Past and Present Futures of User Interface Design

Keyboard, Mouse, and Tactile Interfaces

• Many commenters stress that both mouse and keyboard are indispensable; preference often depends on task and learned skill.
• Some find mouse-based navigation faster for “jump to spot in code,” others argue advanced editor features (Vim motions, plugins like Easymotion) make keyboard vastly faster once learned.
• Several propose richer physical/tactile setups: knobs, macro pads, MIDI keyboards, HOTAS-like controls, and game-style keypads (DataHand, Azeron, Tap).
• Physical buttons/knobs are especially praised in domains needing speed and reliability (video editing, aviation, cash registers, cars).

Command Palettes and “Find Anywhere”

• The rise of global “run command” / “find anywhere” palettes is seen as a major power-user improvement, reducing the need to memorize shortcuts or dig through menus.
• Critiques: in some apps they replace explorable menu bars; one suggestion is to have search highlight the matching menu item to teach the UI’s structure over time.
• OS-level versions (e.g., menu search) are appreciated when they exist.

Why the Keyboard Persists

• Repeated attempts in industry labs to “kill the keyboard” reportedly failed; alternatives looked cool in demos but were worse in real use.
• Consensus: keyboard, mouse, touch, and some voice will coexist, each fitting different contexts; interfaces should be descriptivist (fit existing behavior), not prescriptive.
• Dynamic-key displays (Optimus Maximus, Stream Deck, Flux) are admired conceptually, but some argue they undermine muscle memory if layouts change too much.

Voice, Automation, and Non-Visual UIs

• Many see voice as niche: great for driving, accessibility, or classrooms, but socially inappropriate or privacy-breaking in most public or work settings.
• Ideas like lip-reading or subvocal sensors are discussed but viewed as technically and socially uncertain.
• One thread explores “automation as output”: browser/assistant systems where the UI’s result is actions taken, not screens rendered.

AI Chat as Interface Layer

• One camp expects AI chat to remove large swaths of traditional GUI (IDE refactoring menus, complex 3D tools), shifting toward iterative natural-language workflows.
• Others are strongly skeptical, citing hallucinations, fragile tests, increased bugs and code churn, and diminishing returns from scaling current LLMs.
• Even skeptics concede it can help with boilerplate; disagreement centers on depth of reliance, not existence.

Touchscreens vs Physical Controls

• Touch is praised for flexibility and multilingual input, but widely criticized in cars and appliances: unsafe, slow, hard to use when wet/dirty.
• Commenters note touch is “cheaper for manufacturers” and visually marketable, not necessarily better for users.
• Recurrent theme: knobs, switches, and mechanical keyboards remain superior where tactile feedback and eyes-free operation matter.

Design Philosophy: Familiarity Over Flash

• Strong support for consistency, idiomatic controls, and building on decades of desktop conventions.
• Novel UIs are seen as high-risk, hard to make accessible, and often driven by marketing or sci-fi aesthetics rather than real tasks.