I'm currently working on a photosharing site. It's not up and running yet, but you can find some info at fok.al.
The details of what Fokal is supposed to be a still up in the air. Here are some of the core principles:
- Image Publishing model similar to Noun Project
- Programmatic Access through a simple API
- Machine Learning for Image Search and Recommendation
- Emphasis on Metadata especially geographic data
- Robust Image Filtering
The core thesis is that exposing metadata to the user is useful and that a mixture of machine intelligence and human recommendation can yield valuable results in a very subjective aesthetic domain.
There is a gap between current solutions for managing your own photo collection in say Google Photos, or Apple Photos and looking through what someone has posted online. There is value in applying similar principles to public photosharing websites.
The goal of Fokal is to strike a balance between user curation and recommendation, and using machine learning to do the same.
Update - June 2019
I decieded to shut down Fokal in September of 2018. I wrote about why in a post titled, Fokal: A Retrospective.