Hacker News Reader

Comment by idealmedtech | Hacker News Reader

[Disclaimer: not a lawyer, this is not legal advice]

I understand they're trying to get rid of cascading attribution requirements for enterprise (which is part of the reason NIH syndrome became so prevalent!), BUT I really dislike this incarnation for two primary reasons:

1. It has conflicting branding. How is a newbie gonna know the difference? How many will misclick it because it sorts first, or because they don't know it's different than true MIT?

2. Some users may read 0 suffix as "original", "optimal", or "first" (0-index), further polluting the branding.

If they rereleased under another name these problems would be alleviated, but I don't think their intentions are entirely altruistic here.

View on HN

Releasing under a different name would have the opposite problem. No one would use it because they wouldn't know that it was the MIT license with the attribution condition removed, even if that was exactly what they were looking for.

The 0- convention is also somewhat standard for this category of licenses (like 0BSD).

Not a problem in actual practice, derivative licenses are commonly identified as such and their slight differences are much easier to explain than the whole of any one original license.

See: AGPL, LGPL for very popular examples of such modifications

Neither of which use -0 suffix, however, which I believe to be GGGP's point.

AGPL adds more constraints, LGPL is probably a slightly better example.

Ha, 0BSD makes me realize they could’ve called it 0MIT, as in “omit attribution.” Would’ve been funny and arguably more understandable.

Submit a PR!

I submitted a PR if they dont accept ill just fork and have my own.

This begs the question, do licenses have licenses?

With "MIT" in the title, this might be a trademark issue.

And to answer your question, yes, for example the GPL starts with:

> Everyone is permitted to copy and distribute verbatim copies of this license document, but changing it is not allowed.

This is actually a brilliant idea, may be you should submit a PR.

The proliferation of BSD licenses is exactly why I use the MIT license for things.

Usually when I see 0 tagged on something, it's actually chronologically after, but something more fundamental.

The 0th law of thermodynamics is the main example

Another good reason to throw this license in the trash and just use the older, better named 0BSD license that is shorter and has the same effect.

For what it's worth, my understanding is that the 0BSD license was similarly not authored at Berkeley [1] and, more confusingly, was derived from the ISC license (not the BSD licenses) [2].

[1] https://en.wikipedia.org/wiki/BSD_licenses#0-clause_license_...

[2] https://opensource.org/licenses/0BSD

Yes, but the spirit of "like the BSD license, but go even further and put no conditions on it" is pretty self-evident from the name, as long as you've heard of 4/3/2/1 clause BSD licenses.

I think that is worth something. Something quite marginal, to be sure. But still a nicety.

That’s correct. Rob Landley and I came up with the same license text but called it different things.

I called it the Free Public License and had it approved by the OSI under that name.

He called it the 0BSD, and argued that since he came up with it first then it should be given his preferred name. I think “0BSD” is an awful name for an ISC-derived license, but eventually I gave up trying to argue with him because I felt that having two names was worse than one bad name.

0BSD has the slight advantage of Google placing it on their pre-approved licenses list, which means Google can use a 0BSD project but not an MIT-0 project: https://opensource.google/documentation/reference/thirdparty...

I have similar concerns about the name being confusing, in the sense that the original MIT license originated from MIT, whereas this license was written by someone else. It might appear as if MIT is endorsing this license despite only serving as an inspiration, and we don't know if MIT actually wants to endorse it.

MIT-0 (zero attribution)

Works pretty well in my head. Don’t think it’s particularly confusing.

MIT-ZA

Rolls off the tongue better too.

And next time somebody asks what's the lowest bar to pass for the code to be shared as free software, you can say:

It's the bar MIT-ZA.

Don't forget to tip your server.

--- 1. It has conflicting branding. How is a newbie gonna know the difference? How many will misclick it because it sorts first, or because they don't know it's different than true MIT? ---

Not a lawyer either and I deeply hate MIT(s) but my favourite concept in law is that ignorance is never a justification and is never used so, if you are using a license, don't be ignorant on that license? Or should we avoid doing things because there is people who wouldn't bother getting documented that could get upset? :o

All things being equal, people will use the license that is easiest for them to understand that it meets their needs.

(One of their needs might be that other people also easily understand their license)

Using familiar “branding” like MIT is arguably a way to communicate easily and efficiently.

When people sell “gluten free pizza” they are banking on the fact that people know what a pizza is.

Since I'm one of the ignorant people, would you educate me and share the reasons for your MIT license distaste ?

If you're a big enterprise with lots of open source dependencies, you have to audit your ENTIRE dependency chain to see which of them are MIT licensed, and include the license text for each one.

It's a pain, so many big shops avoid MIT, and more restrictive licenses like GPL

I'd like the OP to answer, because this argument is pretty much usable for most FOSS licences, not specifically MIT.

Besides, legally, you should know all the licences of all your dependancies, and pulling that out is usually scripted in big shops.

So given the strong aversion demonstrating in the comment, I'm expecting something different.

So you're happy to build your commercial product upon the free works of hundreds of FOSS developers, but only if you can't be arsed to give them attribution? Cry me a river! An enterprise has a responsibility to manage its licenses. Complaining that it's not fun is just childish.

I think also it becomes an issue where you have to preclude a source file with ~10+ copies of the same license for each dependencies. It's the way you have to satisfy the attribution requirement (eg "must include this text," rather than "must include a link to this text somewhere in the source tree") that presents the issue.

It's part of why, as I said in the original comment, lots of big shops do _everything_ in house, even when there's a higher quality open source alternative. They simply don't use projects with attribution requirements that are too strict.

Could be the attribution requirement, which this new MIT-0 removes

It should be 2-clause MIT to follow the pattern of 2,3,4-clause BSD.

> I don't think their intentions are entirely altruistic here.

No part of what you wrote before this in any way suggests anything about the intentions.

Lexicographically, MIT comes before MIT-0.

It's the same 0 as in the CC0 license (which is called out in the description). Zero requirements.

Exactly, but I couldn’t figure of how it actually differs from CC0, despite apparent protestation to that effect:

> The CC0 and various "do what you want" licenses and various public domain dedications may be less attractive to the initial developer for various reasons (i.e., a license is preferable to a public domain dedication).

Maybe the difference is more a branding/tribal association than a legal one?

CC0 is public domain. You waive all copyright (and even moral rights!) in a CC0-licensed work. MIT0 only waives the requirement to retain your copyright notice. It doesn't waive your copyright itself.

CC0 contains language that is more suitable for works of art than computer programs, and is intentionally vague regarding which rights are being waived because the use cases are so diverse. (What does it mean to "perform" or "communicate" a program?) Developers might prefer something a bit more precise.

CC0 is long and complicated [1]. You probably don't want to put that in the header of a gist.

[1] https://creativecommons.org/publicdomain/zero/1.0/legalcode

> MIT0 only waives the requirement to retain your copyright notice. It doesn't waive your copyright itself.

Except that it does. MIT-0 "only" waives the requirement to retain your copyright notice in that that is the change it makes compared to the MIT license. But the MIT license has already waived every other aspect of your copyright. You're not retaining any rights in an MIT-0 licensed work.

The MIT license doesn't waive any aspect of your copyright and/or moral rights. You still claim to have those rights, however diluted they might eventually become. Otherwise it wouldn't make sense to grant some or all of those rights to other people.

CC0 on the other hand goes like "I don't have those rights. Nobody does." It falls back to the rights-granting model only in jurisdictions that don't allow unilateral copyright-waiving.

In practice, of course, the difference usually doesn't matter. In theory, though, there might be an edge case where it matters, perhaps in situations where the original author's moral rights are involved, or if someone discovers a novel use case for computer code that nobody had imagined before. (Am I allowed to bunch up your code into a virtual snowball and throw it at a virtual alien kid while cursing your name in Klingon?)

Right? MIT-NOATT would be much clearer. Make it obvious it's about a single difference. MIT-0 reads too cool/"clever".

It's really not too bad with good tooling that can generate your notice documentation.

Reply on news.ycombinator.com