it is safe to use that implementation for any purpose – while still allowing patent holders to enforce their patents against other implementations.

  • I think you're mystifying things here somewhat. If I do something I have been granted the right to do by the BSD (or MIT) license, then that's that. Or - has any court ever ruled that rights from the patent clause are not covered by the general license-granting clause in AL2/BSD/MIT?
    – einpoklum
    Feb 19, 2019 at 11:54
  • I can't speak for courts, but qualified legal commentators certainly have, eg "Permitting one’s software to be distributed under an OSS license that conveys no patent rights involves neither the selling of a product nor the licensing of a patent" (apparently from Kappos and Harrington, The Truth about OSS-FRAND, from the video here, at 120s in).
    – MadHatter
    Feb 19, 2019 at 12:10
  • @MadHatter: What does that have to do with 3BSD or MIT? These are not "OSS license that convey no patent rights".
    – einpoklum
    Feb 19, 2019 at 12:34
  • 1
    @einpoklum Your question asks about the value of the Apache 2 patent grant. My answer argues that irrespective of whether MIT+BSD contain a sufficient patent grant, the A2L's patent grant is preferable because it is obviously sufficient: at least as good as MIT but without any of the doubts. Given that it is irrelevant for the question asked, the answer does not address whether implied patent licenses exist in all jurisdictions.
    – amon
    Feb 19, 2019 at 14:01
  • 1
    @einpoklum no, you're getting caught up in nomenclature. Read the slide, at least: it's clear from context that what Kappos and Harrington mean by an OSS license that conveys no patent rights is an OSS license with no explicit patent grant. You're welcome to disagree with them; I do, and so does the speaker from whose slide I took the quote (though he disagrees with them on a different but related matter). But you asked if anyone seriously thought that 3BSD/X11 didn't convey a patent grant, and I'm just letting you know that some qualified commentators think exactly that.
    – MadHatter
    Feb 19, 2019 at 14:14

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.