If I have a question about a section of code in a library open sourced under a BSD style license is it alright to copy/paste the section of code relevant to my question onto Stack Overflow?
Almost all BSD licenses say something along the lines of
Redistributions of source code must retain the above copyright notice
If I paste the relevant section of code do I need to copy the license over as well?
BSD is a permissive license. It's only stipulation is that you respect the copyright holder's copyright notice, and that you include the disclaimers. So the intent of the BSD license is to be as unrestrictive as possible. Choosing it means you choose to make your code public for any purpose (short of outright theft), and that you basically understand that your code is going to be copied and used everywhere, which is kind of the whole point.
But the license nevertheless does say that you must include the copyright notice with your redistribution. So what should you do? Well, you should probably do what the license says. However...
By way of illustration, I'm going to pick on Dapper for a moment, Sam Saffron's1 brilliant object-relational mapper that drives the Stack Exchange network. I willingly give Sam permission to slap me upside the head if I get this completely wrong, but here goes.
The heart of Dapper is the SqlMapper class. Everything else in Dapper is window dressing and additional features. Within the SqlMapper class is a bit of code that looks like this:
private static readonly System.Collections.Concurrent.ConcurrentDictionary<Identity, CacheInfo> _queryCache = new System.Collections.Concurrent.ConcurrentDictionary<Identity, CacheInfo>();
private static void SetQueryCache(Identity key, CacheInfo value)
{
if (Interlocked.Increment(ref collect) == COLLECT_PER_ITEMS)
{
CollectCacheGarbage();
}
_queryCache[key] = value;
}
What is this _queryCache they speak of? What is its purpose? Is there really a custom, reference-counting garbage collector in there?
Now of course, I could just link to the SqlMapper class directly, which I will do anyway because I'm a nice person and I want to be polite and respect Sam's work. But isn't it also nice that I included the specific bit of code that I was talking about, and not "look at lines 62 to 70 in the file I linked?"
Should I include a copy of the license in this post, or at least the preamble? (It's the Apache 2 license, so it works a little differently) Probably. Am I going to?
No. Here's why:
If it were my work, I wouldn't expect people to have to include the entire BSD license to point out a few lines of my code in a Stack Overflow question.
Nor would I expect people to ask me for permission. It costs you and me more time and effort to ask and answer that question than it does for me simply to assume you're acting in good faith.
Of course, if you decide to copy the entire SqlMapper class into a blog post without so much as crediting Sam or linking back to his Github repository, I think he would be at least a little bit put off by that.
The purpose of redistribution2 clauses is that, if you're going to write your own program and copy my work into it, then you ought to do the decent human-being thing and include my copyright notice so that I can at least get a bit of recognition instead of claiming all of the glory for yourself, and to preserve the disclaimer protections.
1Sam Saffron is the original author, AFAIK. There are several contributors, including Marc Gravell, Nick Craver, and invalid-email-address, to name just a few.
2The BSD doesn't specifically define the term redistribution, but I think it's generally fair to say that redistribution means "incorporating all or a substantial part of a library or its source code into a new program, thus creating a derived work, and then distributing that derived work to others."
If you redistribute BSD-licensed code and do not follow the terms of the BSD license (i.e., reproduce the required notice(s)), then your use is not permitted under the author's BSD license grant. This is generally true of any copyrighted material under any license, however permissive or restrictive: if you don't comply with the requirements of the copyright holder's license grant, then you don't have permission to do whatever the license grant permits.
In the United States, you may be able to defend your otherwise-unauthorized use as fair use. In legal terms, your unauthorized use constitutes prima facie ("at first sight") infringement, but you affirmatively defend your use as "fair" to a court of law according to the four fair use factors.
Fair use factors are notoriously nebulous and their applicability varies from case to case, but if you are using a small amount of previously-published code for the purpose of learning how it works, you have a pretty good case. You won't know for sure, of course, until the copyright holder initiates a lawsuit against you, the suit goes to trial, and a judge or jury passes judgement about the correctness of your fair use defense.
In interpersonal terms, the idea of the author of a BSD-licensed program taking or threatening legal action over educational use of their code seems absolutely absurd.
In ethical terms, you can avoid the risk of plagiarism by identifying the source of the copyrighted material, so people do not wrongly credit you as the author.
Of course you should include the copyright notice. Since this is a web page, you do that by providing a link to the copyright notice. Like this:
//trivial code
Since the BSD licence includes
Copyright <YEAR> <COPYRIGHT HOLDER>
you must link to the actual licence, not the proforma. And it gives you the copyright notice, which is the BSD requirement.
If that was difficult, you might worry about is it necessary. Since it adds just one link to your question, you should do it even for trivial amounts of code.
I find the entire licensing terms on stack overflow bogus....
I have read multiple OP's remarks on multiple meta posts, all leading to the same conclusion:
1) DO YOU REALLY THINK EVERYONE WHO USES STACKOVERFLOW WILL READ META?
Moreover,
2) DO YOU REALLY THINK EVERY PROGRAMMER KNOWS ABOUT LICENSING?
I'm very sure the first thing people on meta will point out is, it is everyones responsibility to do so. However, the truth is far from it.
I quote Murphy's Law here:
"whatever can go wrong, will go wrong."
Even if Stack OverFlow asks you to accept the licensing terms when you have signed up using a new account. The last thing people do when and they are new to something they need to use for a quick solution to a daily problem is read the terms and conditions.
I don't know why:
Releasing all code ever written on stack overflow henceforth under Unlicense ( or the Public Domain ) is harder than creating a new license. Stay with me here.
What I think should be happening is that:
1) All old questions and answers are still under cc by-sa 3.0 (But carry a small, but distinct footer stating these are copyleft licensed)
2.0) All new questions and answers are under Unlicense ( or Public Domain ), unless otherwise stated.
2.1) A nice addition to this, is if stack overflow could provide licensing options before a question or answer gets posted.
2.2) Furtheremore, the licensing prompt could ask the user posting the question or answer if "code" has been taken from other sources with different licenses. If yes, I'm not sure how to handle this, but it's better than no prompt, that's for sure.
This way, those that are more adjusted to licenses have the ability to add them to protect their intellectual rights ( MIT, cc by sa, or whatever). And those that don't, are still educated with being given an option. Much like github's way of doing things. But instead of the default being all rights reserved. It is Unlicensed.
In essence, I conclude by saying that, it isn't about what people should be doing, it's about what they are doing and how to effectively change the wrong attitude.
These are just my 2 cents on the matter. I would love to hear more opinions in the comments.
