A New Code License: The MIT, this time with Attribution Required

Question

Update: January 15, 2016

Thank you for your patience and feedback. The changes proposed here have been delayed indefinitely - we'll be back later to open some more discussions.

---

Important context for those arriving from reddit and slashdot links:
The status quo is not "public domain"; attribution is already required.

See:

• Do I have to worry about copyright issues for code posted on Stack Overflow?
• Can we get some explicit clarification on the *intended* legal usage of code from SO answers?

---

TLDR: This is a follow-up to our initial proposal for transitioning to a more user-friendly code license. The purpose of this post is to address the concern expressed most frequently in response to the initial proposal: no attribution requirement. Also, we want to make sure everyone has ample opportunity to provide feedback and we have time to consider it. We are more concerned with doing this right than doing it fast, so please let us know what you think about this proposed change.

A month ago we proposed new licensing terms to cover code posted at Stack Overflow and across the Stack Exchange network. Hundreds of you voted on the proposal, and many of you let us know how we could improve. We gathered all your feedback, and after careful consideration we have amended the proposal.

2 changes:

1. Attribution now continues to be required when you use code found at Stack Overflow and Stack Exchange
2. The changes will now go into effect March 1, 2016 The changes proposed here have been delayed indefinitely.

Change 1 was made to accommodate contributors who want credit, plus to help developers identify the provenance of a Stack Overflow code snippet when they find it integrated into a project. Change 2 was made to allow you more time to digest this change and socialize it within your organizations.

Both changes were thoroughly vetted internally, with our lawyers, and with the OSI. We think they are an improvement upon the previous proposal, and a vast improvement upon the status quo.

As always, community input has been instrumental in moving this initative forward. We're fairly certain we've arrived at the best possible balance for everyone's needs, but if you spot anything new you think we've missed, please let us know.

The Fine Print:

When the change goes live, new contributions across the network will be licensed to the public under the following terms:

• Non-code contributions will continue to be available for use under the terms of CC-BY-SA
• Code contributions will be available for use under the terms of the MIT License
• You don’t have to include the full MIT License in your code base. Contributors agree to give code users permission to ignore the MIT License’s notice preservation requirement, as long as users give reasonable attribution. This optional exception to the MIT License will live in our terms of service.

The only difference between the terms described above and the previous proposal is the removal of the requirement to add attribution only “upon request of the copyright holder (or Stack Exchange on behalf of the copyright holder)”. “Reasonable attribution” is now required by end users of Stack Overflow and Stack Exchange.

What is reasonable attribution?

A URL as a comment in your code is reasonable attribution.

There are certainly other forms of reasonable attribution, depending on use, and you are welcome to go above and beyond what’s required and include username, date, and anything else if you like.

You are also welcome to use the MIT License as it is traditionally interpreted: by preserving the full license with relevant fields (copyright year and copyright holder) completed.

Full guidance will be provided in an upcoming FAQ.

When will the new terms go into effect?

March 1, 2016 The changes proposed here have been delayed indefinitely.

What about other sites on the Stack Exchange network?

The new terms will go into effect for all new code on all sites in the Stack Exchange network.

We understand that some users feel the new terms are not a perfect fit for certain sites, but we think fragmenting the license across the network will lead to ambiguity – the exact problem we’re trying to solve by updating the terms.

What about code contributed prior to this change?

This change is just a first step in establishing clarity for using code found on Stack Overflow and Stack Exchange. Code contributed prior to this change will be governed by the CC BY-SA license as it has always been. We are exploring ways we might distinguish code covered by each license regime, and we hope to soon roll out an opt-in mechanism to let users relicense their old contributions under the new terms.

But what is code?

We will give you guidance on identifying code in an upcoming FAQ, plus guidance on how best to comply with the attribution requirement. But ultimately, identifying code will be a judgement call on your part. We have full faith in your ability to do this.

What’s next?

We want to hear what you think. Barring any showstopper, these terms and a detailed FAQ will be rolled out March 1, 2016.

Answer 1

85% of votes on the proposal were upvotes.

I upvoted the initial proposal because I believe it is important, but I strongly disagree with it. I suspect a lot of other folks did the same...

Your response to the proposal was positive

I'm sorry, this conclusion is wrong and everyone knows that. At least lets be honest and say:

Your response was not positive but we still believe that this is the right thing to do.

I understand your concerns, and I sympathise that you made an effort to hear us, but I still stand by the most voted answer on the initial proposal.

Answer 2

85% of votes on the proposal were upvotes.

Not everyone active on Stack Overflow is active at SE.Meta. For those users, they can upvote (thanks to cross-network bonus) but not downvote (that requires explicit participation on this Meta, not the SO.Meta).

I'm using SO as an example, but it applies to all concrete SE sites due to how "main" Meta works.

So, your votes are necessarily heavily skewed. And yes, I'm posting this after being unable to downvote.

If you want my opinion on how that may be fixed - use highest reputation across the network as SE.Meta reputation and disable Meta reputation changes just like on concrete Meta sites. This is not a complete fix (novice users still can't vote - I assume you value their opinion as well) but an improvement.

---

Not to mention the fact that you used this as an approval rating without asking for such input - it's not the usual semantics for votes on questions, even Meta questions.

Voting up a question or answer signals to the rest of the community that a post is interesting, well-researched, and useful, while voting down a post signals the opposite: that the post contains wrong information, is poorly researched, or fails to communicate information.

Well, now you're getting a much more accurate picture when we know the "rules of the game".

Update: This part of the criticism may be invalid, since (another) FAQ entry expressly says:

Unlike normal Stack Exchange sites, Meta invites the community to discuss, debate and propose changes to the way the community itself behaves, as well as how the software itself works. On posts tagged feature-request, voting indicates agreement or disagreement with the proposed change rather than just the quality or usefulness of the post itself.

This may be a FAQ problem: various parts disagree.

Answer 3

The year is 2018…

… and the economy is in a slump. Pundits on CNN are attributing the decline to a lack of growth in the computing sector. According to Alyssa P. Bitdiddle, senior analyst with the Society of Innovative Computer Programmers,

This has been a disaster in the making since March 1, 2016, a day that will live in infamy. That was the day when lawyers were invited into the software development process. Ever since Stack Overflow Inc. declared that all code snippets should be licensed as code rather than treated as knowledge, there has been a fundamental shift in programming culture. "Intellectual Property" lawyers throughout the industry have had a field day looking for violators. Nearly every software company outside of Somalia, Chad, and North Korea is vulnerable. By my estimates, software development productivity has dropped to pre-Stack Overflow levels. The average developer quality has declined as well, as programmers aren't putting their code up for review anymore. Software companies now spend a majority of their budget on lawyers, and they simply aren't hiring.

She added,

Ironically, this culture of licensing paranoia started with noble intentions. The founders, Jeff Atwood and Joel Spolsky, wanted to promote remixing and reusing content. Creative Commons licensing was used as a pact with the contributors: since the community owned the content, the knowledge base could never be privatized like IMDb, CDDB, or Experts Exchange. Later, concerns were raised that somehow code should not be Creative Commons-licensed, one thing led to another, and now the entire industry is paying dearly. We will never know for sure, but if Stack Overflow had been a normal forum without the Creative Commons license, software developers today would not have to struggle with licensing code snippets — as if the software patent minefield weren't bad enough already!

How did we get there?

Despite a top-voted analysis that said, basically, "Please don't do this", Stack Overflow staff billed the feedback as "highly positive". This ignored the fact that some people upvoted the question because they thought that licensing was an important issue that deserved attention, rather than because they agreed with the proposal.

I think we've learned our lesson this time. This question has been downvoted off the front page of Meta.SE. Is that a clear enough signal? If not, then what would constitute the "showstopper" that you require?

What did people say last time?

Key points from the previous round were:

• What exactly is the problem that the proposal aims to solve?
• We're sharing knowledge, not code.
• Permissive license is inappropriate for questions.
• Permissive license is inappropriate for Code Review, Programming Puzzles and Code Golf, and most sites other than Stack Overflow.
• License fragmentation is an administrative nightmare.
• Administrative overhead of licensing has the potential to break the Creative Commons culture and our Stack Exchange communities.
• What is the status of pre-transition snippets that get edited after the new policy?
• If code that appears on Stack Exchange is treated as licenseable, can we discuss code for which we don't have the right to relicense?
• Is there a threshold of originality before the MIT license kicks in?
• A crayon license is legally unsound, logically inconsistent, and perhaps unenforceable.
• Attribution is important to many of us.
• If you are serious about license enforceability, attribution and transitivity are essential to cover second-generation copies.
• What exactly did OSI say about the proposal, and how did they miss the legal unsoundness that was so obvious to many of us?
• For a permissive license, CC0 or WTFPL may be more appropriate than MIT.
• If there is a second license, then code should be dual-licensed, not permissive-licensed.
• What constitutes "code"?
• Is permissive-licensed code visually distinguishable from the rest of the content?

Out of all those issues, this revised proposal addresses approximately three. Even if this were a good idea (which I think it isn't), it would still be a half-baked idea.

What "code" are we talking about?

@MarcoAurélioDeleu has pointed out the absurdity of requiring attribution for short snippets. If you're using Stack Overflow as a resource to start learning some programming language, it would be as if you, as a tourist, had to give credit to your foreign language phrasebook for every utterance you made. Yes, the phrasebook is copyrighted. No, you are not allowed to reproduce the book. But using the book for its intended purpose is just not a copyright or a licensing issue. If you string together the lookup results for "Hello", "My stomach hurts," and "Where can I find the bathroom?", that's fine. If the book's appendix has a sample cover letter and you use it to apply for a job, that's fine too, even though the appendix is a non-trivial creative work.

Does this licensing proposal apply to the Stack Overflow documentation project? If so, you might as well kill that project now. Who would want to use that documentation resource if it meant you had to cite all the example code from it?

Do reasonable (i.e. non-troll) Stack Overflow contributors actually post original works worthy of licensing? My impression is that Stack Overflow answers are either going to be short enough that the answerer is giving it away as free advice for no-strings-attached usage, or it's substantial enough that it's going to be hosted on GitHub under the author's own terms.

So, what problem, exactly, does the proposal solve? By declaring code as subject to licensing, aren't we just feeding the trolls?

Answer 4

This is a greatly improved step in the right direction. But I still have problems.

---

BEGIN TL;DR

First, I want a complete proposal. I don't want things that aren't provided. I want to see everything laid out before this change happens and with enough time to have community review of it - the full and complete proposed terms of use / terms of service, FAQs, guidance, mockups of user-facing tools. Preferably things in both human readable English as well as the appropriate legalese deemed necessary by lawyers.

Second, I want whatever license is applied to code to be comparable to CC BY-SA 3.0 Unported:

• I want people to be able to copy and redistribute my source code as source code files or in binaries.
• I want people to be able to transform or modify my source code.
• I want my source code to be available for both open and closed source software.
• I want my source code to be available for both non-commercial and commercial software.
• I want visible attribution.
• I want to ensure that it is clear that I may not endorse the use of or user of my Stack Exchange Contributions.

The only aspect of the CC BY-SA license that I don't necessarily care about is Share Alike - I don't care if people distribute my contributions under the same license or not, as long as I am granted visible attribution.

END TL;DR

---

This is an incomplete proposal.

A lot of this information is still not disclosed. Full guidance for reasonable attribution "will be provided in an upcoming FAQ". You are still "exploring ways" to distinguish code under various licenses. You "hope to soon roll out an opt-in mechanism" for relicensing contributions. You don't define what, exactly, code is.

Based on this post, you'll be releasing these other things on March 1st. That's the same day that the license change will be rolled out. That give us 0 time to review everything and find errors or raise concerns.

Please, come back when you have these details ironed out - they are important for me (and I'm sure others) to decide if this is truly a good idea.

As others have pointed out in the comments, you should also learn from Shog9 and how to persuade other people.

---

You leave defining code up to the judgement call of the person using it.

This is not acceptable. There are many, many users on Stack Overflow. I'm sure there are almost as many definitions as to what should be and should not be code. This is a legal matter with copyright and licensing. It should not be a judgement call, but needs to be clearly defined and documented somewhere.

---

Your definition of "reasonable attribution" isn't reasonable.

First, the MIT license requires two things for attribution: the copyright statement and the reproduction of the license. You are requiring one thing: a link to the post. I agree with the idea of making it easier to attribute a post. However, your scheme of attribution has issues.

The proposed method of attribution requires an Internet connection and access to Stack Overflow. I regularly work in areas without access to the Internet. If I were to follow your proposed rules for attribution, if someone viewed the code in this environment, they would not be able to see the name of the author, the date it was posted/modified, or the license that it was released under.

The proposed method also is effective to no attribution in a closed-source project. The MIT license does allow for use in closed-source projects. This is a good thing, and allows my contributions to SO and other SE sites to be more widely used. However, I would not get any visible attribution in these instances. Again, I think it's OK to not require the full text of the MIT license, but perhaps a requirement for a text file that lists the URLs and author names to all code contributions and a one line statement that some of the code base is MIT licenses would be sufficient for me.

I agree with the modified terms, however any attribution should also include the user name of the author of the code as well as the license that it was released under. I don't think the full license is required, but a two or three line attribution is all it would take.

For me, attribution is the singular most important thing to my participation on Stack Overflow and Stack Exchange. If I submit code that meets the criteria for copyright (and therefore can be licensed), I expect nothing short of a visible indication of the use of my work. I fully support allowing a less verbose attribution than the standard MIT license, but I don't support an invisible attribution or an attribution that requires connectivity to Stack Overflow / the Internet to view. If I don't get that, I don't want my code hosted here.

---

There is no protection against using my name to promote a product.

I would recommend looking at the Apache 2.0 or BSD 3-clause license.

The license currently applied to contributions to Stack Exchange has this segment in it:

You may not implicitly or explicitly assert or imply any connection with, sponsorship or endorsement by the Original Author, Licensor and/or Attribution Parties, as appropriate, of You or Your use of the Work, without the separate, express prior written permission of the Original Author, Licensor and/or Attribution Parties.

I think this resolves some of the concerns brought up by Jon Ericson in the comments, and it would make me feel better, too. I do think some modifications would need to be made to not require the full license text to be copyied (and I'd be OK with these modifications).

Answer 5

We want to hear what you think. Barring any showstopper, these terms and a detailed FAQ will be rolled out March 1, 2016.

Please do not do this. The FAQ should be out and we should have some time to go through it before the changes are actually made. You have also stated that a definition for what is considered "code" would come out at the same time

We will give you guidance on identifying code in an upcoming FAQ, plus guidance on how best to comply with the attribution requirement. But ultimately, identifying code will be a judgement call on your part. We have full faith in your ability to do this.

This is another reason to have the FAQ out before the changes are made. We should have the time to go over all this before the changes are made to the license and terms of service.

Answer 6

Meta meta-Meta

For a post that was intended primarily to address one of the concerns raised (repeatedly) in the previous discussion, this sure has kicked up a lot of dust. I got to talking with a few people in chat earlier, and they... Kinda seemed surprised by a few things that I thought were obvious.

And I realized... I've been sorta following this conversation internally for months, and seeing where all the different pieces arose, why the plan has taken the direction it did. But y'all haven't. There's been one previous post, and both that and this kinda make it sound like a done deal and not checkpoints on a long road toward solving a difficult problem.

So I think it's worth noting a few things, for the benefit of the folks who don't hang out in The Tavern:

1. We're working on this because folks have been complaining about it for years. Search the site - you'll find angsty posts about licensing going clear back to '09, full of fear and frustration. That hasn't gotten any better with time; unfortunately, several recent high-profile court decisions have made a bunch of people all sketchy about software copyright law, and monsters under the bed in the form of database companies in particular - nevermind that nothing anyone does on SO is likely to get you sued, seeing huge companies lose cases over literally a few lines of code has a chilling effect on everyone, and uncertainty surrounding the license used here just fuels the fear.

2. There's no hard time limit on getting this done. We've set a deadline so that we're all motivated to work on it, but it's the third deadline so far and there's no reason it can't be bumped again if there's good reason to do so. We're doing this to be good stewards (see #1), but no one's breathing down our neck about it - if it takes another month or another year to get it right, so be it; the important thing is to get it right and only then get it done.

3. The voting on Meta doesn't really matter; voting here is a means of prioritization, not a mandate. It's fun and interesting and useful when it flushes out useful arguments, but don't get all sad if someone downvotes your opinion or mad because you can't downvote theirs - it's far more important that you take the time to express your opinion here in a way that others can understand than it is that you vote; this is a discussion, not a straw-poll.

That out of the way...

Thoughts on licensing

...I've occasionally participated on these sites too, so I feel entitled to chime in with my own opinions here. Especially since I haven't voted yet. There are a few things expressed here (and in the previous discussion) that I think are... Understandable, but naive. So I'm gonna talk about them. I wanna stress, these are my opinions - they haven't been vetted by any lawyer, any president of vice, or any sane person. This goes for just about everything I write here, but on the off-chance that you haven't learned that already, I wanna make double sure ahead of time that you know I'm as likely to be full of it as I am to be right (and that's pretty damn likely).

Licensing does not prevent careless or malicious use

If you're worried about a more permissive license leading to a heyday for students cheating on their homework or careless workers engaging in Stack Overflow-driven Development, then... You're a real optimist. This nonsense was rampant back when I was first getting into this stuff in the late '90s, and no amount of licensing has made a difference that I've seen. Folks who want to cheat will cheerfully lie about where they got the code they're submitting or committing, strip identifying comments, even obfuscate the symbols. I was dealing with this crap from cheap fly-by-night consultants long before Stack Overflow was even a glimmer in Jeff's eye, and I'm sure the old-timers here have been dealing with it a lot longer. Licensing doesn't stop it any more than laws against stealing protect that $20 you carelessly dropped on your front lawn; laws and licenses - like door locks - exist to keep honest people honest, not to rein in the committed dishonest.

Nor does licensing force folks to treat code as examples for writing their own, much as we might hope it would. When I was still working as a programmer, this was something we had to teach every new recruit, sitting them down and working through the process of taking some MSDN snippet and re-writing it with proper error-handling, structure and testing. Often as not this was the first code review. It's a critically important skill, but not an obvious skill - especially when the forbidden fruit seems so sweet. I'd love to see folks come together to create some guidance for folks struggling with this, but a license isn't gonna force it.

An effective license is one that the licensee understands and can comply with

This has been the achilles heel for CC-BY-SA all along; it's trivial to apply for posts, but folks struggle to apply it correctly to code. Chances are, the vast majority of one-liners out there aren't even copyright-able... But how do you know that? When does the attribution requirement kick in, how does that even apply to compiled or server software, what does the "viral" aspect even mean when all that's being "shared" is a binary? These are all questions that any proper software license answers, but which CC-Wiki ignored because it wasn't meant for software.

Common-sense tells you that - when someone asks, "How do I do X?" on SO and someone else responds, "Use this function" - you should expect the latter person wants you to use that function in your code. And so most folks do just that, and don't even think about the licensing. End result? Licensing punishes the conscientious.

So what do we really need here? Something that folks can apply in a common-sense fashion without being dishonest and without violating a license. This is what Sam's been poking away at for months now, the idea that we can make more people honest by making it more obvious what that even means.

I think we're getting closer. I'm not a lawyer, or even a licensing wonk, but on a gut level "tell people where you got this" seems like a reasonable expectation for using stuff that was posted publicly on The Internet.

But ultimately, I think this has to be something that everyone can get comfortable with. And there are still some rough edges, so I'm gonna shut up for a bit and finish reading what the rest of you have had to say about it.

Answer 7

Thanks for listening to the community. I felt there was a huge resistance and that was primarily caused by not listening. So thanks for taking our opinion in consideration now. (I still don't agree, but at least we can have some influence on the outcome)

To talk in a little more detail about the updated proposal:

I think the proposal lacks a very important part: old posts and updated posts. In short: this proposal isn't ready yet! Okay, there will come a FAQ entry in some time, but how things are handled should be discussed now, and not in some time in the future. We all need time to go over it. In my opinion, old posts should have the old license. An indicator of which license is applicable would be very nice then.

Another suggestion regarding the above: a suggestion has been done a few times already to have a 'copy the code' button. If that can be implemented, can you put in a prefabricated comment in there with the required attribution? That would make users actually follow the guidelines you set now.

Another thought: on SO we had some 'Our answer to your questions' post. I think that would be useful for all the high-scoring answers on the previous post, so we have an official and clear statement on the questions posted there.

Answer 8

I get from your comments ("in response to specific concerns about the license not being appropriate") that the current situation might be less than ideal (lots of people think it is) - but I don't really know enough about the legalities behind it to suggest what to do. But I trust SE enough to believe there must be an issue if they want to change something.

I can, however, read feedback. Which it seems you haven't.

This is an important topic - but you say this:

"Both changes were thoroughly vetted internally, with our lawyers, and with the OSI. We think they are an improvement upon the previous proposal, and a vast improvement upon the status quo."

The main thing that is missing is the community being involved. Sure, you've asked us on Meta.SE (the wrong place) and you've posted a banner containing this:

"Thanks, everyone, for your feedback to this proposal. We're going to digest this one over the holidays and should have a follow-up announcement answering your questions and addressing your concerns after the new year."

Highlighting mine.

I assume this is said follow-up - but I don't think it has done either of those points.

You've not answered any questions here - certainly not mine. Most notably, "What is code"?

And the "addressed the concerns" part? Let's check the first post. It had 592 upvotes and 95 down, so 86% of people "agree". It also has 37497 views, so only around 1.8% of people voted on it. Maybe that could be off by a factor of 5. That's not a lot of people.

You seem to be ignoring the concerns I've read. These comments come from the top answers. I'm not cherry picking - the top 12 are negative.

As much as I love SE (I really do admire what you have made), you seem to have ignored this quite a lot. Maybe this is as SE get's bigger, it's harder to involve the community. Maybe the SE format of Upvote and Downvote is not the right way to get feedback on this. And (almost certainly) Meta SE is the wrong audience.

The top answer has 460 upvotes and 23 down. That suggests that 95% of people agree with them. And disagree with you. That's throws the 86% figure out the window, eh?

Sure, something seems to be done. But you're a) Twisting the facts and biasing the vote by posting here, b) Not suggesting the right thing (I don't know what that is) but this is not based on other answers and c) not doing what you said you would - answering questions of addressing concerns.

---

For anyone interested, here is the Word Cloud of the first announcement:

And here is it for this one:

Answer 9

The only justification you've given so far for this change is the following:

...it’s always been a little ambiguous how CC-BY-SA covers code. This has led to uncertainty among conscientious developers... Uncertainty is a drag on productivity...

So, if I understand correctly, this rather significant change is intended to increase developer productivity.

Would you please expand on this, showing, perhaps, actual instances where developers have actually been hampered by the existing licensing scheme? (ie, not theoretical situations, but actual complaints)

Building this change on a foundation which shows the change is needed, necessary, and will provide significantly more benefits to all users than detriments might go a long way toward convincing the "who moved my cheese" users that this is an improvement.

If this is merely a philosophical choice, with no actual basis in objective needs, that's fine, but you're still not explaining the foundation well, and you're going to get a lot more pushback until you express the why more clearly than you are right now.

Answer 10

Another META META site question that impacts a site I have high rep for that I cannot downvote because its moved to Meta Meta. Thanks.

Also this post does not cover the scenario where a SO user advertises a different license or waving of any attribution requirement in their network profile. How is that to be handled or affected?

Answer 11

Last time this was brought up, I posted this answer. This new version still has many of the problems I set out in it.

1. You're still creating a crayon license.

   If you're going to do MIT, do MIT like MIT was meant to be done. The requirements for using code haven't really changed - all you've done here is say "now you have to have a comment in your code". That's less than is required with the current license.

   More to the point, you haven't actually changed anything here. Let's compare last time round:

   You don’t have to include the full MIT License in your code base. Contributors agree to give code users permission to ignore the MIT License’s notice preservation requirement, as long as users give reasonable attribution upon request of the copyright holder (or Stack Exchange on behalf of the contributor). This optional exception to the MIT License will live in our terms of service.

   to this time round:

   You don’t have to include the full MIT License in your code base. Contributors agree to give code users permission to ignore the MIT License’s notice preservation requirement, as long as users give reasonable attribution. This optional exception to the MIT License will live in our terms of service.

   You have essentially removed two words: upon request. The attribution is still not MIT attribution, and is still creating a crayon license.

2. It's still far too easy to claim an illegal use is legal.

   Essentially, someone can get my code from somewhere I use it - perhaps in a commercial open-source product of mine - and not attribute it to me attribute it to me with a single comment. When I chase them down on that (because they aren't complying with standard MIT), they can simply claim "oh, I got it from Stack Overflow", and get away scot-free.

The fact that I'm able to directly copy-paste most of last time's answer to this one should tell you that the changes made here are not enough.

So now I've had my little rant, how do you fix this?

I've already mentioned it in the first point. If you're going to do MIT, do MIT like MIT was meant to be done.

Answer 12

I've had a scenario pop up in my head in discussions in comments, and I kinda want to see how this is expected to be handled.

• Alice answers a question on Stack Overflow, and provides some code for, say, frobnicating widgets in JavaScript.
• Bob, knowing the code is MIT-licensed, re-uses that code in his company's widget portal software. It's all closed-source code, but the MIT license is okay with that. Normally, he'd need a full copy of the MIT license to accompany his code, but Stack Overflow waives that. A quick comment in the JS at the top of the widget.frobnicate() function is all he needs!
• Carol works on the re-write of the widget portal a year later. Part of this includes putting a proper minifier in place in the build process. Carol looks through the whole codebase, doesn't see that one little comment, and moves forward. Unfortunately, the minifier is stripping out comments, so Carol and Bob's company is now shipping code without attribution (I.E. illegally).

The problem isn't limited to JS and minifiers, obviously. Any time you re-use someone else's code that happens to include a Stack Overflow attribution, this can come up.

Legally, the fault in my scenario lies with Carol and/or the company. Code being re-used without checking it with a fine-tooth comb is always a potential source of copyright violations. But I'm really uncomfortable with Stack Overflow, quite possibly the largest source of programming help/resources/etc. in the world is helping to create landmines like this.

Don't get me wrong, I love the idea of not having to include the whole license text when I use stuff from Stack Overflow. But I worry that not having that license requirement is just asking for people to miss the much shorter attribution line, and break the law. And, given that the whole point of this exercise is to try to align the legal situation with reality, I'm worried it's a bit self-defeating on that point.

I've been mulling this over in my head for an hour or two, and I'm still not sure if it's really a big problem or not. So, what do other people (Stack Overflow employees or otherwise) think about this? Is this a problem, or is it merely something people will have to take note of in the future?

Answer 13

This doesn't really answer any of my concerns that started with the original proposal. A comment from the posting mod on my response to the original proposal said, about determining what's code and what's not (emphasis added):

Just to clarify, we never said code is anything inside a code block. It might be up to a user to decide what's code and what isn't, just like he/she may have to make a judgement about whether something crosses the threshold of originality. Ultimately, always, it's on a code user to decide how he/she wants to use what they find here.
– samthebrand♦ 12 hours ago

This was one of the biggest sources on confusion for me, and it seems like it'd be one of the easiest to address. Leaving it open causes a number of issues:

• Inline code, or just code blocks?
• Documentation strings in code?
• Comment blocks in code?
• What about edits that turn one into the other?
• Why on earth would it be up to the user, not the author to make this determination?

I'd like to say that this revision answers some of these questions. It doesn't though. These issues just get punted a little longer, except that a little bit of the confusion is cemented (emphasis added):

But what is code?

We will give you guidance on identifying code in an upcoming FAQ, plus guidance on how best to comply with the attribution requirement. But ultimately, identifying code will be a judgement call on your part. We have full faith in your ability to do this.

This shouldn't be on the user of the code. If I walk into a bookstore and there are free flyers for local events on a table by the door, I don't get to decide that the contents of the bookshelf nearest the door are also for free; the bookstore gets to decide that. It needs to be clear what content is available under what terms.

What’s next?

We want to hear what you think. Barring any showstopper, these terms and a detailed FAQ will be rolled out March 1, 2016.

I'd say that the lack of a detailed FAQ is the showstopper (negative sense).

Answer 14

The changes will now go into effect March 1, 2016

We want to hear what you think. Barring any showstopper, these terms and a detailed FAQ will be rolled out March 1, 2016.

I would like to see the changes be delayed an extra month or 2, just so there will first be feedback on the FAQ that will be rolled out on the 1st of March and any missing information can be filled in so the community fully understands what this means for their post and the usage of code from SE sites and what exactly changes before implementing it.

Answer 15

Update: I'll take back my claim that SE does not have adequate legal support. However, as mentioned in comments, they are yet to provide a justification for their action.

Why must you dictate?

Even on the previous post, there was a comment criticizing your attitude over the whole issue. And again, even on this post, it does not propose March 1 as a starting date, instead it declares that March 1 onwards, we will bring this license.

Was there any significant support for such a license on the previous post? Is there any list of reasons why this license is better than every other license there exists?

Are you capable of deciding?

And lastly, does SE lack the legal support to actually determine the best license (and justify it adequately to the public), instead of relying on the up-votes (which themselves are disputable) of an online crowd? I have a strong feeling that this is so.

I personally have no opinion on the issue (read no clue what's right or wrong), but it does look bad on your part that the SE admin is boldly declaring their intentions only to receive a flood of down-votes. If you cannot decide (or else justify your decisions), then why do you?

Answer 16

If you are going to fix it, fix it properly; do not apply a band aid

I don't support this, but for different reasons to the remaining answers (as far as I can tell). I have a feeling I'm going to get a pile of downvotes, but...

The original proposal gave (almost) the right answer, but you went about it the wrong way, and it was incomplete (for all the reasons set out in almost every answer). The right answer (as far as I'm concerned) is to MIT licence new posts (not MIT with attribution) as you suggested, but with a dual licence for CC-BY-SA, and to leave old posts as they are. Then ensure the licence is actually listed with the code.

What you're now suggesting is MIT with attribution. Really, that doesn't help. MIT with an exception that the licence does not have to be reproduced with attribution is not an OSI approved licence, and we don't need yet another licence around - licence proliferation is a bad thing. If you are going to require attribution, use an OSI approved licence that has an attribution requirement rather than making another one up.

But is attribution really necessary? In my view, the answer is no.

Firstly, Stack Exchange contributions are to share know-how. They are not to share intellectual property. Typically we are talking about small pieces of code - if not, why are you pasting them on Stack Exchange? Whether these are sufficiently large to constitute an independent work protectable by copyright that does not come under a 'fair use' provision is going to turn on the facts, and be a matter that can be debated by lawyers. So if I want to use a little code that's on Stack Exchange, the licence is of no help to me. As Ian Ringrose pointed out, does a call to create a new GUID require attribution or not? Under the existing CC-BY-SA licence, we don't know (because we don't know whether the one line of code is a 'work', and we don't know whether it is within a fair use extension), and under the proposed new licence (MIT with attribution) we also don't know. I'm guessing not for one line. If it's 100 lines, may be yes (but what's 100 lines of code doing there?). Five lines? You are not actually helping your readers.

Let's not assume that this only impacts people writing proprietary software. Even an open source project using an Stack Exchange contribution has just as many licence compatibility issues as under the previous licence (because that open source project is unlikely to be using your mangled MIT licence). You haven't helped.

And you aren't actually helping your contributors. If someone wants to create a new GUID, and I tell that person how to create a new GUID in the most efficient possible way (with a line of code), I want to do just that. I want to teach them, and to make their life better. I do not expect every time they use that technique to reference an article on Stack Exchange to acknowledge that I was the guy who taught them how to create a new GUID (NB it was actually someone else). This is a qualitatively different action from publishing software on GitHub, where (incidentally) I get to choose any licence I like. I don't need the attribution, so don't give it to me. Just like I don't need it to be a GPL type licence (and you didn't give that to me).

The only conceivable result is those using short bits of code will be no more happy with MIT+attribution than they were with CC-BY-SA. I.e. they will either just ignore it, or be in a position where they don't know what the right thing to do is, and possibly rewrite in order to disguise a copy. This helps no one.

Additional (technical) issue

The licence has the text:

"Contributors agree to give code users permission to ignore the MIT License’s notice preservation requirement"

So, if Alice copies Bob's code, Alice has to copy it with an attribution, but may omit the text of the licence licence. Charles then copies Bob's code, and no longer has to acknowledge the original author (Alice), because the license that requires it has been removed. But had Charles copied it directly from Stack Exchange, Charles would have had to do something different. This makes very little sense.

Even better, unpleasant Stack Exchange scraping site B can legitimately copy every post on Stack Exchange, attributing them properly, but not replicating the licence. Unpleasant Stack Exchange scraping site C can then copy scraping site B without attribution. Yuck.

Answer 17

In case no one has said it already: add a licence selector to a posts where code is detected. If there are multiple blocks, give the option to select fine-grained licences for each one of them. Default it to no licence, or a user-configured one.

I personally consider any code I might write in SO public domain, not MIT or a modified MIT. And I certainly do not want code spammed with SO links.

Answer 18

David Thornley has some good insight on this issue which hasn't received much positive attention:

CC-BY-SA is not an all-permissive license like the modern BSD type. It's much closer to the GPL, in that derivations have to be under a share-alike license.

For some programmers, this is fine. I work on internal software, so it really doesn't matter what Free/Open Source licenses we use. All we need is permission to use, modify, and distribute internally, and that's what we do.

On the other hand, some of us make money by selling software in the traditional sense, and CC-BY-SA isn't compatible with that business model. (I've had a couple of jobs like that.) Some of us work for companies with lawyers or managers who don't "get" free/open source software. (I've had some clueless managers.) In this case, anything short of a BSD-type license might frighten them.

Nor does the "excerpt" idea necessarily help. There is, as far as I know, no official lower bound of what is copyrightable, and there is not necessarily a right to use excerpts. In the US, "fair use" is in the copyright law, but again there's no actual definition: it's a judgment call that should consider several things. There's legal dangers with rewriting snippets also, in that it isn't clear what's a derivative work. All of these would potentially have to be decided in a court of law, and that's expensive.

So, there is a very real problem for individuals or companies that sell proprietary software and don't want to be in the position of having to defend what they include in court.

I think it would help if we had a handy BSD-type license we could slap onto code snippets as we wished.

From What is up with the source code license on Stack Overflow?.

Another example where the existing license is shown to be wholly inadequate:

If you are talking about taking code from Stack Overflow for use in your work, I would be very careful. Even if it is the opinion of the people here that posts to Stack Overflow are covered by Creative Commons, you will need to be able to prove that you got the code from here.

A Short Story To Illustrate:

• Someone else decides the code is a good solution for a problem they are facing, and decides to copy it into their open source GnuFizzBuzz project, covered by the GPL.
• Five years from now, someone is doing an open source audit on your code (maybe you are selling to a nervous customer, or your company is going public).
• The open source audit finds the snippet of code you got from Stack Overflow, and recognizes it as originating in GnuFizzBuzz.
• You are now stuck explaining how/why you have GPL code mixed into your commercial product, or proving that the code was actually from Stack Overflow. Can you prove that the person posting the code to Stack Overflow owned it to begin with (maybe they actually copied it from GnuFizzBuzz to begin wih).

It may sound unlikely, but I have seen this exact situation with my own eyes. The safe thing to do is to regard any code snippets you find as explanations of how something could work. Read, understand, and write a solution based on your understanding. Don't ever cut/paste code you found on the Internet if you are working on a commercial product.

(source)

Answer 19

A few thoughts (that may be familiar to those who read my last answer):

1. The CC BY-SA license should apply to the whole post

Making only non-code contributions be licensed under the CC BY-SA is a completely unnecessary complication. Dual-licensing the code will be much smarter.

2. You need to define what "code" is

Some other posters are saying this will be very complicated, but I don't think it will be. If the dual licensing applied only to code in code blocks, then I think that would be a clear and unambiguous definition. It wouldn't catch all code ever, but that's okay. (Yes there would also be some false positives, but I think the simplicity of a definition like this outweighs the downsides.)

3. Don't pretend this is the MIT license!

Don't pretend you're using the MIT license when you're not. If you need a unique license for a unique situation then just use a unique license. While crayon licenses in general are a problem, I think that this the Stack Exchange network is a situation where a new ultra-basic license would be warranted. And I know you'll use your legal team to prevent the problems with other crayon licenses.

4. It would be best to make a new license

I think it would probably be best to write a new license which is very basic and allows reuse as long as a URL of the post is attached to the borrowed code. It may make sense to make that a non-transferable license, so that those who want to redistribute it again must do so only under the full MIT.

5. The license needs to be self contained

You can't have the license, and then list exceptions to it in the terms of service. The terms of service are an agreement between us users and the Stack Overflow corporation, but the post license is an agreement between us and the post authors. Confusing the purposes of a copyright license and a website terms of service is a fundamental mistake.

If a post is contributed under the MIT license then that is its license, and I doubt it would even be legal for the terms of service to give an exception to it. In addition, the license should apply no matter when you find the post, whether it be here, at some kind of mirror, or at archive.org.

6. Each post should show its applicable licenses

Each post should indicate below it what the relevant licenses are. If the user has selected only CC BY-SA or CC0 then it could show it. Old posts would show only CC BY-SA (unless perhaps their author edits them and ticks a box saying "update to the new SE-Attribution license"). Posts edited by multiple authors would show only the compatible licenses: new posts without any user opt-ins would be the same as if they only had one author, but if the user did opt-in to CC BY-SA only, then the posts would be only CC BY-SA.

7. Fix your footer "attribution required" link

It is against the terms of the CC BY-SA license to require specific attribution formats. It is dishonest and disingenuous to keep that link in the footer.

Answer 20

Reactions to this proposal fall into different camps, because there are different kinds of users of the Stack Exchange sites, in particular Stack Overflow. I am not sure this license change will serve all of them. In particular, I believe other mechanisms might be a better solution to the problems this license change is trying to combat.

---

Jonny can't get his code to work, and dumps it on Stack Overflow

To be clear, this already is a problem. If Jonny works on some commercial project, or is contributing to a copyleft-licensed project, it is quite unlikely they can share that code under CC-BY-SA on Stack Exchange. Of course, the real solution is that they create a MVCE specifically for their Stack Overflow question rather than dumping their actual code. But by changing the default code license from CC-BY-SA to the faux-MIT, this problem is worsened. Whereas sharing a short snippet of internal code with a copy-left license might be forgiveable – it's unlikely to be used in other projects given that the copy-left CC-BY-SA is not compatible with proprietary code –, giving everyone a license to do what the f* they want is far more troubling. I would be surprised if a reasonable company would continue to allow their employees to freely ask questions on SO.

The “what about Code Review?” argument also falls into this category. Here, large complete programs are shared to be improved. This is already unlikely to be done with commercial code, so questions are mostly by people learning a new language or otherwise improving their skills in their own time. The CC-BY-SA license is fairly safe here, but I doubt most question askers on Code Review would be fine with giving their programs away under an extremely liberal license: I wrote this program, and all I got was this lousy link.

---

Janine likes helping people, and is fine with her code being used

Then she can already point out all her code is free. This does not necessitate a network-wide license change. Most positive reactions I see to this proposal seem to fall into this category, but it is the least important. If you'd like to make this easier, you could introduce a per-post license selector rather than forcibly relicensing all contributions to this mind set. Such a license selector would merely add formal support for the already common practice of dual-licensing code in posts via a notice in the user profile.

---

Jerome searches for his problem, finds an SO answer, copies the code

While this may be common practice, Jerome is not allowed to do this under the CC-BY-SA license^[1]. He would be well-served by this change. As already pointed out elsewhere, the only safe way to use code found on SO is to do a clean-room re-engineering, or to rely on Fair Use laws or limits to copyrightability, which are not consistent across the world.

I am not convinced that this should be encouraged. Maybe SE is right and this is a futile war on license violations, and the best way forward is to legalize what everyone is allegedly already doing? Or SO could invest in better education explaining what is OK and what is not OK to do. In particular, this would imply moving the focus away from debugging questions back to knowledge questions.

^{[1]: Jerome can only incorporate CC-BY-SA code in his project if (a) the project is never published, or (b) the project is also licensed under CC-BY-SA (see section 4.b). This rules out usage of SE code in most commercial and open-source projects.}

---

Julia likes sharing her knowledge, not her code

This is where I find myself in this discussion. I love writing stellar answers that provide lasting value. In my experience, the best answers on many sites I participate on (Programmers, Code Review, Stack Overflow) do not actually contain much code. And I am fine with people using this knowledge for whatever purpose. Curing Cancer? Coding Porn Sites? I'm fine with that: knowledge can't be copyrighted.

But expression of knowledge can be, and this extends to code examples I might include in my answers. When I write an answer, I want to be attributed. Properly, with not only a link but also my nick, and with a reference to the license so that people that read my writings know their rights (and restrictions).

In particular, I don't want to feel like I'm doing work for free for other people (which is why I stopped answering all those debugging questions on SO). The copy-left CC-BY-SA license does not forbid commercial use, but it makes sure my answers and all parts thereof will continue to stay free. This is in stark contrast to the faux-MIT license suggested by SE, which would allow my contributions to be made non-free.

Changing the license will alienate contributors like Julia or me. If this change comes, I'll think twice before I use runnable code to illustrate my knowledge – maybe, more text or pseudocode would be a better mechanism.

---

Peter doesn't use Stack Overflow, doesn't care

Why is SE complicating licensing for half the SE network that has nothing to do with code? From Seasoned Advice to Super User, this change adds no value to every site that isn't Stack Overflow. Yes, I'm exaggerating. Other sites that might benefit from this change are Game Development, Tex, Wordpress, …, but these are the minority. I understand the desire to limit license fragmentation across the network, but this is a change that should really be applied on a per-site basis, after discussion on that site's own meta to assure that this change is a benefit. SE will already need a per-post license indication because they can't re-license old posts, so a per-site policy would make it even simpler for many users!

---

In summary, I do see the appeal of this change for the majority of passive SO users, but I don't see the value for active SO users, or non-SO users. I'd very strongly rather not see this license change happening. And I'm saddened I had to do this analysis myself. It would have been nicer, less sneaky, for SE to demonstrate up-front how this change would impact different users.

Answer 21

So you are saying that every time anyone writes

Guid optional = new Guid()

they must include a link to this question in a comment?

Just having the possibility of a company getting into legal problems due to their employees not keeping to this, will be enough to get stackoverflow outlawed by lots of legal departments.

Answer 22

I generally agree that the license change is a good idea. I would have personally preferred something even less restrictive, but I can understand why the MIT license was chosen.

One part that is getting a bit lost in the discussion here is how problematic the current license for code is. The Creative Commons licenses were never meant for code which leads to various issues when you use them for code as SE is currently doing:

We recommend against using Creative Commons licenses for software. Instead, we strongly encourage you to use one of the very good software licenses which are already available. We recommend considering licenses made available by the Free Software Foundation or listed as “open source” by the Open Source Initiative.

Unlike software-specific licenses, CC licenses do not contain specific terms about the distribution of source code, which is often important to ensuring the free reuse and modifiability of software. Many software licenses also address patent rights, which are important to software but may not be applicable to other copyrightable works. Additionally, our licenses are currently not compatible with the major software licenses, so it would be difficult to integrate CC-licensed work with other free software. Existing software licenses were designed specifically for use with software and offer a similar set of rights to the Creative Commons licenses.

I'm not a laywer, but because the CC license isn't meant for code it means I'd probably need one to figure out the implications of using CC-BY-SA code in a project. The license is not compatible with all the common open source licenses, and the SA clause is likely to be very problematic in closed source programs.

For me, code in answers is meant to be used. Having a license that can prevent some people from using it is problematic, and something worth fixing in my opinion.

Using a license meant for code makes sense and removes ambiguity. That is why I like the idea in general. There are certainly issues with the current proposal, but I think they're much smaller than the issue we have right now with the existing license.

Answer 23

Just looking at some points you make.

Attribution will be required when you use code found at Stack Overflow and Stack Exchange

I don't get it. Isn't that the case already?

Either I've missed something, or... nothing changes there?

Next one.

We understand that some users feel the new terms are not a perfect fit for certain sites, but we think fragmenting the license across the network will lead to ambiguity – the exact problem we’re trying to solve by updating the terms.

I can be mistaken, but I don't recall different sites of Stack Exchange family to have their content licensed differently. As the footer stays, "user contributions licensed under cc by-sa 3.0 with attribution required". Currently it says so on any SE site. Then the license doesn't appear to be fragmented? There's no problem then, let's keep the old good traditions... But wait, you're now changing the license and it appears to cause problems. Isn't that a sign that something is done not in the right way?

This change is just a first step in establishing clarity for using code found on Stack Overflow and Stack Exchange

As I've mentioned above, the footer says the attribution is required... linking to Jeff's blog post. Anything you mention doesn't appear to contradict with the point of Jeff and hence the rules which we're complying now. In fact, the blog post is being more descriptive regarding the contribution.

Also, we want to make sure everyone has ample opportunity to provide feedback and we have time to consider it. We are more concerned with doing this right than doing it fast, so please let us know what you think about this proposed change.

...

We want to hear what you think. Barring any showstopper, these terms and a detailed FAQ will be rolled out March 1, 2016.

Woops. Doesn't it look like you won't consider the feedback provided there? I mean, possible at first, and evident now, the "showstopper" is the feedback posted here and in the previous thread of yours. This means that asking for the feedback doesn't make any sense... Or probably the question should be reworded (note that some of us have been already tempted to close it as "unclear what you're asking").

---

I'm not discussing the advantages of one license over another, just because I have no idea what it gives. You change the license and the reason of the change seems to be missing - if you'd like to receive more valuable feedback, probably it'd be better to list the FAQ you're mentioning.

Answer 24

The proposed terms are reducible to any license, including CC0 or WTFPL.

For 'code' (which you can't even be bothered to define), the terms you are proposing possess a gaping license-laundering loophole and are legally reducible to CC0 or any other desired license. I'll explain how.

The post above proposes the following new licensing terms.

Starting March 1, 2016, new contributions across the network will be licensed to the public under the following terms:

• Non-code contributions will continue to be available for use under the terms of CC-BY-SA
• Code contributions will be available for use under the terms of the MIT License
• You don’t have to include the full MIT License in your code base. Contributors agree to give code users permission to ignore the MIT License’s notice preservation requirement, as long as users give reasonable attribution. This optional exception to the MIT License will live in our terms of service.

That "MIT License" (which is really the OSI MIT License or the Expat License) states:

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so.

The only restriction the OSI MIT License imposes is the requirement to include a copy of the copyright notice and the permission notice. However, Stack Exchange's proposed terms waive that requirement (see above: "Contributors agree to give code users permission to ignore the MIT License’s notice preservation requirement").

This means that:

1. A first generation derivative work would have to include attribution, but would be able to be distributed under a license not requiring attribution or sharing alike, e.g. CC0 or WTFPL.

2. Anyone receiving such a first derivative work is therefore not required to attribute or to share alike (e.g. in any second derivative works).

3. Anyone will be able to re-license, in this way, any 'code' posted to Stack Exchange sites.

4. 'Code' posted to Stack Exchange sites can therefore ultimately be used in any way a plagiarist derivative author wishes. As such, it is effectively CC0 (or WTFPL, etc) licensed.

This loophole means that your proposed change to Stack Exchange's licensing terms ultimately undermines any desire you (or we!) may have that Stack Exchange contributors should be required to be treated with respect and acknowledgement by users of their contributions.

Please don't go ahead with your proposed change to Stack Exchange's licensing terms.

Answer 25

_{If there has to be a new code license:}

Please license code additionally under CC BY-SA.

So that everything (like posts and threads as a whole) is licensed under CC BY-SA, and code is licensed under the code license in addition.

Answer 26

This whole thing is moot, because you have no idea about the status of anything that is posted to this site.

Simply demanding that it must be MIT licensed doesn't make it MIT licensed.

#include <std/abe-lincoln-dog-leg-tail-quote.h>

People will do things like, "Oh, we solved that problem here at XYZ Inc. with the following code; I'm sure my boss won't mind if I share a few lines ...". Only they won't say that part, but rather only think it as they copy and paste.

And what about all the people who contribute answers to SO, but are doing it on work time, using company equipment? That code could be owned by the company, according to the employment agreement, even if it is invented on the spot and doesn't appear in any company product.

Therefore, everyone in their right mind will not use anything more than a three-line snippet from some random website without paraphrasing it so that it doesn't resemble the original.

Attributing it is the stupidest thing you can do; you're exposing yourself to legal problems when your program clearly reveals the origin of some section of code, and that section of code is under a legal contention (it was posted to a website, but is in fact proprietary).

As a developer using snippets of other people's code of uncertain origin, you have two main responsibilities:

1. Don't do it.

2. If you must do it, make it yours. Paraphrase the code into your own expression. Work with it. Understand every detail as if you had written it yourself. Basically, use the original as an implementation guide. If there is ever the slightest suspicion that it is a knockoff of that original, you will easily be able to play dumb---by playing smart! You know that code inside out and can answer questions about it as if you had written it yourself, and can probably reproduce most of it from memory.

Basically, the honest thing in StackOverflow is to just have a big disclaimer:

We don't really know where any of this code came from, or what is its licensing status; if you incorporate anything verbatim into your product, you implicitly acknowledge that you do so at your own legal and technical risk. Stack Exchange offers no protection from infringment claims arising from third parties, and makes no warranties about the correctness of any code or its suitability for any purpose. Imagine the most draconian legal disclaimer that your eyes have never seen. Now multiply it by two.

That's it! Everything else is superfluous posturing that doesn't mean a thing and won't hold up in court.

Here is something to consider:

When you contribute original code to a GNU Project project, you will be asked to assign the copyright to the Free Software Foundation. Not only that, but you need some affidavit letter from your employer that the code is really yours to give away; that they lay no claim on it. (I have been through this myself, so I know!) That's what then enables the GNU Project to claim that the code is under the GPL, with a fairly certain copyright status. StackExchange neglects to put contributors through anything of this sort (for understandable reasons, such as it putting a huge damper on the party). But that means there is no basis for asserting any fact about the copyright or licensing status of anything that is posted.

Answer 27

We're talking about a somewhat drastic change and I can't even give my voting input on this; this is problematic.

Additionally, I'd like to suggest the FAQ created not be an incomprehensible collection of legal jargon. Many people are just starting with programing in general, let alone licences. To add to that, this is a site where many people's first language is not English.

You must be able to present things in a way most people can understand.

Answer 28

But what is code?

As non-SO user, this is the central question for me here. My main hub of activity is Computer Science SE, and we have quite a lot of maybe-code there.

You say:

We will give you guidance on identifying code in an upcoming FAQ, plus guidance on how best to comply with the attribution requirement.

Before you do this, it is impossible to give feedback. Please provide this FAQ asap.

But ultimately, identifying code will be a judgement call on your part. We have full faith in your ability to do this.

Who is "your"? Authors? I guess. Other regular users? Maybe. Visitors? No way.

Drive-by visitors can on average not be bothered to adhere to even the most basic things prominently put in the FAQ, and you want them to read the ToS? Do they even have to accept them when they to not register? They certainly don't when they just google upon some content. The footer? Good look expressing the rules clearly in that format.

Even if I do read the rules as a visitor (a popup when I hit CTRL+V maybe?), what am I to do? Risk that somebody sues me?

Summary

Case distinction:

1. Most users of SE (visitors and registered users) are not aware of the license-for-code topic. "SE is CC, right?"

   Consequence: the code license is useless.

2. Users are aware, but we don't actually punish violations of the license, because we have to assume it was by accident ("Ah, they probably thought that was not code") and we don't want to be dickheads.

   Consequence: the code license is useless.

3. Users are aware, and we actively try to punish violations of the license whenever we notice them (an event I estimate would have extremely low incidence).

   Consequence: nobody copy-pastes anything from SE anymore because they won't want to open themselves up to litigation. Plus, we have become dickheads, arguably.

Conclusion

Without strict objective guidelines and visual indication about what constitutes code (and thus which license applies), a code license is useless and/or harmful for non-programming sites that deal with any kind of "code" (read: sites on which there is any amount of ambiguity as to what is code and what is not; some SO answers may even be affected).

Please don't.

Answer 29

The core problem that all of this is intended to solve is that people are copying material from Stack Overflow into their code... and that the license for CC-BY-SA isn't ideal for code.

The question behind this problem is "why are they copying the code?"

---

Let's take an old post on SO: Apache POI autoSizeColumn resizes to minimum width. When one looks at the code in the question, it is clear that it is an MCVE. It does stuff, but it clearly isn't actual code. Likewise, the answer presents knowledge of essentially "you need to call this method."

As the code presented in the question was an MCVE and the code in the answer was in the form of knowledge, it should be fairly clear to people that there will be no copyright or licensing issues generated from this Q&A post - no one is copying any code.

However, I believe that this question is the exception to what SO has become. The vast majority of the questions are not MCVEs but rather "here is my actual code from my homework or real code from work" and the answers are snippets of code rather than information about solving the problem.

Question

I have a problem.
Here is my code.

{large block of code}

Thanks.

Answer

Try

{block of code}

Hope this helps.

And here, the licensing of the question and the answer is critical because the answer will be copied directly into the code outside of SO.

No longer is this about transferring knowledge about a problem - it is a code writing and debugging service.

There are many individuals on Stack Overflow who are providing knowledge to people - knowledge about how to fix the problem rather than the fix itself. Yes, this is a gray area. There are some that are clearly on one side of the line, there are others that are clearly on the other side of the line.

I contend that questions that ask for solutions rather than knowledge are becoming by in far the dominant type of questions on SO. And likewise, answers that hand the actual code that can be copy and pasted into the problem code are likely the dominant type of answers on SO.

This isn't the case on most other technology sites, even where code is presented.

Trying to apply this solution to an SO problem across the entire network, while consistent, makes for many headaches on other sites where it is knowledge that is being transferred and any code is demonstrative of that knowledge and not intended to be copy and pasted into a solution.

---

One thing that should be looked at in part on this is an attempt to redirect the culture of asking and answering on SO back to transfer of knowledge rather than 'fix my code'.

If, on the other hand the 'fix my code' with 'here is the answer' is the intended state of Stack Overflow, reexamine how to fix Stack Overflow's Q&A model to better reflect that along with the necessary licensing changes. Don't change everyone else's licensing because people are indiscriminately copying code from SO answers into their code.

Sites that go more into explaining code may need to use material from other projects that isn't compatible as code into CC-BY-SA, but is acceptable under fair use; that the material there is intended to explain - not be copied into other projects. Pulling a method out of some GPL code to show an of mapping something too large to fit into the return type of hashCode() when talking about the pigeonhole principle. Applying the solution to SO's problem that the code in the post is MIT licensed will have a significant negative consequence on the ability to draw upon other resources for these sites to explain and transfer knowledge.

Answer 30

The terms under which users are willing to share their code can vary a lot.

Some don't care about attribution, others consider it mandatory. Some want to release their code with no restrictions whatsoever.

Votes on OP and answers are biased

Votes on the OP or the top answers are biased for two reasons:

• Many users don't have down-vote rights on MSE, so they can't express their disagreement.
• An answer (especially on a matter we aren't experts on: law) can sound very .. accurate, while on the contrary, is completely wrong. I'm not saying most votes are bandwagon-votes, but some of them definitely are.

Find out what users want the right way.

If you truly want to accurately know how many users will be displeased or happy with this change, it might be best to make a survey.

Ask users how strongly they agree or disagree with specific proposals. Also, make sure you take into account that many (most?) users are unfamiliar with licenses and are not lawyers.