Community Wishlist Survey 2019/Multimedia and Commons/Allow non-CC0 licensed data for datasets

From Meta, a Wikimedia project coordination wiki
Jump to navigation Jump to search
Random proposal ►

◄ Back to Multimedia and Commons  The survey has concluded. Here are the results!


  • Problem: Tabular datasets in Commons cannot contain non-CC0 information because the interface doesn't support the correct attribution
  • Who would benefit: Wikipedians could move the data tables from the article text, and generate dynamic tables or graphs from it through the Graph and Maps extensions
  • Proposed solution: See referenced tickets
  • More comments: See also Village pump discussion
  • Phabricator tickets: phab:T154071, phab:T155290
  • Proposer: Sabas88 (talk) 11:25, 4 November 2018 (UTC)

Discussion[edit]

If this data were in a table, would I have to license it under CC0? How could I claim copyright over basic physical properties of the universe? "Quark masses, copyright User:HLHJ" is absurd.
  • Support. If the blocking technical requirements are those listed at phab:T154071#4323571, they don't look they would require too much to fix, and should be well within the capabilities of the Community Tech team. Shapefiles that are licensed ODBL, eg from OSM, could be of huge usefulness in connection with WDQS queries. Jheald (talk) 22:59, 5 November 2018 (UTC)
  • T200968 documents the current state of play. This will supposedly be worked on soon, though I'll believe it when I see it. Note that (re)fixing this bug was included in my Provide an easier way to create a wikitable from a Commons dataset proposal. Gareth (talk) 05:27, 7 November 2018 (UTC)
Yeah, perhaps if needed, we could consolidate it --Sabas88 (talk) 11:35, 7 November 2018 (UTC)
  • I've already taken data published under another CC license in a scientific paper and made it into a datagraphic which I uploaded to Commons. It's vector-format, so re-extracting the original dataset would be fairly simple. If I've understood, the status quo means that that was OK, but if I put the data in a table and then made a graph I'd be unable to do so legally due to interface limitations. One might argue that scientific data is not copyrightable (I sort of assumed that, and I was following longstanding academic convention), but that's not the same as a CC-0 license. I'd feel stupid placing an unoriginal table of properties of the universe under CC0. And my original work in making a graphic, is that my copyright that I must license? Can I license it differently from the data? Are recently-invented standard statistical visualization techniques copyrighted? I'm confused. I have no idea how to license data correctly, and excluding scientific data from Wikipedia is clearly not the answer.
RStallman (WMF), could you please give some informal guidance as to what license options we might need in the interface? HLHJ (talk) 05:58, 18 November 2018 (UTC)

Voting[edit]