Modeling your domain
summary
Argues that modeling domain data as EAVT (entity attribute value time) facts stored in an append-only log gives immutability, audit trails and one-way flows, while feeding multiple read-optimized indices (AEV, AV, inverted) for fast relational, graph and full‑text queries. Demonstrations reference DataScript in-browser datalog, Datomic-style separation of transactor/storage/peer, Lambda architecture at LinkedIn and Om Next on the client; claim is that this architecture simplifies evolution, testing and querying and that auditing is effectively free.
AI-written summary · about
chapters (39)
- 0:00Start
- 2:43INFORMATION SYSTEMS
- 2:57TECHNICAL DEBT.
- 3:08PREACHED TO COLLEAGUES
- 3:36MVC SOA DRY TDD
- 3:51SNAKE OIL
- 4:08ENCAPSULATION
- 6:51DUPLICATION
- 7:36RELIABILITY
- 7:47CONFIDENCE
- 8:14TYPES / SCHEMAS ASSERTIONS PROPERTY TESTING
- 8:55JUST TOOLS
- 9:06NOT RELIGION
- 9:33PROJECTS FAIL
- 9:46Not modeling
- 10:01SOUL-SEARCHING
- 11:10is immutable
- 12:03semantic
- 12:36recursive
- 12:47string boolean number
- 13:00symbol keyword #identity
- 15:53FACTS
- 16:12FLAT DATA WATER
- 22:14APPEND ONLY
- 22:57BUSINESS DATA SOURCE CODE
- 24:50QUERIES!
- 25:36AUDITING
- 26:03RELATIONAL (ROW)
- 26:31find ?name ?age :in $db ?id : where [?id :actor/name ?name] [?id : actor/age ?age]
- 27:43RELATIONAL (COLUMN)
- 28:34GRAPH
- 30:07inverted
- 31:14STREAM
- 33:26ROI
- 34:01THEORY APPLIED
- 35:02SIMULATION/ PROPERTY TESTING.
- 35:23LAMBDA ARCHITECTURE LAMBDA
- 35:57DATOMIC
- 36:45CRDT
