13

Some framing around content licensing...

As a content creator who contributes to the data repository that is Stack Overflow/Stack Exchange, authors maintain ownership of their works, even though their contributions are freely available on the internet. Those contributions are licensed at the time of publishing in two ways (and, for the sake of this question, let us assume these two licenses both exist, as some folks question the existence of dual licensing):

  • Authors grant a Creative Commons license (slightly different variants, depending on the time of contribution, but let's just call it all CC BY-SA for simplicity). This is a blanket license that applies to anyone and everyone in the world.
  • Contributions are also "perpetually and irrevocably licensed to Stack Overflow on a worldwide, royalty-free, non-exclusive[emphasis mine] basis." (We'll call this the "Stack Overflow License.)

As the author of a contribution, content creators can continue to use their works however they wish. Content Creators can grant additional licenses in conjunction with the Creative Commons license and Stack Overflow license, though those additional licenses must also respect the other licenses--ex, the CC BY-SA is well known to have a "no additional restrictions" clause that prevents an entity from claiming "CC BY-SA" but then further limiting usage. Essentially--additional licenses can be more permissive but not more restrictive.

The Question

If a content creator wishes to apply a more permissive license to their content, what is the appropriate way for the content creator to inform this community as well as anyone viewing that data?

For example

Looking at the extreme case, perhaps a user wishes to apply the CC0 license, which places the content wholly in the public domain--effectively waiving any attribution or other restrictions. Would the content creator label their content with a "disclaimer" in their answer that denotes that the content is freely usable?

A little more specific

Many content creators here are also bloggers, writers, or contributors to other websites & open-source projects. Oftentimes, an answer here may be a derivative work of something that creator has previously published elsewhere (like their blog). Depending on that original location, the license is likely not the same CC BY-SA--So that particular answer would carry the CC BY-SA license, but may also carry an additional license from the original work, which may also be attached to derivative works, such as the Stack Overflow answer.

What is the best way for a creator to ensure that the additional licenses are conveyed to visitors of the Stack Overflow version?

12
  • 3
    I don't think the "two licenses" thing is all that certain. I've always read that as restating certain CC license grants that the Stack Exchange network relies on, just because it's good practice. (The "Stack Overflow license" is surrounded by two paragraphs that talk about the Creative Commons license, and the "[…] and […]" structure is later paralleled by another "[…] and […]" where both clauses definitely refer to CC BY-SA licenses.)
    – wizzwizz4
    Commented Jul 30 at 14:57
  • 3
    "though those additional licenses must also respect the other licenses" ... I don't think this is the case either. I can make something I create available under, say, both CC-BY-SA and CC-BY-SA-NC. Maybe one party will prefer licensing it from me under the former as they want their derivative to be less restricted, maybe another party will prefer licensing it under the latter because they don't want their derivative to be used for commercial purposes. AFAIK nothing prevents this.
    – muru
    Commented Jul 30 at 15:00
  • 3
    That's to say, I can make something I create available under additional licenses, each independent of the others
    – muru
    Commented Jul 30 at 15:02
  • @AMtwo The company's current beliefs do not retroactively alter the license: they're not Darth Vader. Additionally, the company hasn't ever claimed there are two licenses: while they've proposed actions that blatantly violate the CC BY-SA licenses, they usually describe them as "as allowed by CC BY-SA" or "in accordance with CC BY-SA" (both paraphrased).
    – wizzwizz4
    Commented Jul 30 at 15:04
  • @wizzwizz4 Regardless, I don't think sorting that out is necessary to answering this question.
    – AMtwo
    Commented Jul 30 at 15:06
  • 1
    I would recommend removing the part about the "Stack Overflow license" from the question. It adds noise. It doesn't matter if there are one license or two, the fundamental question here is if it is possible to or how to go about adding a more permissive license to content posted on the network when it is accessed through the network itself. So, how can I apply something like CC BY or CC0 to my SE posts without cross-posting them to a different location? Commented Jul 30 at 15:15
  • @ThomasOwens It is relevant insomuch as any official answer from the Company may make reference to it, and in particular the fact that licenses are granted on a non-exclusive basis. Dual licensing, and the terms of that dual licensing is secondary to, and must be compliant with the CC license.
    – AMtwo
    Commented Jul 30 at 15:42
  • 1
  • 1
    A tangent maybe, but I also agree with muru above; by design, the CC license explicitly does not restrict the Licensor from additionally licensing your content under any other terms you want. The "problem" with licensing under more restrictive terms is that the CC-BY-SA terms are still around and available, but as muru described, there may be reasons to prefer a more "restrictive" license in certain situations.
    – zcoop98
    Commented Jul 31 at 18:06
  • 1
    Regarding "non-exclusive", the only thing that means in the context of the supposed "Stack Exchange license" is that Stack isn't the only one with license rights over the content. If it is indeed a dual-license, then nothing about CC licenses requires that it be compatible with their terms, which is exactly why people get so uppity about the debate between whether it exists or not. If a true dual-license exists, it exists separate from the CC terms, and need not comply with them.
    – zcoop98
    Commented Jul 31 at 18:25
  • 1
    The only middle ground that could exist, as I understand it, is if the "Stack Exchange license" that Stack grants itself had, in itself, a condition to also apply CC-BY-SA 4.0 requirements. But, that's something it would have to require in its own text, and not something that dual-licensing with CC-BY-SA 4 requires.
    – zcoop98
    Commented Jul 31 at 18:28
  • 1
    Non-exclusivity is an apparent necessary reminder TO THE COMPANY that they aren't the only ones who can use the data.
    – AMtwo
    Commented Jul 31 at 21:24

2 Answers 2

7

From a moderator's point of view - an additional license added somewhere to a post wouldn't be too dissimilar to a signature. It is noise that detracts from the content of a post. There's nothing stopping you from relicensing your code as you feel fit - there's just no good way to do it, inline with a post or the other way. Even in place - it assumes whomever is using the code actually notices, is looking for it, and benefits. On SE we know everything is on roughly the same license - some variant of CC-by-SA. Usefully finding a poly-licensed post or code with a suitable license is tricky.

There's no real mechanism cause when the core of the site was designed, it wasn't needed and practically, it would add a huge amount of complexity to the platform if formalized somehow.

Codidact deals with this with a drop down menu with license options but honestly that wouldn't scale too well IMO at the volumes SE has. It would certainly make something like the datadump more 'fun' and potentially unusable in certain contexts. Want to stand up a commercial, CC-by-SA compliant alternative if SE falls? You need to filter out all the non commercial licenses. It lets you pick one alternative - which isn't what we want here. We want "CC-by-SA" and "additional freedoms". Assuming SE wanted to do this, this is probably the sanest approach, especially tied into search. (Which leads me to the question of what license a code snippet from a question falls under, should it be used in an answer under a differently permissive license).

I don't think anyone even worried about the license versions being changed when Jeff did it (which resulted, eventually, in a kerfuffle as people complained about retroactive relicensing, during one of many low points of trust). Practically, outside keeping the company from paywalling posts - I suspect the initial choice of license was basically a poison pill to prevent that, rather than anything else.

To borrow a phrase from Jeff:

It's not a license mess, some people are a mess.

And in a sense, the situation is a mess.

I mean, the most practical/safe way is to republish your content, with whatever license you want, and link it back to SE for context. This is an unsatisfying answer. And really this does feel like its about more solidly asserting individual content ownership in a period where trust in the company maintaining access for us to things like the data dump is weak. Where there's a lack of trust in the integrity of the company, it might be worth considering backing up one's own content.

(I believe you can request a data dump of your own posts via the GDPR form but the current design is painfully unintuitive.)

11
  • One issue with republishing your content is that it often doesn't make sense out of context. So you end up with some messy ways of linking to content without including the content or messy ways of denoting mixed license content on your own platform. This is something that I've looked at, but it's not an easy thing to make your own SE content available under a permissive license in a way that makes sense. Commented Jul 30 at 15:50
  • part of it is, its going to get messy no matter what you do Commented Jul 30 at 15:53
  • 1
    A license chooser on the platform wouldn't be messy. At least, in the sense that it would be painfully obvious exactly what license each individual work is under. It would be messy in the sense that a given page will have mixed content, but you would also be assured as to what the most restrictive license is. It would be trivial to make that determination in the data dump, API, or SEDE. Commented Jul 30 at 15:57
  • A bunch of my answers were actually from blog posts I had already written, and because link-only answers are not acceptable, my answers on DBA.SE generally link back to my answer, but don't necessarily include the licensing details. Now that Stack is going down the road of added restrictions, I am trying to map out clarity-of-use for my own creations. Since "no LLMs" seems to be an additional retroactive license change the Company is trying to make,
    – AMtwo
    Commented Jul 30 at 15:59
  • Ah, but see "Which leads me to the question of what licence a code snippet from a question falls under, should it be used in an answer under a differently permissive licence)" . I mean practically it shouldn't matter, but if we're splitting hairs... Commented Jul 30 at 15:59
  • 1
    @JourneymanGeek I think we're probably well past that hair even with the current situation - there are probably thousands or hundreds or thousands of images, quoted posts, etc. whose sources didn't allow CC-BY-SA, but the people who pasted them on the network didn't care.
    – muru
    Commented Jul 30 at 16:04
  • 1
    @AMtwo Practically - the 'old' problem of different licences was solved cause there was the will to do this, and the solution was inspired. Most of the potential solutions involve an awareness by decision makers of the core problems and situation and sentiment on the ground. I mean, the solutions here are "the company acting in an enlightened manner" or "just ignore the noise and chaos". We're either in situations where there's no need (cause the company is on the same page as us, or no good way to solve the issue) Commented Jul 30 at 16:05
  • 1
    @muru the post I quoted from Jeff has a very practical solution for that, that its fair use. Sadly that concept is eroded by abuse Commented Jul 30 at 16:06
  • @AMtwo I'm trying to go in reverse. I actually have a big backlog of answers that I want to convert into something useful/usable and post on my own platform under CC BY. Unfortunately, there's no good way for me to link to that more permissive source without editing each post on this platform, and even that could be seen as spammy. At least you have posts here that point to previously written content so the content is already in a useful/usable state and referenced so people can get it under the permissive license. Commented Jul 30 at 16:09
  • Regarding a license picker, I think continuing to require posted content be licensed as at least CC-BY-SA would alleviate some of the mess, since it would give a consistent baseline license that broadly applied. Allowing users to pick other, additional licenses to put their content under would just, IMO, make something that already happens more visible and user-friendly (e.g. compared to putting "I declare all my content is usable under CC0" in the user profile, which... just isn't that usable in practice, especially since there are no public revisions of profiles).
    – zcoop98
    Commented Jul 31 at 18:42
  • @ThomasOwens Regarding "doesn't make sense out of context": I have a mostly-spec'd system for providing that context, based on schema.org. If there's interest, I could make something for a static site generator. (I'd make a JavaScript button that fetches and inserts the relevant question upon request, but it wouldn't work for questions on SE because of CORS.)
    – wizzwizz4
    Commented Jul 31 at 20:13
3

There's no good way to do this on the SE platform.

The CC BY-SA license is extremely visible and obviously connected to each work. It's in the footer of every page. It's in the "share" menu on every question and answer. It's in the hover text on a comment's timestamp. It's in SE API responses. It's in the data dump. In just about every way a person can obtain a copy of a post, it's clearly linked to the appropriate CC BY-SA license based on when it was created or modified.

I have seen people put a disclaimer in their profile. However, this has a lot of problems, ranging from the fact that there aren't good timestamps for when such a disclaimer is added or edited and therefore what posts it may apply to. There's also a lack of connection between the profile and to the post itself without additional work.

The strongest available solution would be to put text in the body of each post. This is problematic for a number of reasons, as well. Once a post is edited, it becomes more difficult to see when this was added and if it was added by the original creator to ensure it's a valid claim. There's also a strong culture of removing extraneous information and this would have legal implications. If given the choice between what the body of the post says and the clearly marked terms, I would follow the clearly marked terms, even if the contributor is trying to use a more permissive license.

The best solution would be for SE to implement a license chooser. Just to stick with the CC licenses, CC 0, CC BY, CC BY-SA would be valid, assuming that SE believes that CC BY-SA is the most restrictive license possible. There may be other open-source licenses that would be appropriate as well. Unless and until something like this is implemented, anyone using SE data should treat it all as the marked CC BY-SA version.

Also, for the record, if I had the option, I'd mark all of my SE contributions as CC BY.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .