Commons:Curation

From Wikimedia Commons, the free media repository
Jump to navigation Jump to search

Following a discussion on curating files on the Village Pump, let's try to hash out the idea a bit further here.

Why?[edit]

The problem: Commons has tons and tons of files, and it's not always easy to find the best ones. Curation currently consists of Quality Images (limited to Commons contributors), Valued Images (no copyright checks, requires a voting process, not enough media with VI status to make it useful as a search tool) and Featured Pictures which is a different beast altogether. There are galleries as well, but there are very few of them, and they are typically not maintained.

The question: can we make something that makes it easier to find useful files?

What?[edit]

  • All files can be eligible, curation is not restricted to images.
  • Files may be curated based on various factors, some possibilities:
    • Are among the most useful to illustrate a subject
    • Technical quality (how to rate?)
    • Have a checked copyright status and are likely properly licensed: trusted contributors, Flickr users with history, USGov, social media accounts that have been verified, no issues related to FoP or DW, etc. Not getting deleted under COM:PRP isn't enough, the bar is higher than that.
    • Good categorization
    • Good description
    • Geocoded

Who?[edit]

  • All Commons contributors in a voting process? Pro: democratic. Con: slow. This is an important problem with VI.
  • A restricted set of users, like autopatrollers or patrollers who can curate files on their own? Pro: many times faster. Con: less democratic.
  • Who/how can demote a file? The same set of users, who in case of conflict would be dealt with similar to how enwiki deals with edit wars? License reviewers? Admins?

How?[edit]

  • Separate curation for copyright status, usefulness, technical quality, good categorization, etc?
  • A more simplified system, like "preferred", "discouraged" and "unassessed"? (and possibly "neutral"?)
  • Templates/wikitext or structured data? In case of the latter, we'll need tools. (otherwise we could use VFC) A curator should be able to view 200 images in a slideshow/gallery and assess all of them in one go. Or maybe we could simply categorize files with HotCat/Cat-a-Lot and a bot could convert the categorization into structured data?