ironfroggy
commented
11 years ago These can finally be built soon on top of Result watching.
Thinking of special stores, each given the name of another store the schema and either a map or reduce function. When the source is updated, the function is called and the special store updated.
They can be set as Atomic or Lazy
The data stored in them includes meta data referencing the source key or keys used to calculate each object.
The map function runs first, then the reduce. Obviously borrowed from CouchDB.
These would be stores which define handlers to convert SyncStore entries, and would allow for very basic map-reduce usage to clean up data that is synced into formats more suited for the end application. Also, the view store can define indexes on generated values.