They don't make 'em like that any more: Borland Turbo Pascal 7

Nostalgia for Borland Turbo Pascal 7 and its DOS/Windows-era IDE highlights how fast, lightweight, and self-contained developer tools once were, especially compared to today’s RAM-hungry, browser-like environments such as VS Code. Commenters recall learning to program and even earning early income with Turbo Pascal and Delphi, arguing that these tools proved strongly typed, efficient languages could be both accessible and powerful. The conversation broadens into worries about modern software bloat, deepening abstraction layers, over-reliance on AI and complex stacks, and whether fewer people now have the skills to build or fully understand systems as lean and tightly integrated as those older tools.

Turbo Pascal as a formative tool

  • Many learned “real programming” on Turbo Pascal (TP), often as teenagers, after dabbling in BASIC.
  • TP is remembered as fast, integrated, and approachable: instant compiles, inline assembly, strong typing, and a tight TUI IDE.
  • Several describe TP or Delphi projects as first paid work or school capstones, shaping later careers and even major purchases.
  • Some recall TP as the first serious exposure to concepts like OOP and pointers, with clear “click” moments after repeated attempts.

Comparisons to Modern IDEs and Tooling

  • TP’s responsiveness and low resource use are contrasted with today’s heavy IDEs, especially VS Code–style, browser-based stacks.
  • Critics highlight high RAM use, sluggish UI, security issues, and centralized extension ecosystems.
  • Defenders argue resources are abundant, hardware is powerful, and rich features and convenience outweigh inefficiency.
  • Counterpoint: efficient tools allow more concurrent workloads and remain important on constrained systems.

OS Context: DOS, Windows 3.x, and OS/2

  • Nostalgia for running TP under DOS and early Windows, including Windows 3.1 atop DOS.
  • Debate over OS/2’s popularity vs Windows 3.1: some recall OS/2 dominance in specific “bubbles” (IBM shops, certain regions); others rarely saw it outside stores.
  • Limited stats mentioned but no clear consensus; overall popularity remains “unclear.”

AI, Abstraction Depth, and Human Expertise

  • One line of discussion worries that few people now can build tools like TP; software development is becoming semi-skilled.
  • Some see hope in AI as a persistent, stack-spanning intelligence that can manage modern complexity.
  • Others insist on preserving human knowledge: AI may be unavailable in disasters, and over-reliance is risky.
  • There’s concern about centralized model control and whether LLMs, as “beautiful diamonds,” are too brittle compared to messy but resilient “big ball of mud” systems.

Software Bloat and Framework Overheads

  • Several lament modern desktop and chat apps using hundreds of MB of RAM compared to 1990s tools using single-digit MB.
  • Causes debated: deep abstraction stacks vs high-res graphics, GPU pipelines, fonts, and double-buffering.
  • Case studies (Electron apps, GTK2→GTK4 migration) show large memory jumps, sometimes traced to GPU driver initialization.
  • Some argue we’ve optimized for developer convenience over end-user efficiency.

Pascal, Delphi, and Successors

  • Delphi is praised as a technically superior successor (faster, better language than VB) that nonetheless lost to Microsoft’s ecosystem.
  • Some recount cultural bias in shops where C/C++ developers were valued more than Delphi users.
  • FreePascal and Lazarus are noted as modern TP/Delphi-like environments that keep the lineage alive.