Offline MVP
This extends basic feature completeness to encompass excision, noHistory, import and export, schema evolution and migration, logging, documentation, asynchronous writes, concurrent read isolation, and query preparation/caching
Mentat stores a sequence of changes. That makes replication and synchronization more tractable than stores that only maintain the current state of the world. This milestone defines the ability for a Mentat store to be replicated from a primary to a secondary, and for two or more stores to be synchronized via a centralized clearing-house.
One of our core value propositions is exposing the transaction log to applications. This entails a bunch of query work to allow individual statements and the query as a whole to examine the transaction log or a historical version of the database, and a little core work to allow callers to work with database instances.
Richer types
This milestone encompasses a variety of datatypes that might be stored in Mentat:
- Timestamps, UUIDs, interned keywords, bignums, and URIs — Datomic parity.
- JSON, which will be particularly useful for JS consumers, and opens the door to JS-aware predicates and destructuring.
- Geospatial data, using RTree or SpatiaLite.
This milestone covers description, maintenance, and use of materialized views and in-memory caches.