Commons:Flickypedia/About

From Wikimedia Commons, the free media repository
Jump to navigation Jump to search
Translate this page; This page contains changes which are not marked for translation.

About Flickypedia

This is project by the Flickr Foundation in partnership with the Wikimedia Foundation, inspired by the Flickr2Commons software tool.

Flickypedia will streamline the transmission of appropriately licensed Flickr images into the Wikimedia Commons. Think of it as an “endorsed pipeline” from one service to the other.

The idea is to build upon the good work already done on Flickr2Commons (which was developed by Magnus Manske) extending it to add features like structured data import and better duplicate detection.

Files uploaded with Flickypedia use {{Uploaded with Flickypedia}} by default and are categorised in Category:Uploads using Flickypedia.

Our Licensing Framework

[edit]

What’s this?

[edit]

The Flickr Foundation is developing a new version of Flickr2Commons to move images from Flickr.com into Wikimedia Commons. It’s called Flickypedia.

This redevelopment of Flickr2Commons is an opportunity to emphasise responsible licensing and work on sealing gaps that expose Flickr members, Wikimedia volunteers, and each of the host platforms to unnecessary risk. The Foundation worked with The GLAM-E Lab to develop this framework to more carefully steward the movement of Flickr images to Wikimedia Commons.

Current challenges to responsible licensing

[edit]
  • “License Washing” is when an image is duplicated and the newer copy is licensed differently from the original. For example, a Flickr member may publish her photograph using an All Rights Reserved license, which could then be washed away into a CC-BY license on a copy. The copy would then be allowed for upload into Wikimedia Commons (and the photographer may not be aware this duplication has even happened.)
  • Sharing personal or private information could be an issue given that Flickr photos often show identifiable people and metadata which can reveal personally identifiable information like names or locations. Care must be taken when sharing photographs like this, particularly in countries where frameworks like the EU’s GDPR exist, but also in general, as the subjects of photographs may not want their image used in unintended circumstances.
  • Different licensing information for the same thing across platforms can happen wherever license information is editable, as it is on both Flickr and Wikimedia Commons. If an image’s license information is changed after copies are propagated outside Flickr, these “off-Flickr” copies may not get updated. If license information is edited off-Flickr, that change doesn’t find its way back to Flickr either. Theoretically, you can have the same image with a CC BY 2.0 license on Flickr, and a CC BY 4.0 license on Wikimedia Commons.
  • “No known copyright restrictions” (NKCR) is not a license. It’s a special assertion only used within the Flickr Commons program on Flickr. Images with NKCR attached do not sit easily in the Wikimedia Commons since Wikimedia Commons requires images uploaded to be licensed as CC0, Public Domain Mark, CC BY, or CC BY-SA. While it’s not documented as a formal decision, many Wikimedia volunteers feel that NKCR images do not meet precautionary principle in the project scope of Wikimedia Commons. (Flickr2Commons stopped making NKCR images available for transfer in 2020.)
  • Copyright infringement can happen with descriptive text. Directly copying something that someone else has written to describe their pictures can be seen as copyright infringement when it goes beyond the more ‘objective’ information about pictures like the date taken or its camera settings.

What we’ll do in Flickypedia to encourage responsible licensing

[edit]
  • Create a public list of all the uploads that come through Flickypedia. That way, people can look at the list to see what’s happened and if their Flickr pictures have been moved across.
  • Ensure we check for duplicate Flickr pictures that already exist on Wikimedia Commons (so they’re not uploaded again), extending what Flickr2Commons already does. This should help with license washing.
  • For every upload, methodically show Flickr-based photo metadata like creator, date taken, date uploaded, and Flickr URL in a structured format that can be read by people, and queried by machines.
  • Backfill the 10 million-plus already-uploaded Flickr images on Wikimedia Commons with new structured data. This will help with license washing and the duplication problem.
  • Help people using Flickypedia understand Wikimedia licensing requirements, like the CC license types and CC0 for “short caption”. We will not import the Flickr Description automatically either, to remove the risk of copyright infringement.
  • Flickypedia will not offer NKCR images as options for transfer to Wikimedia Commons, since NKCR is not a license, and doesn’t meet the Wikimedia Commons’ precautionary principle.
  • Notify Flickr members via comments on flickr.com when photographs are used on Wikimedia Commons, providing a link to the photo’s new page there to see what’s happened.
  • Design a Flickypedia template on Wikimedia Commons to explain all this, and provide links to the various references so people can check and explore things more easily themselves.

We expect these improvements will help Wikimedians by reducing duplication of Flickr uploads, focusing on images that are clearly available for reuse, and making Flickr’s great photography more discoverable through structured data on Wikimedia Commons.

At the Flickr Foundation, Flickypedia lives in our Content Mobility program, which is about facilitating sharing of Flickr images, and knowing where they go across the web. A delightful unintended consequence of the Flickypedia work is weaving several million Flickr images more tightly into the web through our linked open data updates to their records in the Wikimedia universe. We also hold ourselves to a high standard regarding open licensing and reuse, which we’re pleased to demonstrate in Flickypedia.

We hope these improvements to the original Flickr2Commons will make it simpler for everyone using Flickypedia to respect image creators and the time of Wikimedia volunteers.

Flickypedia UI Outline for Version 1

[edit]
This is a flowchart illustrating the basic workflow of the proposed Flickypedia software application.

Basic flow of Flickypedia

[edit]
  1. Hello and instructions
  2. Log in to Wikimedia Commons (required) and Flickr (optional, recommended)
  3. Find Images (scoped to CC-BY, CC0, CC-BY-SA, PD and NKCR)
    • We have built a toy to demonstrate this, launched 4 Oct 2023, called Flinumeratr
  4. Make Selection from results
  5. Prepare Metadata for your upload
    • Show detailed field-to-field mapping - 2 columns: Flickr → Wikimedia Commons
    • Set language for majority of short caption
    • Use Flickr Title field for filename(s)
    • Connect each image to WM category (either new or existing)
    • Append Flickr image ID to filename(s)
  6. Upload and show progress
  7. Review in Wikimedia Commons
  8. Notify Creators in Flickr
    • If authenticated on Flickr and not the creator: leave a comment telling the creator their photo is also now on WM Commons
    • If not authenticated on Flickr and not the creator: Flickypedia Bot leaves comment on Flickr photos with Wikimedia links (but may not have permission to comment)

Scope for Flickypedia V1.0

[edit]

We are developing a new version of the Flickr2Commons called Flickypedia. It will build on Flickr2Commons’ main features, and we hope to add one or two new features as time allows. We expect to iterate on it into 2024.

Key parts of the V1.0 tool will be:

  1. Authentication with Wikimedia Commons - Done! alpha
  2. Retrieve appropriate Flickr images with acceptable CC licenses - Done! alpha
  3. Help the uploader to check metadata / licensing information - Done! alpha
  4. Reconcile structured data (maybe) - Done! alpha, we think!
    • Prevent “license-washing” (maybe)

Note: The “maybe” things may prove hard to launch by the end of this year. We’re still doing our scoping work at time of writing, June/July 2023. But we’d love to include these features at some point as we have already learned they’ll address some gaps.

Features we've added that weren't anticipated:

  1. Backfillr Bot - to add SDC to more than ~10 million Flickr uploads on Wikimedia Commons
  2. Say thanks! - a way for Flickypedia users to notify and thank Flickr members for their photos
  3. Flickr Photo ID has become a formal Wikidata property (which helps with our robust duplicate detection)