The MIT License, Line by Line (2016) | Hacker News Reader

The MIT License, Line by Line (2016)

writing.kemitchell.com

143 pointsTomte20 comments

This part is hilarious and sad:

> The canonical horrible, no good, very bad example of this is the JSON license, an MIT-family license plus “The Software shall be used for Good, not Evil.”. This kind of thing might be “very Crockford”. It is definitely a pain in the ass. Maybe the joke was supposed to be on the lawyers. But they laughed all the way to the bank.

Indeed they did. It's a good reminder to not pull this kind of “very Crockford” stunt and just use a standard, well understood and accepted license.

https://en.wikipedia.org/wiki/Douglas_Crockford#%22Good,_not...

https://news.ycombinator.com/item?id=5138866

I recommend all readers to follow footnote [20] in the Wikipedia article. This is bonus round good.

Direct link for the uninitiated: https://web.archive.org/web/20170722132351/https://dev.hasen...

Sometimes I wish that the basics of copyright law, (in the US: softare patent law as well) and open source lienses were a mandatory part of formal education in field related to computing. This has clearly become essential, but knowledge still seems to be here and there and everywhere.

https://fieldguide.kemitchell.com/

https://oss.kemitchell.com/

Previous discussion: https://news.ycombinator.com/item?id=12559169

I remember reading this article back when it came out, and it made me think a lot about software licenses.

I encourage everyone here to invest some time to read the different libre and open source licenses out there and decide which one would be the best for the software they produce.

The choice of a software license is important and is a political choice.

I recommend not spending too much time researching the license. For copyleft there are not that many choices that are standard enough for people to be comfortable using your code (GPL, LGPL, MPLv2 and maybe some ecosystem specific licenses like the EPL).

For BSD-style licenses there is Apache for large projects (which has your back) and a myriad of others that does not.

> For BSD-style licenses there is Apache for large projects (which has your back) and a myriad of others that does not.

I don't know, the BSDs seem to be doing pretty well without Apache licenses.

They are, but if you start a new project and you want a permissive license you should probably choose Apache 2.0. It grants the same rights as an MIT or BSD license but also includes an explicit patent grant and a section that deals with the licensing of contributions.

As a law prof, I have mixed feelings about this.

- On the one hand, as a piece of legal education, this post is brilliant and lovely.

- On the other hand, I wonder that it is just part of the regrettable tendency in the legal profession as well as the tech industry to take intellectual property as well as funny contractual-esque terms far too seriously.

The exemplar of point (2) is the Crockford clause (along with the discussion in this comment[1]). Nobody in their right mind would litigate this; there's no universe in which "use for good not evil" is actually a legally enforceable constraint. So what harm does it do?

Like seriously. Look at this[2] complaint:

Without a clear definition of “good” and “evil”, people need to seriously consider whether they are safe to use the code – because if the developer and users’ interpretations of the terms differ, there could be hell to pay. Would you want to ship a hardware device – let’s say you’re making a smart TV – and have some developer of a library send his lawyers around to tell you “sorry, TV rots kids’ brains, it’s clearly evil” to get your entire distribution channel shut down by injunction?

No. Just no, no, no. Find me a judge who is going to issue that injunction. Worrying about that is paranoid in a really unhealthy way. But, alas, we lawyers do that all the time, and that kind of worry is contagious, and we've spread it to everyone else, and it's just silly.

I find it really uncomfortable to think that there are people who are seriously like "THIS DOESN'T COUNT AS AN OPEN SOURCE LICENSE BECAUSE IT FORBIDS US FROM DOING EVIL." Like, really? What litigation are you actually imagining? The only way I can ever imagine someone with a Crockford clause actually convincing any court in any non-dysfunctional jurisdiction to award damages for copyright infringement for unauthorized evil use is if the use is so deeply unambiguously evil that copyright liability is by far the least of the unauthorized user's problem. I mean, we're talking about linting the JSON containing the list of the peoples to be genocided level of evil.

And, you know what? If you use JSLint to do a genocide, I hope Crockford does sue you.

[1] https://news.ycombinator.com/item?id=19014052

[2] http://apebox.org/wordpress/rants/456

> actually convincing any court in any non-dysfunctional jurisdiction to award damages for copyright infringement for unauthorized evil use

Is it sane to assume all potential jurisdictions are non-dysfunctional, though?

You are probably right in many respects. But I don't think the objection to the Crockford clause is the worry that he might sue you if you commit genocide.

The objection is that it muddies the waters and costs users money for no good reason.

If a clause like this is so inherently unenforceable, then why put it there in the first place? To make a statement? To annoy people? To make your potential users have to pay their lawyers extra money to find out if they agree that it is something you can safely ignore?

A public company can't just rely on your opinion or mine. They have to do at least some due diligence on an issue like this. And that's money they wouldn't have to spend if the author used a standard license.

This is all fair and true. I'm mostly just lamenting that the world is that way. We have structured our collective existence so that it is unsafe for people to draw common sense conclusions like "of course nobody can successfully sue you for copyright infringement for a do no evil clause." It's a horrible side effect of our creating horrible giant forms of human activity that require massively bureaucratized forms of organization in order to function. It's necessary, but still tragic.

Very well said, and I share your sorrow.

Businesses founded on intellectual property protection tend to take it pretty seriously!

I don't think "Find me a judge!" is the right response. We could count the number of US opinions on breach or infringement under any public software license terms on hands and fingers.

A lot happens before we get to that point. Companies set up elaborate open source license compliance policies and processes to make sure they don't get that far. And not just to control the legal risk. Especially for companies hiring software developers, or benefiting from "open source" in their marketing, making the trade press for trampling over some independent developer's LICENSE file won't go over well.

But if you'd like to focus on doctrine: Are you sure voiding "The Software shall be used for Good, not Evil." for vagueness gets the licensee-defendant off? Maybe for a breach claim. How about the infringement claim? Write one vague sentence in a copyright license, lose your copyright?

Here's a wizard which helps to choose a license: http://beza1e1.tuxen.de/licences/

(You'll need to enable JavaScript, though...)

I think the author significantly misrepresents the argument against copyright assignment. The issue is not the overhead, the issue is control. If you assign copyright to e.g. FSF then FSF has the exclusive right to relicense to whatever they want. Maybe in the case of FSF I trust them not to abuse this power, but in several recent cases this has been used to relicense projects to be effectively closed source---which is almost certainly not what any independent contributor wanted done with their contribution. The DCO is a compromise that retains protections for the project but also avoids stripping contributors of their rights, which is a safety mechanism of sorts against unilateral relicensing on the part of the original developer.

I don't recall covering copyright assignments in this piece. I do recall touching on DCO, but only briefly.

If you'd like my views on contributor license agreements and assignments: https://writing.kemitchell.com/2018/01/06/CLAs-Are-Not-a-Sha...

If you're interested in new ideas on relicensing projects with many contributors: https://writing.kemitchell.com/2018/09/30/Contributor-Counci...

Wow. That was 2016?

Reply on news.ycombinator.com