Florida Wikidata: Knowledge from different points of view: A response from Wikidata.← Back to ContentsView Latest Issue9 December 2015
Discuss this story
Compatibility of Wikidata's CC0 licence with Wikipedia's CC-BY-SA licence
Lydia, Wikidata imports large amounts of data from Wikipedia infoboxes, templates etc. which it then republishes it under the CC0 licence.
Unlike Wikipedia's CC BY-SA licence, which retains the right to attribution and imposes on re-users the obligation to ShareAlike, assuring that Wikipedia will be named as the source, CC0 waives all authors' rights. This means that content compiled in Wikipedia under the CC BY-SA licence is republished by Wikidata under a licence that does away with the rights contributors here were told they possessed when they contributed to Wikipedia.
Given this background – and the information on database rights provided by the WMF Legal Team in https://meta.wikimedia.org/wiki/Wikilegal/Database_Rights – could you comment on these recent mailing list posts [2][3]? Originally people were told that Wikidata did "not plan to extract content out of Wikipedia at all", but would "provide data that can be reused in the Wikipedias".
Note also this post on de:WP calling for an RfC to determine whether the community would be willing to agree to waive its CC BY-SA rights for Wikidata, and that DBpedia, engaged in a similar endeavour of data extraction from Wikipedia, made a conscious choice to use the same licence as Wikipedia. Thank you. Andreas JN466 21:23, 12 December 2015 (UTC) expanded by --Andreas JN466 16:52, 13 December 2015 (UTC)[reply]
Usefulness of unsourced Wikidata imports to Wikipedia
Lydia, most Wikipedias do not allow users to cite one Wikipedia article as a source in another. It's a basic principle of Verifiability. If Wikidata imports unsourced content from various Wikipedia language versions' infoboxes, it seems to me that these imported Wikidata contents can only be used to populate infoboxes in another Wikipedia if they contain an external source (something that is generally not the case in Wikidata today). Moreover, this source would have to be visible in every Wikipedia that draws on the relevant Wikidata content, as otherwise the content would be unverifiable for the reader. Does that match your vision? Andreas JN466 21:23, 12 December 2015 (UTC)[reply]
I am concerned that some might view the Wikidata material as an acceptable "reliable source" whether directly or indirectly for actual articles. Such a position, I fear, would be a wondrous Pandora's box indeed. Instead, I suggest that a wall be placed here at the outset - let those who are completely outside Wikipedia be able to use Wikidata, but estop those within this walled area from using it. Collect (talk) 13:01, 13 December 2015 (UTC)[reply]
Building a multi-lingual repository of useful data
Speaking as a Wikidatan, I enjoy working on Dutch 17th-century paintings and last year I joined the "d:Wikidata:WikiProject sum of all paintings" group. Wikidata is constantly improving the number of properties that can be used on paintings and I have helped propose and model the usage of these. Today I add reference statements because in the beginning I didn't know how. If all statements are from the same reference (like a museum website) I just used the "Described at url" property to link back to the website entry because I was used to working that way on Commons. We are still searching for ways to describe paintings in terms of movement, style, and period. I have re-used a number of references added by others on Wikipedia projects which were very beneficial to me as a Wikipedian, so now it's my way of giving back, by adding these to Commons images or Wikidata items where appropriate. Each project has its own community of volunteers and there don't seem to be many who venture out into the others like I do. I have been a member of the Wikipedia:Visual arts group on Wikipedia as well, but on Wikidata the interaction with like-minded people happens less on talk pages, because we don't speak a common language. Wikidata enables data-sharing by leveling the playing field to all mono-lingual players. I am used to working on image files of paintings on Wikimedia Commons, where we have similar language issues, but there we also have lots of complicated discussions about copyright problems. On Wikidata it doesn't matter whether you are working on a painting collection of modern art or 17th-century art, because the data model is basically the same. We may not be able to show you an image, but we can tell you where it is and all sorts of other things about copyrighted images. In some cases we can link out to a picture of it somewhere. As far as data quality goes, what I think a lot of people don't understand is that if one Wikipedia is in disagreement with another Wikipedia on any issue (such as a painting attribution to its creator), on Wikidata both statements can reside side by side with the "Normal" rank. Currently Wikipedia only has two ranks for "points of view" in statements; namely published and deleted. As it is now, everything deleted from Wikipedia just disappears, whether it is an alternate point of view or pure vandalism. On the other hand, everything that is published has a ring of "truth" to it, whether or not it's under discussion. The Wikidata item, like any wiki, has these published ("Normal") and deleted state for statements, but it also allows these two extra states ("Preferred" and "Deprecated"). We see the deprecated state used a lot for past designations, so for example in buildings from the project Wiki Loves Monuments where old buildings are re-purposed over the years and so on. The "Preferred" status is used to indicate the value that is considered as coming from the latest, or "most reliable source". This may be seen as a controversial issue, namely choosing which source, and one can argue that this is of course a matter of opinion, but it is much more often a matter of consilience. I have used the "Preferred" value several times for painting attributions, using catalogs by leading art historians as references. When in doubt, I allow multiple statements to reside side by side with the "Normal" rank. We see this occasionally on Wikipedia with a lead statement such as "...is a painting by XXX or associated workshop". Wikidata can add precision to this statement by actually naming the individuals of that workshop to whom the painting has been attributed in the past. Jane (talk) 16:44, 13 December 2015 (UTC)[reply]