Commons:Requests for comment/Licensing of metadata

The following discussion is archived. Please do not modify it. Subsequent comments should be made in a new section. A summary of the conclusions reached follows.

No consensus to support different licenses in namespace 6. –Be..anyone (talk) 11:48, 2 January 2015 (UTC)

---

This RfC is only about the license of the file metadata/description, not the license of the files themselves.

(As recently mentioned on the Village pump and in a parallel discussion on the Cultural-partners e-mail list):
The default licensing of the text on the file description pages and thus (much of) the metadata about the file is automatically licensed under CC BY-SA 3.0 and GFDL (WMF's default) as are any subsequent changes to these pages. We presently apply this license independently of the source of the metadata; be it a GLAM content donation, an image from Flickr or extracted from the Exif data; and independently of how this metadata was previously licensed.

This causes problems for some of our GLAM partners which would require the metadata to be published under a {{CC0}} license. This is for example the case if they want to import updated metadata back into their collections (for a file they have released together with the original metadata) or if they want to include Commons material, such as WLM images, in their database.

During previous discussions the issue of database rights and how these might also be a hindrance for reuse (independently of the license) was also raised, so I have included this aspect below as well.

Questions:

1. Should the WMF explicitly release any database rights that they might be eligible to claim on the Commons metadata in the EU (and other jurisdictions)?
2. Should metadata that has been released under a less restrictive license than CC BY-SA + GFDL or that is deemed non-copyrighteable in a given jurisdiction be tagged accordingly on Commons?
3. Should the upcoming Wikibase integration on Commons treat licensing (of metadata) differently compared to the present situation? E.g. having only non-copyrightable/CC0 information in the wikibase and leave other information on the file description page?
4. And partially related to the previous question: Can the community (and possibly the WMF) agree on any parts of the file metadata which could be deemed non-copyrightable and then pass a policy/resolution on this (similar to the PD-Art decisions/discussion)? E.g. can the choice of license of the file always be deemed to be non-copyrightable?

Discussion

In my mind great majority of file metadata falls under the {{PD-text}} license and is ineligible for copyright. One can not claim copyright on facts, like year of creation, author or size of a painting, and file metadata is suppose to be mostly facts. There are few exceptions where file descriptions might cross that threshold:

• People writing articles or essays about the subject of the photograph. In many cases such descriptions fall under Commons:PS#Excluded_educational_content and are included on Commons only because they can not be added to Wikipedia due to en:WP:NPOV or en:Wikipedia:Notability. However there are legitimate cases like artwork history (conservation history or history of ownership) might be in a form of an essay.
• Occasionally metadata transcribe the text in an image to allow better searching. In many cases those should go to wikisource, but it might not be in scope there.

I think that using multiple licenses for Commons metadata could be quite confusing, but I would support re-licensing the file description metadata (and any database rights) as a default to PD or CC0 unless marked otherwise. If we run into some metadata licensed under different free license we should be able to still use it provided it is marked as such. --Jarekt (talk) 12:16, 2 April 2014 (UTC)

While I agree that the majority of the metadata is ineligible for copyright the lack of guidelines means that resuse becomes more problematic and any effort such as relicensing metadata (where appropriate) upon importing it to a Commons wikibase would have to be determined on a file-by-file basis which is not doable considering the vast amount of files we have. /André Costa (WMSE) (talk) 08:21, 3 April 2014 (UTC)

I would like to see us making more metadata available under CC-0, and it would be good to identify content that can be released under that license (at the very least, image metadata such as the camera model should be CC-0-able as per Jarekt's comment). My main worry would be image descriptions, which may have a variety of licenses, but perhaps that issue can be avoided by clearly identifying a license for those descriptions in the database (which could be bot-imported and then tweaked later on). It would definitely be good to have a more nuanced approach than just CC-BY-SA/GFDL for all text here. On database rights: I think this is already waived by the CC license? Thanks. Mike Peel (talk) 18:58, 6 April 2014 (UTC)

Only 4.0 explicitly deal with the sui generis database rights [1]. Assuming we'll shift over to that at some point the database issue should be cleared up. /André Costa (WMSE) (talk) 07:41, 15 April 2014 (UTC)

About assumptions

• (1) You assume that because mediafiles are licensed in a particular way, the associated meta-data is automatically licensed in the same way. Why is that?
• (2) When a GLAM donates stuff to us. What WE get is a license, they retain the copyright and can re-license it as they see fit.
• (3) You state that meta data is automatically licensed. Did you really check this with a specialised lawyer? Remember, for all intends and purposes the WMF is US law.
• (4) In the way you phrase the RFC, you claim that the Commons community is entitled to decide on this ... first there is the law but why would it be Commons ?
• (5) Did you consider the wider ramifications of making all this copyrightable and licensable. What does it imply when your "copyrighted" material is to be mixed with information from Wikidata, does the mix become automatically licensed more than CC-0 ??

Sorry, but this RFC is wrong in so many ways.. GerardM (talk) 12:15, 3 April 2014 (UTC)

Dear Gerard,

• Ad 1: By default, all text on Wikimedia Commons is licensed under the CC-by-sa 3.0 license and the GFDL. For any edit I make on this wiki, I make a corresponding copyright claim. As most of the content on Wikimedia Commons pages is factual metadata that wouldn't be deemed copyrightable in the first place in many jurisdictions, I guess it perfectly makes sense to question this practice and to suggest that a CC-0 license be applied to all the metadata by default.
• Ad 2: I think we should encourage GLAMs to publish their metadata under a CC-0 license. In many cases, claiming copyright over metadata amounts to copyright overreaching anyways (this practice is comparable to applying a CC-by-sa license on a public domain work and should generally be discourgaged by the community). In cases, where it could be argued that parts of the metadata might be copyrightable because of the individual character of some formulations, asking institutions to provide the data under CC-0 terms would add clarity.
• Ad 3: It is licensed automatically, but the license may be considered as void by the users if the material is not copyrightable. The situation is similar to the one where CC-by-nc licenses are applyied to public domain works. Many users would think twice before using these works for commercial purposes, although from a legal point of view, you could just ignore them (at least that's the legal situation in Switzerland, I haven't questioned U.S. lawyers about this). - Don't you think Wikimedia Commons has a strong interest to avoid such grey zones where Wikimedia Commons claims one thing and the effective legal situation is something else? In my opinion, such practices undermine the CC-by-sa licenses. I think the best way is to apply them only where we can claim with good faith that they are enforceable.
• Ad 4: I guess that's a valid point. - Who decided to apply a CC-by-sa license to metadata on Wikimedia Commons by default? What would be the right procedure to change this policy?
• Ad 5: I guess the intention is to avoid exactly this kind of issue. We should avoid a situation where metadata on Commons is licensed differently from Wikidata, because we should be able to use them together without bothering about incompatible licenses. Furthermore, as André pointed out, heritage institutions which make their metadata available on Wikimedia Commons should be able to re-import the modified/enriched metadata into their databases without bothering about any Share-Alike terms.

Cheers, Beat Estermann (talk) 13:13, 3 April 2014 (UTC)

• When the only reasonable claim to copyright is, the right to the database, would anyone believe the WMF to be restrictive in this way?
• When you consider database rights, what has been donated by a GLAM can be seen as a database in its own right. All the little bits of additional data is consequently a derivative. Who in his right mind will deny our GLAM partners least of all the WMF itself ? Who else has standing ?
• Commons is not exactly the darling of many Wikimedians by its over aggressive approach to copyright and license issues. What is needed is a more granular approach, an approach that is not the current all or nothing. To me, this RFC feels like another land-grab. It is rich in assumptions and it does not have its eye on playing the ball but more on the rules of a game where a ball is played (what ball, what game). Thanks, GerardM (talk) 14:42, 3 April 2014 (UTC)

This RfC is meant to be the opposite of a land grab. Today all the land is already grabbed, what I want is a senisble way of giving back the land which should never have been claimed to start with.

As for the database rights I don't believe WMF would intend to be restrictive enough to claim such rights, BUT many potential re-users would hesitate unless such a clarification was explicitly made. /André Costa (WMSE) (talk) 13:19, 5 April 2014 (UTC)

My assumption: The metadata is under the same license as the file/media it belongs too. No matter what WMF does. --Hedwig in Washington ^(mail?) 18:15, 12 April 2014 (UTC)

I can see how that would be a valid assumption for the EXIF-data (and any info derived from it). The rest of the information on the file description page is entered separately from the file though. /André Costa (WMSE) (talk) 11:04, 13 April 2014 (UTC)

Another assumption is that the uploader doesn't want different licenses for text and image. It wouldn't make any sense. Further, if we assume that text and imaga are one entity, altering the text would technically require a DW template and new upload. --Hedwig in Washington ^(mail?) 19:37, 13 April 2014 (UTC)

---

The above discussion is preserved as an archive. Please do not modify it. Subsequent comments should be made in a new section.