Grants:IdeaLab/Make it easy to embed Wikimedia content
What is the problem you're trying to solve?
One of the directions described for the Wikimedia Foundation in our 2016-2017 annual plan is the idea of "Reach".
"Reach; everyone should be able to freely share in the free knowledge on our projects. We are making it a priority to reach new readers and contributors around the world and improve the experience for all, including those on mobile devices. To do this, we will better understand and respond to the needs of our global readers and contributors through research, new technologies, effective programs, strategic partnerships, and more."
From the mission statement:
“The mission of the Wikimedia Foundation is to empower and engage people around the world to collect and develop educational content under a free license or in the public domain, and to disseminate it effectively and globally.”
There is one achievable way to both further this direction and the mission that has not been strongly considered. There needs to be a consistent, easy way to embed and reuse Wikimedia content outside of WMF-supported wikis. The ability to embed content has been incredibly successful for other online resources of media - Flickr, Soundcloud, YouTube, TED, Instagram and more. Especially with the declining importance of Google in mediating the internet, exposing the content within our movement outside of our corner of the world can draw more people back to our projects.
What is your solution?
Imagine you are working on an online project that utilizes Wikimedia contributions. Say a journal, news report on a topic, or even a casual blog post. You want to reference information on Wikipedia, but you want something more than a hyperlink. You have to:
- Copy and paste the content
- Copy and paste the wiki URL
- Write up the proper attribution
- Implement that in whatever technical solution (like a CMS) you or your organization uses to create a stylized version of the content.
All of this, every time.
This is a pain, inconsistent, and you're not 100% sure you'll get the attribution correct.
What if you could instead easily copy a normal wiki URL and something like this would appear:
And what about images from commons?
Embedded in a content management system (CMS) you might see something like this:
What would this work with?
- Articles - Providing a link to xx.wikipedia.org/wiki/Article_Title and the first freely licensed image (with the option to remove the image if it's not relevant)
- Images - Commons images with attribution and a link to the file page.
- Maps - Maps could have a link to the fullscreen map (similar to how maplink works)
- Graphs - Ditto
- Article excerpts - Just a few sentences or a section - think Wiktionary definitions for a word.
Where would this work?
Everywhere that can ingest the meta data schema to present a rich interface to content. Content management systems, social media, even other MediaWiki wikis could be a consumer.
How would this work?
There are many existing ways to implement from a technical standpoint, and nothing preventing us from building a solution on our own (although "not invented here" is a concern). A few possible APIs to investigate:
- schema.org - used by many known sites to makeup their metadata
- oEmbed - a popular format for embedding content
- Some crazy new iframe thing we cook up
- http://embed.ly - not a solution for Wikimedia per se, but does support Wikipedia. :)
Note on Commons
Imagine if we were unable to embed images from Commons on Wikipedia. That our only access to that information was a hyperlink. That would be insanity; but that kind of difficulty is what other CMS users face. Our current strategy for embedding on MediaWiki, ought to be expanded to be included in other platforms. Integration with WordPress alone would invite reusers to enrich nearly 25% of the web with our content.  Engineering for this work, would have to account for structured data on commons, which is currently on the roadmap for the Wikidata Development.
Concerns to address
- The biggest is attribution. Our implementation would need to ensure that reuse would include proper attribution for all media types, after all -- most of our content use free licenses that require attribution.
- Open vs. proprietary - I propose we do consider adopting schemas that are popularized by for-profit organizations - specifically social media services such as Facebook and Twitter. We've seen good success elsewhere in the movement with using these services to further awareness of the work being done. The cost is minimal, the reach is great.
- Create an easy-to-use, consistent way to share Wikimedia content with attribution across the rest of the Internet.
- Initially this could should focus on a single content type. For a initial minimally functional feature I suggest we focus on the common Wikipedia article link
- We should be able to measure these embeds to track their usage (while respecting user privacy)
- The Amelia Earhart article has been embedded N times in the last 30 days
- 1 Project idea
- 2 What would this work with?
- 3 Where would this work?
- 4 How would this work?
- 5 Note on Commons
- 6 Concerns to address
- 7 Project goals
- 8 Get involved
- 9 See also
- 10 References
- 11 Expand your idea
- 12 Endorsements
- 13 Participants
We have yet to identify designers, developers and other technical folk who could implement this idea. If you are interested add yourself to the participants below, and describe how you could help.
- The Wikipedia Library/Embedded discovery
- https://www.mediawiki.org/wiki/Wikimedia_Apps/Share_a_fact - A very similar feature in the Wikipedia native mobile apps
- 50 Shades of Wiki: New Tools to Share, Remix, and Reuse - Session from Wikimania 2015
- RDP Wiki Embed - A WordPress plugin for MediaWiki URLs
- Stockphoto - A Gadget that provides code to share Commons media
Related Phabricator tasks
- https://phabricator.wikimedia.org/T56829 - Implement support for sharing protocols (tracking)
- https://phabricator.wikimedia.org/T27854 - Expose image thumbs, embedded video players via oEmbed (API + discovery <link rel>)
- https://phabricator.wikimedia.org/T142243 - Add open graph meta tags
- https://phabricator.wikimedia.org/T142090 - Add hover-card like summary to open graph meta data
- MediaWiki could be its own generator and consumer of metadata from itself and other systems.
- "Usage Statistics and Market Share of WordPress for Websites, August 2016". w3techs.com. Retrieved 2016-08-11.
Expand your idea
Would a grant from the Wikimedia Foundation help make your idea happen? You can expand this idea into a grant proposal.
- I think this kind of external discovery, is really important for our long-term relationship with not only other major websites (like Google and Facebook which already embed our content), but with the wide range of educational reusers (GLAMs, schools, teachers, students, etc) who want to use our content. Astinson (WMF) (talk) 16:44, 26 September 2016 (UTC)
- I think this is a brilliant idea to extend reach. MKramer (WMF) (talk) 17:41, 27 September 2016 (UTC)
- It´s a great idea to expand the content. I totally suport this project! Giselle Bordoy (WMAR) (talk) 16:43, 28 September 2016 (UTC)
Support Wikipedia has always been about finding easy ways for people to create and disseminate free content, and this is idea is firmly in that tradition. Gamaliel (talk) 18:19, 28 September 2016 (UTC)
- yes nice scope of work to concentrate the mind. wikidata may make it easier, but yet another reason for metadata cleanup. 184.108.40.206 22:22, 28 September 2016 (UTC)
- This is an important part of the role of Wikimedia which we have not been very good at yet. If organisations start to use Wikimedia regularly for content they will experience the benefits of open licensing and perhaps be more open to contributing content. John Cummings (talk) 14:45, 3 October 2016 (UTC)
- This sounds like a really really good idea! Especially considering how this can draw traffic not just to the specifically embedded content but to lesser known wiki's. Perhaps another really cool thing would be a way to highlight a phrase or a paragraph in your embedded link, kind of like how Medium does, so you could even easily embed a specific section of a specific resource. Omarshehata (talk) 02:15, 5 December 2016 (UTC)
- Support - though some sort of stable revision system should be worked out so that vandals aren't incentivized. Mark Schierbecker (talk) 13:50, 14 December 2016 (UTC)
- Advisor I would love to help with an outreach strategy and/or any documentation to help people as this project is underway. MKramer (WMF) (talk) 17:42, 27 September 2016 (UTC)
- Endorse and offer to advise/test - An idea that's long overdue. We haven't been active in providing this kind of service, and have left it to others to format our content, when we probably know best how to do it. -- Fuzheado (talk) 17:00, 28 September 2016 (UTC)
- Developer I just created a wrapper to extract attribution information for images from Wikimedia Commons: https://packagist.org/packages/gbv/image-attribution. See https://image-attribution.herokuapp.com/api.php?image=Amelia_Earhart,_circa_1928.jpg for an example. JakobVoss (talk) 17:40, 28 September 2016 (UTC)