The legacy of NeXT lives on in OS X (2012)

Original Article ↗ Hacker News Discussion ↗

NeXT legacy in macOS

  • Commenters note how pervasive NeXT-era APIs remain (e.g., NS* classes, NeXT-style streams in prefs, NX artifacts in headers).
  • Early Mac OS X felt almost like NeXTSTEP with a new window manager and frameworks on top; filesystem layout and many tools were nearly identical.
  • Debate over what “NS” stands for: NeXTSTEP, NeXT Software, or “New System”; consensus is that it predates Sun’s OpenStep collaboration.

Frameworks: SwiftUI, AppKit, UIKit, tools

  • Apple clearly pushes SwiftUI as the future, but AppKit/UIKit are still actively maintained and interoperable with SwiftUI.
  • Some find SwiftUI powerful but immature: weak docs, poor guidance when diverging from standard look, rough edges on gestures and platform parity.
  • Others report success with highly customized SwiftUI UIs and see it as the natural next-gen framework.
  • Strong split on Auto Layout and Interface Builder: some love constraint-based layout and dislike nib/storyboard indirection; others still rely on storyboards despite editor bugs and slowness.

Documentation and platform strategy

  • Many complain Apple docs have deteriorated: sparse, little rationale or examples, often worse than archived docs.
  • Some believe Apple underinvests in third‑party dev experience because they increasingly clone or bundle key apps themselves.
  • Nostalgia for an era when Apple’s docs were considered industry-leading.

UI / UX design and nostalgia

  • Extensive nostalgia for early–mid OS X (Tiger–Snow Leopard, Mountain Lion): colorful icons, clear affordances, “gel” buttons, rich but readable chrome.
  • Others counter that people romanticize the past and ignore slow hardware and limited functionality.
  • Strong criticism of modern flat/minimal UI: reduced contrast, white‑on‑white, tiny/hidden scrollbars, inconsistent first‑party apps, and loss of visual joy.
  • Some argue flatness and simplification are driven by high‑DPI screens, mobile constraints, and fashion rather than necessity; others insist rich UI still looks fine on Retina.
  • Theming in classic Mac OS (Copland, Kaleidoscope) is remembered fondly; modern macOS is seen as locked-down and visually monotonous.

Unix, development, and distribution model

  • Early OS X’s “real Unix” shell was a major draw; macOS is now seen by some as stagnant or less dev-focused compared to modern Linux desktops.
  • Others still view macOS as the best “just works” Unix desktop, especially on laptops, with Linux praised but still hit-or-miss on hardware and power management.
  • Package management is contentious:
    • Some argue the Mac App Store is the official “package service” and works well for end‑users but is sandboxed and limited.
    • Developers rely on Homebrew, Nix, etc., but worry about bloat, fragmentation (multiple updaters), and potential for “messing up” systems.
    • Others see unified system+user package managers (like pacman) as either a strength (single update command) or a liability (risk of breaking core tools).
  • On distribution/signing:
    • iOS still requires a paid dev program for practical distribution; sideloading is constrained.
    • macOS can run unsigned apps but with increasing friction; some see this as security and spam control, others as a slow funnel into Apple’s signing/fee regime and erosion of desktop freedom.

Terminal and tooling UX

  • Terminal.app is praised as stable and “done”; some fear unnecessary rewrites.
  • Others want better ergonomics (cursor movement, per‑tab history, transparency, font rendering) and point to third‑party terminals (iTerm2, wezterm) as evidence of unmet needs.
  • Several note that many desired niceties already exist via keybindings (Emacs/readline shortcuts, Option-click, Option+arrows), suggesting discoverability rather than capability is the issue.

Objective‑C and language evolution

  • Learning Objective‑C in the 2000s/2010s is described as a history lesson in NeXT and Smalltalk-influenced design.
  • One camp thinks ObjC hasn’t “aged well” versus newer systems languages (Rust, Zig) but still had clever ideas (message passing, reflection, event loop).
  • Another camp argues ObjC has aged very well as a “C with objects” that’s cleaner and more dynamic than C++:
    • Full C superset, easy interop with C/C++, dynamic runtime, good collections and Unicode handling, and ARC relieving most memory-management burden.
    • Verbose syntax is seen as an asset for long-term maintainability.
  • Disagreement over how much ObjC makes sense outside Apple platforms; some point to GNUstep and cross‑platform use, others say practical, non‑Apple usage is negligible and the real alternative is Swift.

WebObjects and web framework influence

  • WebObjects and EOF are praised as ahead of their time: early ORM, “direct to web” no‑code tooling, component‑based server-side UI similar in spirit to modern HTMX plus server-rendered components.
  • Some still use WO or WO‑inspired frameworks and argue many later ORMs feel worse than EOF.
  • There’s mention of historical influence on Java EE and ongoing efforts to recreate WO-like frameworks today.

Apple OS history: Copland, BeOS, NeXT

  • Broad agreement that Mac OS X is essentially NeXTSTEP/OpenStep underneath with:
    • Quartz/Display PDF replacing Display PostScript.
    • Cocoa (NS*) plus Carbon and Classic as a carefully staged transition path.
  • Copland is cited as a classic case of overreach and poor project management: too many simultaneous reinventions (filesystem, OO API, UI toolkit, help, OpenDoc Finder) and failure to say “no.”
  • BeOS is remembered fondly for responsiveness and multimedia, but viewed as immature for Apple’s needs at the time (weak printing, single-user design, poor networking). Price expectations from Be were high relative to its completeness.
  • The NeXT acquisition is framed as primarily about getting a mature, Unix-based OS; Steve Jobs’s return and later dominance are seen as consequential but not necessarily the original driving motive.