Keyboards no longer jam, yet layouts stay locked by network effects rather than engineering constraints

What it is

After mechanical jamming disappeared, letter-to-key mappings still clustered on a handful of layouts. What pins them is not switch travel or the scan matrix, but network effects: the more people type a layout, the more textbooks, job ads, rental machines, support scripts, and printed keycaps converge on it, cancelling the private gain of switching. Engineering can remap freely; social complementary assets cannot be remapped as cheaply as a switch.

Why it happens

A layout is both a skill that must be automated and a standard that is valuable only if many parties adopt it. Factory-printed caps, preinstalled input methods, exam-room PCs, and hot-desk hardware form one installed base. Remapping a machine to Colemak or Dvorak immediately hits three frictions: legends no longer match, a visitor cannot type, and mnemonic shortcuts drift with letter positions. Friction rises with the number of users of the incumbent layout; that is lock-in. A new layout can still win a lab speed test. Lab speed is individual performance; network effects are system complementary cost. They are not commensurate, so a typing contest cannot adjudicate them.

Studying it

Observe installed base and complementary goods, not only laboratory WPM: rank of a layout in the OS default list, share of retail keycaps printed for it, whether hiring and training materials name it, and whether people dare remap public machines. Novice-from-scratch experiments measure learnability, not conversion friction for an existing population. The economic-history debate on whether QWERTY is inefficient lock-in (David versus Liebowitz–Margolis) is a reminder that even a faster layout can remain rare if coordination fails. Stability in the data is not evidence of ergonomic optimality.

Where it stops holding

Network effects are not unbreakable. Closed devices—game handhelds, industrial panels, single-purpose terminals—do not share a typing skill, so nonstandard alphabets are safe. Touchscreens can swap layouts per app, thinning the installed base. On open desktops, software remapping can be nearly free and lock-in still holds if caps, classrooms, and docs stay printed to the old map. Small language communities also struggle to accumulate tutorials and correction dictionaries, which amplifies rather than dissolves the effect.

Applying it

• For mass-market hardware, ship alternative layouts as downloadable maps and swappable caps rather than expecting retail boards to reprint.
• In settings, offer a layout preview and one-tap restore to the public default, lowering the social risk of remapping.
• If an internal tool mandates an uncommon layout, supply physical legends and onboarding practice; otherwise the complementary cost is dumped on every collaborator.
• Verify by sampling rental machines, meeting-room keyboards, and support scripts: after a remap, can someone still complete a collaborative typing task? If not, lock-in is in the complements, not the switch.

Related

• Same group: C6.17.1 QWERTY originally separated keys to avoid typebar jams, not to match modern key actuation · C6.17.3 In multi-person collaboration, sharing one layout outweighs individual switching gains · C6.17.4 OS and app defaults that preinstall a layout reinforce the same path dependence
• Adjacent: C6.01 QWERTY and layouts · C6.03 Shortcut systems
• Search: network effects · installed base · keyboard layout lock-in