Storing knowledge in a single long plain text file

Original Article ↗ Hacker News Discussion ↗

Perceived Idea & Goals

  • Proposal: store all tabular/knowledge data in one (or logically one) plain-text file, parsed via an indentation-based “ScrollSet”/Tree Notation grammar.
  • Data model: “concepts” ≈ records, “measures/measurements” ≈ fields/cells; strongly typed, hierarchical, git-backed, compiled to CSV/TSV/JSON.
  • Claim: system scales to large datasets and enables fully auditable, schema‑rich knowledge bases using only spaces/newlines plus a small syntax.

Novelty vs Prior Art

  • Many readers see this as a reinvention of old ideas: plain-text storage, Unix “everything is text,” and semantic/structured data formats.
  • Specific predecessors cited: GNU Recutils, Plan 9’s ndb, RDF/semantic web, CSV/JSON/YAML/TOML, TiddlyWiki, Wikidata.
  • Some argue the paper underplays prior work and should foreground comparisons more explicitly.
  • The article is updated to add Recutils and to list claimed advantages: easier hierarchies, less encoding overhead, first‑class comments, stronger integrity via “parsers.”

Tone, Presentation, and Reception

  • Title and style are widely read as tongue‑in‑cheek; some interpret it as satire or even “TimeCube‑like” grandiosity.
  • Others think the core idea is sincere but oversold, and that the ambitious framing distracts from the technical merits.
  • There’s debate over whether provocative tone attracts useful critique or just emotional backlash.

Alternative Tools and Related Approaches

  • Comparisons drawn to: Org mode, Emacs workflows, Obsidian (Markdown graph of files), “one big text file” note systems, Canon Cat, Notion “vault” tables, and conventional databases with views.
  • Several participants mention that high‑earning bug bounty hunters reportedly use a single large stuff.txt plus grep.

Technical Questions and Critiques

  • Concerns about namespace and indexing complexity; risk of “namespace hell.”
  • Questions about write safety, corruption, and reliance on git; suggestions to add hashing/copy‑on‑write ideas.
  • Confusion over terminology (“parsers,” “concepts,” “measurements”) and how definitions vs. data are distinguished.
  • Some feel the focus on the indentation trick and minimal syntax harms readability and visual salience compared to formats like Markdown.
  • Others argue this overcomplicates what databases already solve while losing features like robust querying and integrity constraints.

Enthusiasm and Potential

  • A subset of commenters are excited by text‑centric, programmable knowledge bases, especially combined with AI and Emacs‑like environments.
  • The extreme simplicity of plain text as the base abstraction is praised, even by some skeptics.