Page MenuHomePhabricator

Search through visual identification
Open, LowPublicFeature

Description

Instead of typing, I would love that the camera of my phone could be able to identify a word or a sentence in the frame and provide me suggestions of contents related to it in the wikiverse.

When I am walking in the street and wonder about the name of the people honored by the street's name, I need to stop and type on my phone to get the answer. Then, I get to a library and want to check if an old book I like is available on Wikisource, I need to type the whole name on my phone. In the same book, there is a word I don't understand, so I type it in Wiktionary to get a definition. Tired of typing on my phone, I wonder what's more interesting to do in my city and I type the name on Wikivoyage search bar. I typed a lot. And I don't even had a look at the gorgeous gallery of pictures in Wikimedia Commons associated with my city!

I imagine it's a very specific technology here. But Google Translate can identify and suggest a translation, so why we couldn't have a similar way to enter into the wikiverse?
I suggested this idea into the Community Wishlist Survey 2019 and MSchottlender-WMF decline it in the behalf of the Tech Team because it's a huge project, but she suggested to copy the idea here, so here it is! Fell free to develop it if you like it!

Event Timeline

Is this about a Wikimedia mobile app? Or about using a MediaWiki website in a browser?

It could be through a dedicated app or included in the research options, I don't know which solution is preferable. As it is very crosswiki, it may be a separate app. As I am not familiar with the way this kind of technology work, I am afraid I can't be more specific.

Aklapper edited projects, added Technical-Tool-Request; removed Discovery-ARCHIVED.
Aklapper changed the subtype of this task from "Task" to "Feature Request".