※ Korean documentation is available at Korean README.
Roman Orthographic Simultaneous-input for English by eekdland (Sinseiki)
Alias: ROSE
(Project status: Experimental / On hold
ROSE demonstrates the feasibility of orthographic simultaneous input for English, but further development is currently paused pending better corpus-based and ergonomic analysis of English chord patterns.)
ROS-e is an experimental simultaneous-input method for English.
Unlike traditional stenography systems, ROS-e does not require dictionaries, briefs, or phonetic theories for general typing. Words can be generated directly from their spelling, allowing users to type unfamiliar words, identifiers, technical terms, and newly coined words without memorizing special shorthand codes. ROS-e can be used without a dictionary.
The project explores a different approach to English text entry:
- Orthography-based input instead of phonetic input
- Simultaneous key presses on a standard keyboard
- Direct spelling-to-output correspondence
- Dictionary-free typing for ordinary words
- Adaptive chord depth
- Onset–Nucleus–Coda structural organization
- Left-to-right ordering within each group
- Developer-friendly input for identifiers and technical vocabulary
ROS-e is inspired by Korean simultaneous-input keyboard research, particularly Semo-e and Sebeolsik-based designs.
Sebeolsik is a Korean keyboard family that separates onset, nucleus, and coda components into distinct input groups.
The central idea is that multiple letters can be entered together while preserving the natural spelling order of a word.
This project is currently in the Alpha stage as an AutoHotkey prototype. (The AutoHotkey code was created with the help of AI.)
The layout, grammar, and key mappings may change as practical testing continues.
The current layout was developed through repeated analysis of approximately ten million English characters, followed by practical typing experiments and iterative refinement.
Most English stenography systems achieve high compression by relying on phonetic theories, dictionaries, and memorized shorthand forms.
ROS-e explores a different direction.
ROS-e is not intended to compete directly with professional stenography systems.
Instead, it aims to provide a practical increase in typing efficiency on a standard keyboard while preserving direct correspondence between spelling and input.
The goal is to occupy a middle ground between conventional typing and full stenographic systems.
Instead of translating words into shorthand codes, ROS-e attempts to preserve the original spelling structure of English while allowing simultaneous input.
-
Download and install AutoHotkey from the official website:
-
Download
ros-e.ahk.https://raw.githubusercontent.com/Sinseiki/ROS-e_keyboard/refs/heads/main/ros-e.ahk
(If the file opens in your web browser instead of downloading, right-click the page and select "Save As" to save the file.)
-
Run
ros-e.ahk.
Antivirus False Positive Notice: As this program intercepts, rearranges and re-enters keyboard input,
it may trigger false positives in the behaviour-based detection mechanisms of some antivirus programs.
It is recommended that you add this script to the exceptions list of your antivirus program.
ROS-e uses Scroll Lock as its primary mode switch.
- Scroll Lock ON → ROS-e enabled
- Scroll Lock OFF → ROS-e disabled
- F9 → Alternative toggle key
The Scroll Lock indicator can therefore be used as a visual status indicator for ROS-e.
When ROS-e is disabled, all keys behave normally according to the active keyboard layout.
ROS-e attempts to preserve the intended spelling order even when individual key presses are not perfectly synchronized.
For example:
researchTeam
cambridgeUniversity
humanMind
A ROS-e chord is interpreted according to the intended letter order rather than the exact timing of individual key presses.
This approach may reduce a common class of typing errors known as the Transposed Letter Effect, where letters appear in the wrong order even though all required letters are present.
For software developers, mistakes such as:
resaerchTeam
cambrigdeUniversity
huamnMind
can be surprisingly difficult to notice during code review and debugging.
By preserving orthographic order within a chord, ROS-e aims to make such errors less likely during fast typing.
ROS-e is based on the following principles:
- Spelling comes first.
- Simultaneous input should remain readable and predictable.
- Users should be able to type unfamiliar words without consulting a dictionary.
- Abbreviations should be optional rather than mandatory.
- Standard keyboards should remain usable without special hardware.
And the current ROS-e layout was not derived from existing stenography layouts.
Instead, the key arrangement was developed through repeated analysis of approximately ten million characters of English text, combined with practical typing experiments and iterative layout revisions.
The goal was not to reproduce traditional stenographic theories, but to explore a simultaneous-input system that preserves English spelling order while remaining practical on a standard keyboard.
It is not to replace stenography, but to explore a practical middle ground between conventional typing and full stenographic systems.
ROS-e is not based on strict finger assignments.
Unlike many traditional keyboard layouts, ROS-e allows a degree of fingering flexibility in order to reduce awkward hand movements during simultaneous input.
The vowel area is shared between both hands.
Users may choose whichever hand provides the most comfortable movement for a particular chord.
As a result, the same vowel may be entered by either the left hand or the right hand depending on the context.
ROS-e generally follows conventional touch-typing principles.
However, simultaneous input occasionally permits alternative fingerings when they reduce awkward hand movements or finger collisions.
The objective is practical comfort rather than strict adherence to a single fingering method.
When a chord becomes uncomfortable, users are allowed to use a neighboring finger if it produces a more natural hand motion.
The goal is to minimize unnecessary strain rather than enforce a fixed typing technique.
The thumb is normally not required.
However, ROS-e allows occasional thumb participation when:
- a chord would otherwise cause finger collisions,
- a shorter finger would need to be stretched while a longer finger bends at the same time (like 'ac', 'al', 'li', 'mi', 'ur', 'ar', etc.),
- a chord would require an unusually wide hand stretch,
- or an alternative movement is significantly more comfortable.
Thumb usage is therefore optional and situational rather than mandatory.
In practice, the thumb may be used to reduce strain when simultaneous input would otherwise force the hand into an awkward or excessively stretched position.
ROS-e prioritizes practical comfort over theoretical finger assignments.
If multiple fingerings are possible, users are encouraged to choose the one that produces the most natural movement while preserving the intended spelling order.
ROS-e introduces two mechanisms called Onset Shift and Coda Shift.
Because English contains 26 letters but simultaneous input requires multiple positional groups, letters are separated into three functional categories:
- Onset
- Nucleus
- Coda
ROS-e reconstructs words by processing Onset, Nucleus, and Coda groups separately. Left-to-right ordering is applied within each group rather than across the entire chord.
Onset Shift and Coda Shift expand the effective capacity of the Onset and Coda groups.
This allows ROS-e to represent a larger set of letters than would otherwise fit within the available key positions while preserving the Onset–Nucleus–Coda structure and orthographic order.
English contains many common multi-letter vowel combinations, such as:
ee
ea
oo
ou
ie
oi
To support these patterns efficiently, ROS-e uses duplicated vowel positions for certain vowels.
This design allows common vowel combinations to be entered naturally while maintaining orthographic consistency.
ROS-e includes an optional physical-key chord abbreviation system, primarily intended to reduce the number of key presses required for frequently used words.
Abbreviations are not required for ordinary ROS-e input. Words, identifiers, and other arbitrary strings can still be entered directly from their spelling without relying on an abbreviation dictionary.
The default abbreviation set uses physical-key chords from the A+X(Q) family. This is only a convention for the distributed defaults, not a restriction of the abbreviation system itself.
Because assigning an abbreviation to a chord overrides the normal ROS-e output of that chord, it is generally preferable to choose chords that are unlikely to represent useful letter sequences in ordinary English spelling.
The distributed defaults use chords corresponding to Q without U for this reason. In ordinary English spelling, Q is most commonly followed by U, so Q combinations that do not contain U provide a relatively convenient space for optional abbreviations while reducing interference with normal orthographic input.
This is only a design choice for the default abbreviation set. Users may change, remove, or replace any of the default abbreviations according to their own needs.
Users may also add abbreviations by assigning text to any physical-key chord. If a chord is assigned to an abbreviation, that chord is reserved for the abbreviation and its original ROS-e output is no longer available through that chord. Users should therefore consider possible conflicts with normal ROS-e input when choosing their own abbreviation chords.
Abbreviations are matched by physical keys rather than by the letters that those keys would normally produce. This allows abbreviation definitions to remain independent of ROS-e's Onset, Nucleus, and Coda interpretation.
The order in which keys are written in an abbreviation definition does not matter. Chord IDs are canonicalized automatically, so different written orders of the same physical-key combination are treated as the same chord.
The purpose of this feature is not to turn ROS-e into a dictionary-dependent stenographic system. It provides an optional layer of user-controlled input compression while preserving direct orthographic input as the core method.
These words were selected based on frequently used terms by developers. As a general rule, default abbreviations used base forms for greater flexibility. ( https://zenodo.org/records/285293 )
(QWERTY Physical Keys)
- a + x + k → return
- a + x + l → function
- a + x + ; → public
- a + x + y → import
- a + x + u → void
- a + x + i → array
- a + x + o → static
- a + x + p → true
- a + x + / → false
- a + x + k + ; → private
- a + x + l + ; → break
- a + x + y + ; → double
- a + x + u + ; → class
- a + x + i + ; → final
- a + x + o + ; → protect
- a + x + p + ; → boolean
- a + x + m + ; → struct
- a + x + / + ; → extends
See this Using Custom Abbreviations for instructions on how to add, change abbreviations.
ROS-e does not introduce a separate mechanism for capitalization or system shortcuts.
When Shift, Ctrl, Alt, or Win is pressed, ROS-e temporarily falls back to the underlying keyboard layout.
As a result, users can continue to use familiar keyboard behavior without learning additional shortcut rules.
For example:
- Shift + letter → capital letter
- Ctrl + C → copy
- Ctrl + V → paste
- Ctrl + Z → undo
- Alt + Tab → task switching
- Win + R → Run dialog
- Win + E → File Explorer
This approach allows ROS-e to coexist with existing operating-system and development-tool workflows while minimizing interference with established shortcuts.
Some ROS-e letter assignments occupy positions normally used by punctuation keys.
As a result, punctuation may require modifier-based input.
For example:
- Shift + ; → ;
- Shift + ; twice → :
- Shift + / → /
- Shift + / twice → ?
The exact behavior may change during future development.
ROS-e is designed to remain fully usable without dictionaries or shorthand systems.
Future development may explore additional productivity features, experimental input methods, or optional extensions.
The exact direction of future development has not yet been finalized.
ROS-e can be viewed as a simultaneous orthographic input system that occupies a middle ground between conventional typing and traditional stenography.
- Semo-e Keyboard (Korean) — A Korean simultaneous-input keyboard project that inspired ROS-e.