Pyspread – Pythonic Spreadsheet

Original Article ↗ Hacker News Discussion ↗

Documentation & Onboarding

  • Several commenters like the idea and UI but say the docs are too thin, especially for semi-technical users.
  • Suggested model: follow Pandas-style “install + 10-minute tutorial” docs.
  • Documentation is seen as a good entry point for new contributors, but some push back on devs implicitly offloading “unglamorous” work to others.

Installation & Python Packaging

  • Multiple users report pip installation failures on mainstream distros (Mint, Ubuntu), especially when using system Python.
  • Others insist users should avoid system Python, use virtual environments or pipx, and blame user setup; this is countered as unfair when the official instructions say “pip install”.
  • Some success stories: works in a venv on Fedora, via pipx, via Debian packages, and with nix run nixpkgs#pyspread.
  • Broader frustration with Python’s fragmented packaging/venv story and unclear guidance for end-user applications.

Use Cases & Positioning vs Other Tools

  • Compared to Jupyter: spreadsheets are reactive graphs where cell order is irrelevant; notebooks have order-dependent, mutable state.
  • Pyspread is seen as powerful for “improvisational” analysis for non-developers, similar to why Excel dominates.
  • Some ask how it compares to Emacs org-mode spreadsheets and to Python-in-Excel or services like RowZero and Grist.

Spreadsheet Model & Technical Design

  • Praised for numeric row and column indices (matrix-like), avoiding Excel’s A-Z column scheme; others defend named ranges/tables in Excel instead.
  • A key criticism: it appears not to track cell dependencies and recomputes all cells on change.
  • Commenters discuss that proper dependency tracking with arbitrary Python is hard, but suggest runtime tracking of S[] accesses or explicit dependency declarations; side effects and dynamic ranges make this tricky.

Language Choice & Cell Semantics

  • Some argue Python is awkward for cell formulas: one-liners are clumsy, multiline code is visually ugly in cells due to indentation.
  • Others suggest Lisp-like languages might better suit a code-in-cells model.

Licensing, Platform Support, and Ecosystem

  • GPLv3 license is noted as preventing inclusion in some projects (e.g., FreeCAD’s spreadsheet workbench).
  • macOS is officially “unsupported” in docs, which worries non-technical users, though some report pip install working there.
  • The clearly written “Target User Group” section on the site is widely praised for stating who the tool is and is not for.