MIT-0 License
github.com
github.com
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.
The 0- convention is also somewhat standard for this category of licenses (like 0BSD).
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.
The 0th law of thermodynamics is the main example
[1] https://en.wikipedia.org/wiki/BSD_licenses#0-clause_license_...
I think that is worth something. Something quite marginal, to be sure. But still a nicety.
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.
Works pretty well in my head. Don’t think it’s particularly confusing.
Rolls off the tongue better too.
It's the bar MIT-ZA.
No part of what you wrote before this in any way suggests anything about the intentions.
> 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 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
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.
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?)
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
(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.
It's a pain, so many big shops avoid MIT, and more restrictive licenses like GPL
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.
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.
Like the MIT license but you must put 2 instances somewhere in your attribution. It is per-license, so the whoever includes your work in their own project must put 4 instances, and so on.
Also you'd have to sum up the entire series, since the licenses at every level are included in the folder, not just the top-level one.
Nevermind, I see your point! In reality it would be something relating to N sums of 2^k, k being highest depth dependency for each of the N projects. Obviously we're talking bounds and not exact calculations here.
I mean... yes it would, but isn't teaching proper care and attention of licensing important?
If it's not the developer's original code then they ought to think at least a few minutes about where it came from and whether they have consent to use it and all that nuisance.
This license has proven useful for code that is intended for developers to use as reference, teaching samples, examples, or templates that other developers may modify for their own purposes.
I don't think teaching proper care and attention of licensing is incompatible with wanting to release teaching materials under a licence that does not require attribution.
Which was the usecase for MIT-0 that the previous commenter seemed to hint at.
The longer answer is that CC0 is not a licence – it's a public domain declaration. With CC0, you relinquish all rights to the work, including copyright. With MIT-0, you still hold the copyright; you're just allowing people to do whatever they want with the work.
Another difference is the one you alluded to. CC0 is designed for written content, images, videos, etc. It's not really designed for code/software. That's not to say you can't use it for software, however. MIT-0, on the other hand, is designed for code.
So, in essence, these two legal instruments are in a sense equivalent, but there is enough nuance there that it could make a difference in a court of law, depending on how they are interpreted.
There's also the fact that CC0 is quite a mature piece of text at this point, whereas MIT-0 is the new kid on the block. That might be something to consider, too.
Also, obligatory disclaimer that I'm not a lawyer and that this does not constitute legal advice.
With CC0, you relinquish your copyright only, but explicitly retain trademark and patent rights (!).
> 4. Limitations and Disclaimers.
> a. No trademark or patent rights held by Affirmer are waived, abandoned, surrendered, licensed or otherwise affected by this document.
This makes it sort of iffy for software when software patents exist unless coupled with some kind of patent waiver or assertion of not holding and not being aware of patents. While you can argue with MIT-0/0BSD/whatever else there might be that there is an implied patent license, there is an explicit non-license for patents with CC0.
Then in turn downstream users can legitimately obtain it without knowing that you're attempting to disclaim liability.
Consider some libfoo, and in libfoo you include example.mylang as a "quickstart" for the common use case, and many people use this as a starting point, modifying it a bit to suit their needs but not much. Technically you should attribute this file/function/class too, but in practice no one does, and no one really cares either.
I've had people ask me to put a license on 2-line Stack Overflow licenses (it's already CC-BY, but people still ask for some reason, I guess to comply with company legal dept.)
It's useful for stuff like that: you typically don't really care what someone does with your fairly simple examples/demos/etc. (which may not even meet the threshold of originality to apply for copyright protection) and whether they attribute it.
Whether something like this is even copyrightable is questionable. But releasing it under a license like MIT-0 removes any ambiguity -- whether it's copyrightable or not, the original author is allowing you to use it for anything at all, no strings attached.
Can I make my own license and call it MIT-2, to make my license sound much more established and prestigious than it really is? Or is this a case of "it's OK because Amazon does it and nobody will sue them for it"?
Software licenses don't just fall from the heaven and get adopted on their own.
Absolutely not.
Public domain code, I can do whatever I want with it, even publish it with all indications of its origins removed. Legally fine. But it seems like the same thing here-for a downstream consumer at least.
From the original developer’s viewpoint - the warranty disclaimer may give some additional legal protection, but you can put a warranty disclaimer in a public domain dedication too. People will debate whether warranty disclaimers in public domain dedications have the same legal force as those in copyright licenses-but maybe that too is a theoretical debate, because how big is the risk of someone being successfully sued by a code reuser who has zero commercial relationship (support or consulting contracts, etc) with the original developer-has that ever actually happened? Or are all these warranty disclaimers just lawyers saying “it’s never happened, but we can’t totally rule out the remote possibility that one day it might, so let’s just include some legalese to improve our position if anyone ever tries to make it happen - a disclaimer doesn’t cost anything significant anyway”?
Releasing under an extremely permissive license does not relinquish copyright. Any future derived work by the author would still be unambiguously licensable under any conditions they see fit.
Isn't that just the normal case of a derived work? If you modify a public domain work, you don't have copyright on the parts which are identical to the public domain original, but you do have it on any non-trivial modifications/additions, and you can license the later (whether under an open source or proprietary license).
A real world example is the Ghidra reverse engineering suite - it was originally written by US government employees in the course of their employment, and as such that original code is ineligible for copyright in the United States and automatically in the public domain in the US (although its legal status in non-US jurisdictions is rather more complex, variable, and at times even unsettled), and hence cannot be released under an open source license (or any other copyright license for that matter). But, they also accept outside contributions, which generally would be under copyright (whether by the contributor or their employer), and it is a condition of accepting those contributions that the contributor agrees to license them under the Apache 2 license. Hence, Ghidra is an amalgam of public domain code (or at least, public domain in the US) and copyrighted code licensed under Apache 2. I'm not aware this causes any problems in practice.
That seems unlikely to be true. Recipes can't be copyrighted because they're functional. Legal licenses are also functional.
(What's that? You think source code is functional? Shows what you know.)
Copyright © 2007 Free Software Foundation, Inc. <https://fsf.org/>
Everyone is permitted to copy and distribute verbatim copies of this license document, but changing it is not allowed.Never before have I been so motivated to write open source software.
For an open source evangelist, this is an odd way to engage the community. Really makes the FSF's principles look good, with the four freedoms and such.
MIT No Attribution is included in the comparison.
Are we that easy to fool by lawyers? The public domain is the public domain. Before someone comes out and says that the concept of dedicating IP to the public domain doesn't exist in some countries or is full of hidden perils, I'll ask you to provide a single case where code was attributed to the public domain and that became a problem for the author or for the users.
“I place this in the public domain” has a significantly different meaning in British English.
So if you say “the facts are in the public domain” in BrE, that might mean that they’ve been published in a newspaper. It doesn’t mean anything about their copyright status.
Personally a permissive license probably makes more sense but I appreciate someone saying they don’t want to have anything to do with this copyright stuff.
If they're being responsible, they only have to note the provenance of the code in version control. After that, they can basically treat it like their own.
Pretty much any open source license would be that, the main difference this one provides is that Amazon wont even have to acknowledge you worked for them for free :-P
So 0BSD is to MIT-0 as ISC is to MIT. Just a less verbose alternative.
StackExchange answer summarizing the OSI mailing listing discussion - https://softwareengineering.stackexchange.com/questions/1471...
The discussion itself - https://web.archive.org/web/20170301020915/https://lists.ope...
In practice, the wording is much more legally problematic in the WTFPL. And licenses such as MIT-0 or 0BSD contain liability waivers that are not present in the WTFPL
Sticking it to users who care about compliance is part of the appeal of licenses like the WTFPL, but it's ironic that the money flows to lawyers.
When the ASF finally banned the JSON license for use in dependencies, the legal cost of figuring out compliance was one of the motivations cited by the people who proposed the ban.
There's a good reason why GPL and contracts for renting an apartment, buying a car etc are so long: it's to reduce ambiguity.
- reused from previous rental terms, incorrectly restating basic local laws which have since been revised
- downloaded from some website, with terms which are obviously illegal locally
These contracts just introduce more ambiguity. They really should just reference the relevant legal codes. And at least in places with good protections for renters, that should maybe be mandated so the actual lease document is reduced to meaningful terms the renter needs to know.
[1] https://tsdr.uspto.gov/#caseNumber=73797219&caseType=SERIAL_...
Who says that and why? Why should I care about the complicated terms of a license - especially when such a license expresses essentially the same as "public domain"?
Or in other words: is there any real rationale behind the above statement?
I thought it was something released recently but saw on Github that it's from 2018.