3 ms·

Interesting, I always assumed it just meant derivatives had to be MIT as well, but it makes sense. If I made changes to the software, would I just add myself t

Interesting, I always assumed it just meant derivatives had to be MIT as well, but it makes sense. If I made changes to the software, would I just add myself to the Copyright list of authors?

aidenn0 11y ago
That's how I've always done it; not sure if it's correct though.
haylem 11y ago
Derivatives absolutely don't have to be MIT as well. MIT/X11 is not viral. It only requires attribution, but you are free to modify, redistribute and even commercialize as you like, and to re-license your derivative work, as long as you simply give attribution for the original work. But it doesn't mean "claim it as your own", as long as it's not significantly altered.
- jordigh 11y ago
  > re-license your derivative work, No, you cannot relicense. Only the original authors can do that. You can add on top of the existing code stuff of your own under a different license, so that the combination of the two is under both licenses. Practically, it amounts to almost the same, but for example, you cannot remove conditions such as attribution, which only the original authors can do via a license that does not require attribution.
  - haylem 11y ago
    Right, I didn't word that part very well indeed. You still have to retain the original license in any case but it only covers that particular bit of code. Any product you create built on a MIT component can be licensed as you pleased. It depends what was meant by derivative and how this impacts the original code (encapsulated, modified, etc...). I was thinking more of when you use an MIT-licensed lib or 3rd party tool, or when you would reuse a code-snippet licensed as MIT and modify it. License and attributions remain, but only for that part of the code.
jmtulloss 11y ago
You do add yourself to the copyright list. If your contributions are accepted, you are considered an author and the other authors must get your permission before changing the license. This is why a lot of projects (ie clojure) require some sort of contributor agreement so that the original authors can retain control while still accepting contributions.
DannyBee 11y ago
"If I made changes to the software, would I just add myself to the Copyright list of authors? " Here's a super-long answer: You can if you want attribution. the problem with this is that the licenses refer to the "above copyright notice", and most people put the license in headers. So as you might imagine, if you have 50 files, with different contributors, and they all list their names in the appropriate notices so that you have different notices on each file, well then, Congrats, now people using the software have to reproduce 50 notices! This becomes a mess, and it also turns out that other than the license referring to it, the notice is legally pointless. It is cargo cult lawyering to have copyright notices for legal reasons. It isn't required in any berne signatory country (all 168 of them: http://www.wipo.int/treaties/en/ShowResults.jsp?treaty_id=15 http://www.wipo.int/treaties/en/ShowResults.jsp?treaty_id=15). It has basically no legal usefulness[1] As a result, the better practice is to just use something that is legally okay, but more inclusive. This is why when chromium was released, i had them put "The chromium authors" as the copyright owner. It's legally correct/valid, but means nobody has to add stuff to it. [1] In the US, it prevents the innocent infringement defense. But the innocent infringement defense requires that A. you bought a license B. from someone you reasonably believed owned the code because of the incorrect copyright notice Because of the legal standard, I cannot find a single successful innocent infringement defense since Berne.

Interesting, I always assumed it just meant derivatives had to be MIT as well, but it makes sense. If I made changes to the software, would I just add myself to the Copyright list of authors?

aidenn0 11y ago

That's how I've always done it; not sure if it's correct though.

haylem 11y ago

Derivatives absolutely don't have to be MIT as well. MIT/X11 is not viral. It only requires attribution, but you are free to modify, redistribute and even commercialize as you like, and to re-license your derivative work, as long as you simply give attribution for the original work. But it doesn't mean "claim it as your own", as long as it's not significantly altered.

jordigh 11y ago
> re-license your derivative work, No, you cannot relicense. Only the original authors can do that. You can add on top of the existing code stuff of your own under a different license, so that the combination of the two is under both licenses. Practically, it amounts to almost the same, but for example, you cannot remove conditions such as attribution, which only the original authors can do via a license that does not require attribution.
- haylem 11y ago
  Right, I didn't word that part very well indeed. You still have to retain the original license in any case but it only covers that particular bit of code. Any product you create built on a MIT component can be licensed as you pleased. It depends what was meant by derivative and how this impacts the original code (encapsulated, modified, etc...). I was thinking more of when you use an MIT-licensed lib or 3rd party tool, or when you would reuse a code-snippet licensed as MIT and modify it. License and attributions remain, but only for that part of the code.

jmtulloss 11y ago

You do add yourself to the copyright list. If your contributions are accepted, you are considered an author and the other authors must get your permission before changing the license. This is why a lot of projects (ie clojure) require some sort of contributor agreement so that the original authors can retain control while still accepting contributions.

DannyBee 11y ago

"If I made changes to the software, would I just add myself to the Copyright list of authors? " Here's a super-long answer: You can if you want attribution. the problem with this is that the licenses refer to the "above copyright notice", and most people put the license in headers. So as you might imagine, if you have 50 files, with different contributors, and they all list their names in the appropriate notices so that you have different notices on each file, well then, Congrats, now people using the software have to reproduce 50 notices! This becomes a mess, and it also turns out that other than the license referring to it, the notice is legally pointless. It is cargo cult lawyering to have copyright notices for legal reasons. It isn't required in any berne signatory country (all 168 of them: http://www.wipo.int/treaties/en/ShowResults.jsp?treaty_id=15 http://www.wipo.int/treaties/en/ShowResults.jsp?treaty_id=15). It has basically no legal usefulness[1] As a result, the better practice is to just use something that is legally okay, but more inclusive. This is why when chromium was released, i had them put "The chromium authors" as the copyright owner. It's legally correct/valid, but means nobody has to add stuff to it. [1] In the US, it prevents the innocent infringement defense. But the innocent infringement defense requires that A. you bought a license B. from someone you reasonably believed owned the code because of the incorrect copyright notice Because of the legal standard, I cannot find a single successful innocent infringement defense since Berne.