I’ve been building a semantic workbench called Reckoner over the past year.
The core idea is treating datasets less like rigid tables and more like navigable semantic spaces you can define, refine, and compare through constraints and set operations.
The demo currently supports:
* semantic narrowing across 5W1H dimensions
* trie-style value exploration
* set ops (union/intersection/diff)
* SRF semantic bundle import/export
* DuckDB/Postgres-backed datasets
A major goal is reducing the cognitive friction between humans and structured information systems, especially for domain experts who understand the data deeply but aren’t SQL specialists.
Influences are very much Engelbart / Kay / Hopper style augmentation thinking.
Currently running as a Railway demo; a packaged Tauri desktop build is the next step.
abk9811•37m ago
The core idea is treating datasets less like rigid tables and more like navigable semantic spaces you can define, refine, and compare through constraints and set operations.
The demo currently supports:
* semantic narrowing across 5W1H dimensions * trie-style value exploration * set ops (union/intersection/diff) * SRF semantic bundle import/export * DuckDB/Postgres-backed datasets
A major goal is reducing the cognitive friction between humans and structured information systems, especially for domain experts who understand the data deeply but aren’t SQL specialists.
Influences are very much Engelbart / Kay / Hopper style augmentation thinking.
Currently running as a Railway demo; a packaged Tauri desktop build is the next step.
Demo: https://reckoner-production.up.railway.app/
GitHub: https://github.com/peirce-lang