Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The MIT License, Line by Line

writing.kemitchell.com

Post reply on HN
Print view View original

81–90 of 148 posts

Re: The MIT License, Line by Line

#81

Post by detaro »

fogof wrote: post #76
martincmartin wrote: post #18

Earlier quoted context omitted.

Suppose I sell you an oil filter for your 2019 Mazda. When you get it home, you realize there's a big hole in it, and it can't be used to replace the one currently on your car. You call me up and I say "I never said it would work in a car, I just said it was an oil filter (true) which would fit into a 2019 Mazda (true)." The law says "nice try, you can't be that pedantic. The ordinary purpose of oil filters it to act…

What if the oil filter doesn't have a hole in it, but just has 10,000 miles on it so it doesn't work well anymore. Would selling that be ok? What about a filter with 9,000 miles, or 5,000 miles, or 1,000 miles? Who decides what the cutoff is?

> Who decides what the cutoff is?

ultimately, a judge.

Re: The MIT License, Line by Line

#82

Post by nicoburns »

cyral wrote: post #56
6gvONxR4sf7o wrote: post #49

Earlier quoted context omitted.

Is that minification?

I thought minifiers were supposed to keep licenses (they can detect the license header). I suppose this could get turned off or messed up by accident with the gigantic custom webpack bundlers even stuff like create-react-app ship with.

Lots of libraries don't put the license in the header, only in a separate LICENSE file.

Re: The MIT License, Line by Line

#83

Post by ghaff »

xeeeeeeeeeeenu wrote: post #57
jraph wrote: post #53

Earlier quoted context omitted.

When I read your comment I initially thought "that can't be right". And then I found [1]: > CC0 was not explicitly rejected, but the License Review Committee was unable to reach consensus that it should be approved, and Creative Commons eventually withdrew the application. The most serious of the concerns raised had to do with the effects of clause 4(a), which reads: "No ... patent rights held by Affirmer are waived,…

>I was surprised because my first impression was that this clause does not legally do anything more, anything less than a license that does not explicitly grants the rights to use the related software patents anyway. But I can imagine that without an explicit mention, a judge could still decide that a user receiving a program could reasonably feel allowed to use the related patents from its authors. Yes, many lawyers…

Here's a piece by Scott Peterson (Red Hat legal) that analyzes this with respect to the MIT license specifically: https://opensource.com/article/18/3/patent-grant-mit-license

Re: The MIT License, Line by Line

#84

Post by rectang »

kemitchell wrote: post #73
paxys wrote: post #70

> 171 words every programmer should understand Except there really isn't any agreement on some of the finer details of the license, no matter how much you want to study and understand it. Does the license text have to be included in only the source code or the compiled software? What if the source code of the derived software isn't made public? What is "substantial portions" of the software? Can a MIT project be reli…

I was involved with a group of open licensing lawyers that ended up publishing a "model" permissive license: https://blueoakcouncil.org/license/1.0.0 It's not everything anyone ever wanted, but we think it's a lot closer.

I like the way this license is much clearer about rights and obligations. From the standpoint of software developers being able to grok it, it looks like a strong improvement on the MIT and BSD families.

If you are serious about adoption, I suggest providing an official header to embed in files. Something simple like this:

   // Licensed under https://blueoakcouncil.org/license/1.0.0
Also "instructions on how to use this license in your software project" would be helpful. (Paste the text of the license into a file named "LICENSE", "LICENSE.md", or similar. Add the header as a comment to every file...)

Re: The MIT License, Line by Line

#85

Post by ignoramous »

detaro wrote: post #75
kemitchell wrote: post #67

Earlier quoted context omitted.

I doubt there's any single answer that isn't tautological. In other kinds of development, like embedded programming, notice files are a part of the cycle. Web developers simply haven't developed the habit. Which is a real head scratcher, because as a rule, JavaScript developers use JavaScript package managers, and JavaScript package managers, especially npm, provide good license metadata and auditability. You can hav…

I've seen at least some minified/bundled JS files that neatly preserved the copyright notices for each minified blob they included.

Minifiers and obfuscators are one reason newer licenses like MPLv2 don't require minified JavaScript to carry attribution and copyright notices.

Re: The MIT License, Line by Line

#86

Post by owenmarshall »

detaro wrote: post #81
fogof wrote: post #76

Earlier quoted context omitted.

What if the oil filter doesn't have a hole in it, but just has 10,000 miles on it so it doesn't work well anymore. Would selling that be ok? What about a filter with 9,000 miles, or 5,000 miles, or 1,000 miles? Who decides what the cutoff is?

> Who decides what the cutoff is? ultimately, a judge.

This is where technical people hit a wall with the law.

Many legal systems, especially the American one, aren’t based on strict enumerable lists of what my lawyer friends would call “bright line tests”.

This is probably a good thing, all in all.

Re: The MIT License, Line by Line

#87

Post by detaro »

ignoramous wrote: post #85
detaro wrote: post #75

Earlier quoted context omitted.

I've seen at least some minified/bundled JS files that neatly preserved the copyright notices for each minified blob they included.

Minifiers and obfuscators are one reason newer licenses like MPLv2 don't require minified JavaScript to carry attribution and copyright notices.

MPLv2 does require you to inform the recipient that the software is under MPLv2 and where to find the source code for it.

Re: The MIT License, Line by Line

#88

Post by jacobr1 »

CameronNemo wrote: post #45
sidpatil wrote: post #40

Earlier quoted context omitted.

> I think, as others have pointed out in the replies, 0BSD or MIT-0 is more suitable for me. Why not go with [CC0] in that case? It's a lot more robust, and recommended by the FSF over options like the Unlicense. [CC0] https://creativecommons.org/publicdomain/zero/1.0/

Where is the FSF recommendation? And did they only consider unlicense?

https://www.gnu.org/licenses/license-list.html#CC0

Re: The MIT License, Line by Line

#89

Post by ghaff »

rectang wrote: post #72
ghaff wrote: post #16

Earlier quoted context omitted.

The MIT No Attribution License (MIT-0) is also OSI approved now which is effectively a public domain license. The caveat is that it won't be an approved license at a lot of companies, both because it's new and because many companies shy away from public domain software.

My understanding is that companies are leery of public domain because it means disclaiming copyright . It looks to me as though with MIT-0 you still retain copyright and are just granting an extremely liberal license? That may prove to be a distinction without a difference if it's ever tested in court, but "giving up copyright" is scary to management so it's meaningful in practice.

IANAL but my understanding is you're pretty much right. The problem is that there is no standard established way to disclaim your rights. And, in continental Europe in particular, it's not totally clear that you can disclaim "moral rights" to something you've created.

A lot of this is mostly theoretical legal stuff but lawyers worry about that sort of thing with the result that a lot of companies won't touch public domain software even if they know its provenance.

Re: The MIT License, Line by Line

#90

Post by kemitchell »

rectang wrote: post #84
kemitchell wrote: post #73

Earlier quoted context omitted.

I was involved with a group of open licensing lawyers that ended up publishing a "model" permissive license: https://blueoakcouncil.org/license/1.0.0 It's not everything anyone ever wanted, but we think it's a lot closer.

I like the way this license is much clearer about rights and obligations. From the standpoint of software developers being able to grok it, it looks like a strong improvement on the MIT and BSD families. If you are serious about adoption, I suggest providing an official header to embed in files. Something simple like this: // Licensed under https://blueoakcouncil.org/license/1.0.0 Also "instructions on how to use thi…

Thanks. We didn't originally set out to write a new license, but when we did, we put real time and thought into it. It's nice to see that appreciated.

We've shied away from declaring one "right" way to use the license, because norms, conventions, and expectations vary so much across development communities. What's right and normal for, say, Maven artifacts would seem a bit strange in, say, an npm package or a C library.

Post reply on HN