11 insights after 11 years with the functional database Datomic
summary
Eleven years with Datomic yields concrete takeaways: Datomic models data as EAV plus transaction not time so facts never change and every write is appended, allowing full history and as-of queries to replay past states; transactions are handled by a separate transactor while peers read directly from pluggable storage (DynamoDB, Cassandra, SQL), letting reads scale without blocking writes; Datalog queries traverse entity graphs, retractions record intent instead of deleting, and transactions can carry metadata like actor and reason.
AI-written summary · about
chapters (18)
- 0:00Introduction
- 1:25Advice for programmers
- 3:34What is Datomic
- 5:45Entity Attribute Value model
- 8:12Datomic as graph DB
- 10:21Live Clojure demo: Datomic queries
- 13:57Datomic’s immutable facts
- 15:46Timestamps misleading
- 19:00Datalog queries
- 21:10Temporal queries & history
- 22:58Reports from past states
- 24:22Operational model & scalability
- 27:57Datalog’s power
- 29:04Transactions as entities
- 31:55Adding & retracting data
- 33:00Working with transactions
- 35:11Treat user data like code
- 36:43Summary
related talks
37m11 insights after 11 years with the functional database Datomic
40mThe power of Datomic database
46mThe Datomic Architecture and Data Model
51mDatomic
52mEnhancing Datalog for Datomic
40mDatomic: The Most Innovative DB You've Never Heard Of
42mYet & Datomic Immutable Facts Mutated Our Stack
41mExploring four hidden superpowers of Datomic