Blog
All articles. All pillars. All authors.
The Never Type and Exhaustiveness Checking
How FLIN's Never type and exhaustiveness checking work together to guarantee that every code path is handled -- the compiler-enforced safety net for pattern matching and control flow.
Generic Bounds and Where Clauses
How FLIN implements generic bounds and where clauses -- constraining type parameters with traits, merging inline and where syntax, and validating constraints at compile time.
While-Let Loops and Break With Value
How we implemented while-let loops and break-with-value in FLIN -- pattern-driven iteration and loops that produce values, inspired by Rust's control flow.
Labeled Loops and Or-Patterns
How we implemented labeled loops and or-patterns in FLIN -- breaking from outer loops by name, combining match arms with pipe syntax, and the compiler changes that support them.
Every Entity Remembers Everything: The Temporal Model
How we designed FLIN's temporal model so every entity automatically tracks its complete history -- zero configuration, zero boilerplate, full time-travel out of the box.
Transactions and Continuous Backup
How we implemented ACID transactions with savepoints, full and incremental backups with Zstd compression, continuous WAL streaming, and scheduled backup rotation in FlinDB.
Tree Traversal and Integration Testing
How we tested hierarchical data, recursive queries, and tree traversal in FlinDB -- and why integration tests that mimic real applications matter more than unit tests.
Database Encryption and Configuration
How FlinDB implements AES-256-GCM encryption at rest with Argon2id key derivation, and a native FLIN configuration system with environment modes and variable overrides.
System Prompts That Teach: Anti-Cheating, Socratic Method, and Grade-Level Adaptation
Six-layer prompt assembly: root identity, school hours detection, grade overlays from CP to Terminale, subject expertise, and exam mode. How we taught the AI to teach.
The EAVT Storage Model
How FlinDB's Entity-Attribute-Value-Time event sourcing model provides complete audit trails, temporal queries, and entity replay -- inspired by Datomic and built in Rust.
FlinDB Hardening for Production
How Session 308 hardened FlinDB for production with CRC-32 WAL checksums, auto-checkpointing, cross-platform file locking, per-entity-type data files, history deduplication, and schema persistence.
Graph Queries and Semantic Search
How FlinDB implements graph traversal algorithms and AI-powered semantic search with BM25, vector similarity, and hybrid Reciprocal Rank Fusion -- all built into a single embedded database.