Glamorous Toolkit

Original Article ↗ Hacker News Discussion ↗

What Glamorous Toolkit Is (as described in the thread)

  • Positioned as a “moldable development environment”: a place to build many small, custom tools (“contextual micro tools”) that explain aspects of a system.
  • Focus is on reading/understanding systems, not just editing code: exploring data, control flow, dependencies, APIs, logs, etc.
  • Tools operate on live objects, not static snapshots, so visualizations/debuggers/inspectors are just alternate object views in the same environment.
  • Used on real projects (including 100+ dev teams and legacy modernization) and supports multiple languages via bridges (not just Smalltalk).

Comparisons to Existing Tools (Jupyter, Emacs, Unix, IDEs)

  • Many see it as “like Jupyter/IPython notebooks” or “supercharged Emacs/Spacemacs,” with richer GUI and deep inspect/visualize abilities.
  • Proponents argue GT goes further:
    • Unified environment where thousands of tiny tools per system coexist and are reused.
    • Dynamic exploration (context-following inspectors, driller, debugger) vs mostly linear “defined” notebooks.
  • Skeptics say similar outcomes can be achieved by gluing standard tools (Emacs, Unix, ggplot, R, Python), and question if GT’s benefits justify switching costs.

Smalltalk, Integration, and Adoption Concerns

  • Smalltalk/Pharo basis is seen as both enabling (live image, introspection) and a major barrier (unfamiliar culture, fewer libraries, non-native UI, corporate skepticism).
  • GT authors stress:
    • It’s meant as an environment for arbitrary systems, not a pitch to rewrite everything in Smalltalk.
    • Sources live in Git; there is integration with Python, JS, Rust, LSP/DAP, web views, SVG, etc.
  • Some worry there are few high-profile applications beyond GT itself; others note Smalltalk historically powered serious but often non-public systems.

UX, Documentation, and Messaging Critique

  • Strong recurring complaint: website and terminology (“moldable development”, “contextual micro tools”, “systems”) are opaque; users can’t tell what GT does in 15 seconds.
  • Long videos and an in-tool book are seen as too heavy for initial onboarding; people want short, concrete examples (“I have X code, what do I see?”).
  • Authors acknowledge communication problems, have tweaked the homepage and point to online docs/books, Discord, and recorded sessions; they explicitly seek help improving messaging.

AI / LLMs and Future Direction

  • Some argue tool-using AI may compete with this paradigm; others (including GT authors) see LLMs as complementary engines, with GT providing richer human-facing interfaces and exploratory tools.