One-handed layout and font evidence

Jianlai key feedback: compare four states of bubbles and bottom borders

Arrange the real key-bubble and bottom-border toggles into four independent states, checking transient feedback and static edges with one fixed card. Deliver a state matrix and actual input. A bubble is not committed text, and a border change does not establish changed touch areas, sound or haptics.

English translation of the Chinese original, translated on . Original publication dates, attribution, code and version notes are retained.

Arrange the real key-bubble and bottom-border toggles into four independent states, checking transient feedback and static edges with one fixed card. Deliver a state matrix and actual input. A bubble is not committed text, and a border change does not establish changed touch areas, sound or haptics.

1. Fix the on-screen keyboard and both original values

Confirm an actual Jianlai on-screen keyboard in an ordinary note, recording device, orientation, scheme, theme, sizes and both original toggle values. Hardware typing without observable key taps cannot substitute for this exercise; wait for suitable screen input. The image shows independent controls, but visibility under the current theme/layout still needs observation. Decorative styles may obscure differences. Preserve documents, use a private draft and leave sound/haptics unchanged. If installed labels differ, identify scope before acting. An unavailable control is a limitation, not permission to toggle neighbors to obtain four pictures.

2. Separate press feedback from a persistent edge

Bubble feedback concerns what appears during a key action, while a bottom border is a persistent edge that can be inspected without pressing. One still screenshot need not prove both. Enter the card below in the original state, check body text and record feedback actually seen. Missing a fleeting bubble in a screenshot does not prove the toggle was off; preserve observation conditions. Candidate text is not a key bubble, and theme decoration is not automatically the controlled border. Compare the same identifiable letter key, checking digit entry separately without requiring identical behavior from every symbol or function key.

B35

3. Use a four-cell plan, not an imported configuration

Draw four external cells A through D using the table below, giving each toggle on and off states. These letters are trial labels, not Jianlai configuration files, theme identifiers or commands. Read actual state back rather than rely on tap counts. The order A, B, D, C changes only one toggle per transition, but results retain their cell identities rather than redefine the table by execution order. Return to the same host and summon the keyboard after each setting. Pending composition need not survive; do not let an old candidate obscure the current key observation.

A: bubble on;  bottom border on
B: bubble on;  bottom border off
C: bubble off; bottom border on
D: bubble off; bottom border off
One-toggle sequence: A -> B -> D -> C

4. Compare borders only between A and B

In A, verify both toggles on, enter the card identically and record press feedback separately from the resting edge. Turn off only the bottom border for B, retaining bubble state, font, theme and layout, then repeat the same letter and text check. Record a clearly changed edge or uncertainty, not an imagined thin line. Observe B’s bubble too rather than copying a pass because its setting is unchanged. If decoration obscures borders, a more legible theme may be a separate later experiment, not an unrecorded change between A and B presented as a one-variable comparison.

5. Compare bubbles and borders through B, D and C

From B turn off only bubbles to reach D, both off, then use the same actual key and check text. Turn on only the bottom border for C, leaving bubbles off, and inspect the resting edge. Keep settings, visible feedback and committed card separate. No bubble does not establish no input, and committed text does not establish a bubble. If this layout exposes no recognizable bubble, record the limitation: settings and borders can still be logged without passing bubble observation. Hold sounds, haptics and key images fixed so their effects are not attributed to these toggles. List any extra action as a condition change.

6. Verify text separately from feedback differences

Check B, case, three and five character by character in each actual committed result, not a candidate preview. Record casing or substitution errors before repairing a practice copy instead of overwriting observations for a correct screenshot. Matching text across four trials establishes only this card’s commitment under those conditions, not improved feedback, error rates or speed for everything. A bottom border does not measure hit targets, and a bubble is neither haptic nor audio feedback. Separate these judgments so two visual controls do not become a certificate for dictionaries, permissions or every device interaction.

7. Diagnose values, theme and observation timing

For unclear differences, verify readback, Jianlai identity and unchanged theme, then whether the transient key action was actually observed. A released-key screenshot can miss a bubble and cannot prove missing implementation. Indistinct borders call for recording decoration and region, not automatic failure or a crash diagnosis. Change one condition at a time; a new theme needs a new matrix instead of results pasted into old cells. Invent no forced-bubble screenshot or feedback-duration setting. Do not uninstall or clear data, and retain recoverable documents and daily settings.

8. Restore both values and retain all four cells

Restore both recorded values and reopen ordinary input to check the card and feedback, passing restoration only if performed. Save readback settings, observations, text and not-applicable states for all four cells rather than tick the whole table. With unknown originals, label a chosen replacement instead of automatic restoration. New themes, schemes or systems can change visibility and need renewed comparisons. One letter’s transient feedback does not establish every key’s behavior. This is an evidence matrix separating visual controls from input, not sound/haptic configuration or touch-performance testing.

Separate sound and haptics; Keycaps, candidates and body text.

Checked on 18 November 2026: Apple’s Finland lookup and China lookup both list Jianlai 1.0.5, requiring iOS / iPadOS 16.0 or later. The official store describes themes, backgrounds, fonts and multilingual input. Use controls actually present in the installed version; system editing and host processing are distinguished separately. These are repeatable instructions and paper expectations, not claims of completed device input, host compatibility or format-parser tests.

Method references: Official Apple keyboard-settings image at full reading size.