Digger began in late 2020 as a project to categorize music for contextual retrieval, where a context is defined as any listening situation you might otherwise create a playlist for. The idea was to take a myriad of contexts, factor out the common elements, and support autoplay without having to make playlists. The belief this was possible grew out of the SlideMatch project in 2009, which used 10 scalar characterizations to match music for retrieval. The challenge for Digger was to create similar retrieval power with far fewer song data requirements.
By combining dual scalar values with well chosen keywords, Digger was able to achieve substantial retrieval power. For example Instead of playlists for "Brunch", "Beach", "Dinner", "Party" and such, a single "Social" keyword filtered by approachability and energy level provides equivalent and more flexible selection. Retrieval processing prefers what was least recently played, so the app will "dig" through all matching songs in the least boring order. Digger was initially tested on a diverse collection of 10k+ songs to verify the rating process was reasonable and contextual retrieval worked well.
Switching between playback views while listening is seamless. In album view, playback proceeds sequentially through tracks and then stops. In deck view, playback is continuous. Digger prefers songs you have listened to least recently, and it also avoids overplaying in deck mode by respecting the playback frequency of each song. The maximum deck playback frequency is once every 24 hours. If you flag a song as tired, frequency moves from once every 90 days to once a year depending how often you skip the song. Because Digger prefers least recently played songs, frequency tuning is not usually necessary, but when you need it, it's there. Album view and search do not follow frequency constraints.
The Digger player panel lets you capture much of what you feel without having to come up with descriptive words for every song. The knobs, sliders, and keyword toggles are tools for active listening that connect music to past, present and future you. They make it easy to record the basics without having to put your feelings into words. It's an enhanced listening experience that works for you and grows with you over time.
Deck range selectors and three-way keyword toggles let you pull music from your collection matching a wide range of listening situations.
Initially created as a mechanism to synchronize Digger data across devices with bare minimum personal identification, DiggerHub also provides for structured collaboration that respects your time:
Find friends and music fans by their Digger name. If you haven't added any friends yet, and you want to get a feel for how things work, you can click "Connect Me" to add whoever the hub picks for you.
See how much music you have in common. If you have songs in your collection that you haven't rated yet, your friends provide default rating information you can use for contextual retrieval. The "rcv" tab shows how many default ratings you've received. The "snd" tab shows how many default ratings they have received from you. If you remove a friend, their default ratings are also removed. When you play a song, any default rating information for that song becomes yours.
See what your friends have shared recently. Get automatic recommendations matching their ratings to your listening parameters.
The Digger project, mobile platform apps, and DiggerHub are all OPENSOURCE.