Community Wishlist Survey 2017/Archive/Take the subject of licenses seriously

Random proposal ►◄ Archive

Take the subject of licenses seriously

N Primarily a policy/licensing change rather than a technical feature

• Problem: Wikidata withdraw licenses while performing massive imports from miscellaneous sources, including Wikipedia. This put the project itself, and every project using it at large scale, at a heavy legal risk.

• Who would benefit: Anyone using wikidata at some expand.

• Proposed solution:
  • Integrate a field "licenses", enabling to document this at the same extent as references
    • Update this field for every Wikidata entry which obviously came from a massive import (be it through bots or crowd sourcing)
    • If any, remove data which came from such a massive import backed on sources without at least one explicit free license
    • Allow every user to select a set of free licenses under which they agree to publish their own original contributions, so that the "licenses" field is set automatically for this user contributions. As on Commons, user should select at least one free license, but neither CC0 nor any other license should be an mandatory choice.
      • Of course if the data is already licensed under a license not compatible with some of the user set of license, any attempt to contribute should lead to a warning and, after user approval, a publication only under the modified data license set and its intersection of compatible licenses within the user license set.
    • For the sake of equity that our movement is highlighting as one of its core value, the default license set should only include copyleft licenses.
      • For the sake of compatibility with other Wikimidia projects that Wikidata is suppose to back, this default set of licenses should include CC-BY-SA-3.0-unported and possibly GFDL 1.3.
    • Whatever the default set of licenses will be selected, the existing data should move to it, so it will cover further updates of this entries. But CC0 might stay relevant for data imported so far when they didn't came from sources with incompatible licenses – like Wikipedia. Note that this move is completely allowed by CC0 and was even advanced as part of the plan beyond the test phase.
  • Conduct an open legal inquiry with experts of the field to create a clear up-to-date report of what is – under current laws – definitely legal, definitely illegal, and what is in the realm of legal blur.
    • Following this report, adopt clear policies for Wikidata to only accept contribution that stands in the field of definitely legal practices
      • If any, clean Wikidata entries which are already recorded and don't conform to this policy
      • Possibly, develop some tools which help making this policy respected

• More comments:
  • This solution does not prevent any reuse at all. It only makes more clear what are the real potential legal issues, were the current practise is to pretend there is no issue without presenting any serious report on the subject. Giving a fine grained license information allow reusers to remix accordingly with full background knowledge. Querying features make it easy to filter data which are covered under compatible licenses.
  • On a side note, a policy regarding import of Wikipedia content should be created in order to avoid circular references which make a large portion of Wikidata references useless for Wikipedia and sister projects, all the more when the reference doesn't provide a precise passage of an article in a given version.
  • Enlarged community consultation on some or all of this point would be appreciable.
  • Most points are independents, or the list depth reflect this dependencies, so it's not a take or leave it all proposal.

So do I understand it correctly that during development and testing, we can can go with CC-0, and later relicense to whatever seems suitable, which is possible with CC-0 [Wikidata-l] Is CC the right license for data? Denny Vrandečić Tue Apr 3 09:42:57 UTC 2012

• Phabricator tickets:

• Proposer: Psychoslave (talk) 10:53, 8 November 2017 (UTC)

Community discussion

IANAL but it seems like a good idea to have Wikidata content doubly-licensed under CC-BY-SA and ODbL, to ensure that no compatibility issues arise from importing stuff into Wikidata. It's a win-win: if you take the stance we should ignore the fact that certain jurisdictions uphold database rights, no ill will come from using our data. If you accept the (regrettable) existence of database rights, this will have the added benefit of multiplying free cultural goods. NMaia (talk) 00:42, 9 November 2017 (UTC)

IANAL but it seems like a good idea to have Wikidata content doubly-licensed under CC-BY-SA and ODbL

You can not legally relicense CC-BY-SA under ODBL without explicit agreement of each author. To be clear, I'm not hostile to ODBL nor any free license, although I would recommand non-copyleft licenses only in very specific cases, in which I would not include Wikidata as a whole. But I would argue that you just can't take a whole data set such as Wikipedia, parse it into every predicates that automation state of the art enable you to extract, and put the result under whichever license fit your own personal preferences. All the more when after that you propose to generate stubs for smallest Wikipedia, stubs which could be then be licensed under CC0 as they come from synthesis of CC0 data, isn't it? And in the same time, you can't directly translate an Wikipedia article, even a stub, and say the translation is CC0. But I'm also not a lawyer, hence the need of creating a report by experts of the field as a prior requirement for some points of the proposal.

Also note that with this proposal, users which would like to do so, might indeed add ODBL as one of the license set under which they agree to release their works, and that massive import from ODBL sources could lie within Wikidata, but they would be flagged as such (as well as under other licenses if it does apply). Responsibility for mixing in synthetic ways a large set of data which are coming from misc. sources would be buried by the mixer.

It's a win-win

if you take the stance we should ignore the fact that certain jurisdictions uphold database rights, no ill will come from using our data.

No it's not a win-win. It's a "I withdraw your license clauses which doesn't accommodate me by withdrawing the whole license" case. Those who licensed their work under CC-BY-SA virtually lose all their rights in the process. I said virtually, because until they are serious investigation done which prove it otherwise, this is just a plain illegal practice, and everybody who use this data are just building their works on highly doubtful legal ground.

Please stop propagating the obviously fallacious "certain/some jurisdictions" offer monopoly on database. So far, the most official peace of claims made by the foundation regarding database rights states that their are indeed monopoly accorded on databases, based on misc. legal grounds, both in USA and Europe. Ok, that's not the whole world, but it's not just a few marginal cases. All the more the previous link doesn't say there is no monopoly on data in other countries, so until someone come a complete report on the state of the law on this topic, please assume that it is not legally safe to make as if there was issue. Plus the motto for our movement is "a world where every human have access to the sum of all knowledge, not those who have the luck to live in a favourable jurisdiction.

• This seems more a policy issue, which probably doesn't really match the goals of community tech tasks. Shouldn't this be discussed within the Wikidata community first ? —TheDJ (talk • contribs) 07:54, 9 November 2017 (UTC)

This seems more a policy issue, which probably doesn't really match the goals of community tech tasks.

This is a policy issue, which do lead to technical needs to help its tracking and compliance. Of course the root requirement is not technical, I would not expect a single requirement to start from purely technical needs. Technical solutions are there to solve problems of our community, aren't they?

The proposal make it clear that there are points which do have extra-technical requirements, but the dependency tree always finish with a leaf that is technical or have at least one technical dependency. Plus, once more, this is not a take or leave it all proposal.

Shouldn't this be discussed within the Wikidata community first ?

Wikidata doesn't impact only Wikidata, but all Wikimedia projects and even beyond as the Wikidata team is actively pushing exo-wikimedian reuse. So, of course Wikidata community is warmly welcome to discuss the topic, just as the rest of our community. Sure the discussion might have been conducted completely out of this technical wishlist, but since some possible technical requirements have already been identified, this just make plain sense to report it there, doesn't it?

--Psychoslave (talk) 09:11, 9 November 2017 (UTC)

Having the discussion at this place is a good way to have a discussion that won't affect policy as the technical wishlist as it's not within the powers of the Community Tech to affect policy.

This is essentially about removing the feature that all of Wikidata is CC0 and the wishlist documentation says "One more note: Proposals that call for removing or disabling a feature that a WMF product team has worked on are outside of Community Tech's possible scope. They won't be in the voting phase."

Saving that it's a win-win also ignores the fact that not having data in CC0 creates problems for Wikidata. To take one example, Quora couldn't easily exchange data with us if we would have a copyleft license. The same goes for other projects. ChristianKl (talk) 10:43, 9 November 2017 (UTC)

the technical wishlist as it's not within the powers of the Community Tech to affect policy

I agree, the technical community don't needs to establish policies (beyond its own right to discuss it as part of the community), if that is what you mean. Nevertheless it can provide tools which help to track respect of policy, whatever it happens to be.

This is essentially about removing the feature that all of Wikidata is CC0 and the wishlist documentation says "One more note: Proposals that call for removing or disabling a feature that a WMF product team has worked on are outside of Community Tech's possible scope. They won't be in the voting phase."

That's not a bug, it's a feature. Well, no, this is a serious issue that have to be resolved yet with serious investigations on the one hand, and it happens that on the other hand a technical feature might lower the heavily doubtful legal state of Wikidata items. All the more the feature requested might also serve to explicitly state that an item is under CC0, if that is not legally doubtful in view of the data source and the amount of data extracted from it that was imported within Wikidata. I'm not aware of any feature on this topic for which the WMF product team that would be removed by such an addition of feature. At worst it might require deletion of items whom provenance raises licensing doubts, but no such thing can be added in the database by the WMF product team as they surely respect that the foundation terms of use that they merely host the content, and that editorial control is in the hands of [contributors] who create and manage the content.

Saving that it's a win-win also ignores the fact that not having data in CC0 creates problems for Wikidata. To take one example, Quora couldn't easily exchange data with us if we would have a copyleft license.

You are making a wrong assumption if you take this proposal if you take for hypothesis that it tries to prevent any use of CC0 within Wikidata. If Quora did obtained from their users to publish all their works under a CC0, then we don't need any further permission to import that in Wikidata, and specifying a source with timestamps and specify it was published under CC0, then its fine. Or if they have or their user granted them a non-exclusive, transferable, sub-licensable, royalty-free, worldwide license to do whatever they want with their works, surely there is no problem with Quora including data under CC0 within Wikidata. But if the problem you mention is that with copyleft license Quora can't do whatever they want with the work of the Wikimedia community, then yeah, that's the intended goal of the copyleft license. Just like people are not legally allowed to upload random files from Commons on Facebook. What you are talking about is all but win-win. So far, it looks more like Wikidata was transformed into a license laundering machine, gathering tremendous amount of data which is pretended usable under CC0 without providing any serious clue to ground that claim. The fact that Google founded half the sum of the initial project and that Denny Vrandečić which have been leader of the project and a fervent CC0 promoter now works on the Google Knowledge Graph which – with other sources – relies on Wikidata, is worrisome. Not because they use Wikidata community works, which is fine, and copyleft license allow that too. But because tomorrow they might just improve their own internal database, curated with custom non-free curation interface or simply through their sprawling worldwide tracking system of users who contribute in an unaware passive manner. And as they will have extracted all the value of Wikimedian CC-BY-SA projects through the Wikidata license laundering machine, our projects will be left as a hopeless hijacked victim, and our then invisible community will be dispelled as slaves in the realm of the all mighty for-profit tech giants. Well, sure, that's a grandiloquent way to paint a really dark dystopia picture. But, sadly, it's not a unrealistic scenario.

So, really we shouldn't care much if Quora or anyone else out there might have difficulty to use Wikimedia works as a result of our movement knowledge equity commitments. If the price asked for some agreements is to approve inequity, then our movement must reject it.

But for now, the main technical point of this proposal is not asking anything but to provide better license curation tools to our community. --Psychoslave (talk) 13:28, 9 November 2017 (UTC)

Serving data under CC0 is allowing data users a maximum of freedom. This license choice allows Quora to participate by adding free data to our project. The statement says "We will break down the social, political, and technical barriers preventing people from accessing and contributing to free knowledge". Having Wikidata under a more restrictive license would be a barrier for some stakeholders to participate.

Disagreeing with policy decisions is your right, but this isn't the forum where policy is supposed to be decided and pretending that this doesn't go against the decisions for the Wikidata project is ignoring the realities. ChristianKl (talk) 13:53, 9 November 2017 (UTC)

Serving data under CC0 is allowing data users a maximum of freedom.

But that's not the goal of Wikimedia movement. The goal is knowledge equity. We can promote freedom wherever it begins to confirm freedom of everybody, but beside this threshold freedom, that knowledge ought to canalise, is blind pulse with no safeguard against the worst behaviours that humans are apt to.

This license choice allows Quora to participate by adding free data to our project.

That wouldn't change with this proposal. If Quora do have the right and the will to add data under CC0 to Wikidata, this proposal will let this possibility intact.

Having Wikidata under a more restrictive license would be a barrier for some stakeholders to participate.

The proposal wouldn't prevent what you described. Plus the converse of your statement is also true: having a less restrictive license can also be a repulsive barrier for those willing to contribute, but only within a fair mutual equal right beside works that will be derivative from their contributions.

Admittedly, I fail to read "We will break down the social, political, and technical barriers preventing people from accessing and contributing to free knowledge" as "we will encourage individual to withdraw all rights that civil societies ever granted them so that for-profit corporation can exploit freely their works, being assure that they would have no legal duty if they ever arrived at a winner-take-it-all economical monopoly position". If you want to insist more on the case of Quora, I would ask you to provide more information about what do they give, what do they take, who created what they give (benevolent or paid editors?), under which condition (license, EULA) this work have been created, and how Wikidata is used by Quora if it is in any way (do they credit their sources?).

All the more, if they can release under CC0, then they can release it under a copyleft license, so there is no legal mandatory reasons to refuse to do so if where interested in an equitable cooperation, istn't it? --Psychoslave (talk) 14:37, 9 November 2017 (UTC)

Example needed

 Oppose adding a license field. Wikidata is CC-0 licensed and it should remain that way. Anybody addding data should be aware of that. If there has been a copyright violation, the copyright owner can request the associated data to be removed; of course we can also be proactive in such cases if we become aware of them. Have we had any example of such a request? Can you envision a specific case where it would happen? Why wouldn't it be handled the same way copy vio's are handled in wikipedia and Commons? I don't see any need to change the licensing in wikidata itself. ArthurPSmith (talk) 13:26, 9 November 2017 (UTC)

Wikidata is CC-0 licensed and it should remain that way.

Wikidata pretends it can re-license anything under CC0 regardless of the legal status of the source and the amount of extracted data from it. That's really different. It's more like if I pretended that since I only send you a sequence of bytes which, taken alone, represent a at best some random integer, that the fact that the whole file you built with this signal happen to be a perfect copy of the last Hollywood blockbuster blue ray makes you legally free to do whatever you want with this file, like dealing with partners. However database comes with additional special legal information monopoly which vary between countries, that's basically a good analogy for the approach that so far Wikidata took regarding licensing issues.

If there has been a copyright violation, the copyright owner can request the associated data to be removed; of course we can also be proactive in such cases if we become aware of them.

Or we can be proactive in just respecting licenses traceability and let end user take the responsibility to mix them in synthesis works, or to filter data by compatible licenses.

Can you envision a specific case where it would happen?

Well, just as many cases as there was performance of massive data import which didn't compelled with the source term of use. So far, as far as I know, misc. Wikipedia versions are among the main source provider, so it makes a lot of contributor potentially able to sue for license violation. That doesn't mean this would necessarily lead to a condemnation. I just don't know, really. That's why I said it's a doubtful legal status, and not a clearly illegal practise.

Why wouldn't it be handled the same way copy vio's are handled in wikipedia and Commons?

Because Wikidata is designed for scalability through automation, not only legal issues are also scaled, but without a proper tracking system, it will be difficult to analyse what data are impacted by a given legal issue.

I don't see any need to change the licensing in wikidata itself.

I hope my answers might enlighten somewhat. --Psychoslave (talk) 13:56, 9 November 2017 (UTC)

if the issue is copyright on large-scale collections of data and not individual facts, then adding a license field on individual statements cannot help. Perhaps this issue needs to feed into bot approvals and could be associated with some sort of tagging of large-scale collections of bot edits. Putting a license field on statements is definitely the wrong approach here. ArthurPSmith (talk) 15:29, 9 November 2017 (UTC)

How this would not help when this would result in copyleft license compliance? This is indeed a tagging of large-scale collection import, whether through bots or crowdsourcing, with each item keeping a tags for licenses (and matching references), if you want to name it this way. The result is that one might select items which are available under a given license. --Psychoslave (talk) 21:37, 9 November 2017 (UTC)

It's not true that Wikidata says that any data can be imported. On the other hand, facts aren't copyrighted. When the American Chemical Association told Wikipedia a decade ago that they aren't happy with Wikipedia to store their CAS numbers, Wikipedia position was that CAS numbers are facts and can be freely shared even if the ACA didn't like it. WikiCommons made decisions like hosting the monkey-selfie to assert a permissive interpretation of copyright. Wikimedia's position was never to stay aware from the grey-zone of copyright as that removes the abilities of us to do what we do but it's rather to fight the lawsuit for the monkey-selfie.

Having the view that facts are copyrighted wouldn't only make it more complex to import from Wikipedia but also to import from elsewhere. ChristianKl (talk) 15:52, 9 November 2017 (UTC)

Recently, a person on our project chat asked whether he can contribute data about when software was released. He used our data on software release data as input for a machine learning product. Given that our data about software release dates isn't complete, he pays people to complete the data and is willing to give that data back. This is the kind of commercial project that benefits from Wikidata. If we had a license that would force him to relicense everything his machine learning algorithm produces as CC-BY-SA, he probably wouldn't build on our dataset.

Just like corporations rather use Apache licensed libraries and contribute back to them then doing the same thing with GPL, it's easier to work with more freely licensed data. ChristianKl (talk) 15:53, 9 November 2017 (UTC)

Given the rules of share-alike combining multiple share-alike licenses that aren't compatible in one project of linked data as you propose might also create it's own legal risks. ChristianKl (talk) 16:06, 9 November 2017 (UTC)

It's not true that Wikidata says that any data can be imported.

I don't think I said otherwise, and of course I hope you can't import the indefinite stream of bytes that generate your local /dev/urandom stating that your personal computer generated this peace of bytes at some timestamp, which would be factual but completely useless as far as I can imagine.

On the other hand, facts aren't copyrighted.

This is just a too fuzzy statement to have any relevancy. At large, this is just plain false, as to retake the Harry Potter example, you can't enumerate the complete list of words and their ordinal occurrence positions that appear in the saga. Yes, this would be a lot of factual data, and as they provide enough data to rebuild to whole work, this would be just plain hold copyright violation. The same is true with Wikipedia articles. Imagine you can automatically generate a whole article on a linguistic Wikipedia which doesn't cover the subject yet thanks to data stored in Wikidata which were extracted from an other Wikipedia. We are not yet there, but there are already stubs which are created that way, according to what I red somewhere. Now, imagine that enough information was extracted from the article and that the generator is so efficient that the result is equivalent with a direct translation. If you believe Wikidata claim about its license laundering ability "because it's only facts", then you could publish this synthetic article under CC0. But a direct translation would require CC-BY-SA. Even better, imagine your generator actually generate the article within the same language as the single original article from which all related data were extracted from, with a 100% match between the original and the synthetic. Ta ta, this generated article is now CC0, so is the original one, according to Wikidata claim that factual data can't have copyright issue at any scale. Clearly there is a problem. Sure with predicates stored in Wikidata, such a 100% matching will be more difficult to obtain than with an enumeration of factual statement about ordinals where a word appears in a text, but conceptually the problem is exactly the same.

Wikimedia's position was never to stay aware from the grey-zone of copyright as that removes the abilities of us to do what we do but it's rather to fight the lawsuit for the monkey-selfie.

Well, I like this attitude. But right now, we are in the purest grey grey-zone. I would be fine with a Wikipedia vs Wikidata case with WMF as both complainant and defendant (if such a thing is possible). That would minimize the risk and make it clear whether license laundering through the factual data argument is legally receivable and at which scale. Until we do have such certainty, then we should just respect the licenses of source works.

Having the view that facts are copyrighted wouldn't only make it more complex to import from Wikipedia but also to import from elsewhere.

Sure, law is always cumbersome when it stands in our way. But making as it doesn't exist where it annoy us don't make it vanish. Also you seems to continuously confuse facts, data carrying statements about facts and large aggregation of that kind of data. That might be OK in casual conversation, but for such a topic, please avoid to mix everything up.

Recently, a person on our project chat asked whether he can contribute data about when software was released. He used our data on software release data as input for a machine learning product. Given that our data about software release dates isn't complete, he pays people to complete the data and is willing to give that data back. This is the kind of commercial project that benefits from Wikidata. If we had a license that would force him to relicense everything his machine learning algorithm produces as CC-BY-SA, he probably wouldn't build on our dataset.

Well, first I don't plaid for forbidding CC0 in Wikidata, although I do plaid for a default set of license which only include copyleft licenses. If there are users who do want to withdraw every single right they have on their work when they have other options, that's fine. But forcing people to contribute under such a license when all other Wikimedia project where Wikimedians have grown their communities use copyleft license or at least let the user choice.

Now, back to your example, I just don't care to release my work under licenses that are not practical for people who are not willing to play equitably and publish back under the same conditions. Either they would have comply with CC-BY-SA, or they would have search an other solution, both being far better than to see them using a free work to create derivatives and publish the result as a non-free work. Actually to me your example means that we missed a possible occasion to foster free works, so it's an example against non-copyleft license.

But once more, it's OK to let people publish under CC0 if they really want to do so. But making it mandatory to contribute to Wikidata is not OK, as it would be unacceptable to have a mandatory CC0 on Commons or all Wikimedia projects. All the more, mass import of non-CC0 material while nothing proved it's legal is clearly an issue which should be solved.

--Psychoslave (talk) 22:54, 9 November 2017 (UTC)

• CC-0 is fine for Wikidata. Facts can't have a copyright, and I am against a copyright for a collection of facts. --Yann (talk) 15:23, 9 November 2017 (UTC)
  • Maybe you mean that you feel fine with CC0 for Wikidata. It doesn't mean it is legally fine to make massive import of non-CC0 sources within a CC0 database. The same apply regarding copyright and factual data collection, a personal mood is not a law. Here are some facts: The first word of the w:Wikimedia article is The, the second is Wikimedia, the third is movement, and I could factually enumerate each word within this article. Enumerating a few words and their respective positions in the text doesn't raise any legal problem. Enumerating a large part of them, or all them, does raise legal concerns. Actually it's plain old copyright violation. Otherwise circumventing any information monopoly would be trivial, and related laws would be pointless. While I do personally think that no monopoly should ever be granted on any kind of information, this is my personal opinion which unfortunately doesn't match legal reality. I don't publish Harry Potter saga in Wikisource, nor a factual position of each word occurrence used in Harry Potter within Wikidata because I think it would be illegal, not because I think that law is fine or that this works and fact are culturally worthless. Therefore, as long as they are law which grant monopoly on information and that it's unclear to which degree they impact a database who aims at gathering facts, it would be more careful to track finely sources of statements and their licenses. Wikimedia is not sci-hub, we follow the path of legality, whatever our personal feelings are about this laws. --Psychoslave (talk) 20:31, 9 November 2017 (UTC)

• This is not a community issue. It is up to the WMF legal team to decide if they are happy with the possible liability of licensing bulk copyrighted data under CC-0. The variation by jurisdiction is immense, and I am taking the view that although in some instances we are probably causing a copyright violation, this is not a major issue, and the WMF and partners are perfectly able to deal with any possible lawsuits that result from it. I assume it will be reasonably possible to remove any datasets that are litigated if and when, given that this is structured data, and the database owner could simply run a search query to ascertain the affected items in any given dispute. In short, if it ain't broke, don't fix it. A Den Jentyl Ettien Avel Dysklyver (talk) 16:13, 9 November 2017 (UTC)

This is not a community issue.

Please provide a method to decide what is a community issue according to you. Here is my own proposal: 1. a person have concerns that according to this same person have an impact on the community. 2. this person report the said concerns 3. the community decide whether a. it doesn't care, b. share this concerns or c. explain why this concerns are mere delusions supported with a large rejection consensus indicating that the community find the problem is irrelevant. I don't feel like we are nowhere near 3.c so far, so I will take you claim that this is not an issue as an indicator for such a decision process, but not a conclusive fact, even in bold character.

It is up to the WMF legal team to decide if they are happy with the possible liability of licensing bulk copyrighted data under CC-0.

My opinion is that the Foundation is is accountable to the community. So they are not free to experiment whatever might be the fancy of the day at the expanse of putting the community sustainability at risk and certainly not without a clear transparent report which evaluate this risks. Assume good faith is fine, and trust balanced with known past behaviour is OK. But blind trust is misguided, don't blind trust the foundation, me or even yourself: keep your mind open to critiques from and toward everybody. But avoid paranoia (even if we indeed have a worldwide secret evil plan against you, mouhahaha).

I assume it will be reasonably possible to remove any datasets that are litigated if and when, given that this is structured data, and the database owner could simply run a search query to ascertain the affected items in any given dispute.

A goal of this is proposal is to ease that kind of solving and even prevent this kind of litigation. However it doesn't go as far as systematic deletion of all problematic material under the current lack of license tracking system.

In short, if it ain't broke, don't fix it.

Yeah, so, my opinion is that the current state of Wikidata is broken. The license problem apart, it didn't hold its promise of avoiding circular dependencies with Wikipedia references. This is an other related problem of source tracking laxity. Denying the broken state won't change the fact that it's here.

--Psychoslave (talk) 21:26, 9 November 2017 (UTC)

Archived

This proposal is primarily about changing the licensing policy and while it includes some development work, we can't start this work before there is agreement that licensing should change. This is a technical survey and thus it's not suitable for policy discussions. I suggest you to start a discussion on Wikidata because until the community is on board , this is not happening. Considering all of the above, I've archived this proposal. Thanks for participating in our survey! Max Semenik (talk) 23:56, 9 November 2017 (UTC)

Well @MaxSem:, I'm a bit disappointed of course, but I prefer to concentrate on constructive actions rather than complaining. So your suggestion to put the community as a whole on board, not on Wikidata alone because this is a problem which impact the whole Wikimedia ecosystem. What are the proper processes to engage the whole community about a topic with such a large impact on its future for a mere contributor like me? --Psychoslave (talk) 09:05, 10 November 2017 (UTC)

Wikimedia has decentralized decision making. Apart from the movement strategy discussion, there's no venue for whole community discussions. ChristianKl (talk) 21:05, 10 November 2017 (UTC)