Plojo

Note: uses enigo for computer control. Linux users may need to install libxdo-dev.

Immediate todos

• check drop-in compatibility (dictionary format changer)
• write documentation
• for corrections (browser URL bar), add space and then delete
• add orthography (-Z) for numbers? (to add "00" to a number)
• call function (like for date)
• mouse control

Plover compatible

Text

• add support for plover modes (caps, camel, snake, etc.)?
• add orthography (-Z) for numbers? (to add "00" to a number)
• suffix strokes on multistroke words

Other

• support for escaping characters (\\{, \\})
• for stroke lookup, search also with first letter capitalized/lowercased
• BUG?: "okay, okay" has 2 spaces
• BUG?: capitalize last word for: (| indicates cursor)
  • how should these be handled?
  • said, 'hi
  • get frank's
• BUG: 1BGS works but 5BGS doesn't

Features

• record and replay macros

• a stroke that can add something after the next stroke

• allow comments to be added to the dictionary

• add support for multiple dictionaries that can have their order changed

• something that can suggest briefs based on usage

  • calculate stroke speed and average strokes per word
  • find which strokes happen quickly one after the other (for brief suggestion)

• translation mode for verbatim strokes for brief creation

• all lowercase text filter for texting

• calculate translation frequency

  • also calculate stroke speed and strokes per day
  • calculate strokes per word
  • calculate frequency of certain individual keys and two keys

Test on Plover

• SEUFRPL + ER/ROR
• EPB/TKOUD be "endowed" or "endo youed"
• STREUP/-D be "stripped" or "striped"
• UPB/TPA*EUGS be "unification" or "uniification"
• TPAUR/OUS be "forous" or "forrous"
• (space after): H-L/TK-LS/* be "hello " or ""
• R-RS: be "s" or just "s" (suffix folding for commands?)

Later todos

Optimization

• I probably shouldn't worry about performance because it is already really fast
• use a bloom filter to prevent need to lookup a long stroke
  • instead of looking up 10, 9, 8, ... 1 strokes joined together
  • 10..n (where n is around 4) could be looked up in a bloom filter
    • could be done in parallel as well
• store prev_strokes in a VecDeque instead of a Vec
  • only diff the last 15 or something strokes instead of all the strokes
• find out what text was deleted to allow for delete by word optimization
• limit number of strokes sent to translate_strokes
• possibly optimize hashmap lookup by turning steno keys into a u32
• initialize vecs and hashmaps with capacity to improve performance
• alternate log format that stores time deltas and strokes as ints for low size

Cleanup

• look for plojo config folder in multiple places (instead of just ~/.plojo)
• write dictionary parsing as a serde deserializer
• check for stroke validity with a regex and warn if a stoke is invalid
• refactor machine to use more traits
• use macros for raw stokes parsing
• implement feature flag for serde deserializing in plojo_core
• consolidate Lit and Attached (and maybe even Glued)

Plover compatible

• ignore dictionary unknown special actions
• add orthography rules aliases
• add config to customize undo strokes
• some strokes (like O-RBGS) have a dash when it doesn't need a dash
• should be usable as a drop-in replacement for Plover

Documentation

• write somewhere about how commands are dispatched without modifying any text
  • even if a correction is required, it will not press any backspaces
  • command will only be dispatched if it has been newly added
• document the keys available for pressing and how raw key codes are allowed
• grep for all the NOTEs and document them
• note that numbers in a stroke must have a dash where necessary
  • if 0/5 not present and there are digits 6-9
• note that translations with only numbers will be interpreted as glued
• document how undo removes all strokes that only have text actions and commands
  • also removes text (attached, glued) that is empty
• keyboard shortcuts must use the "raw" version (eg: [/] instead of {/})
• capitalize prev will capitalize the previous word that appears on screen
  • for translations with multiple words, the last word will be capitalized
  • if space prev is suppressed, the whole thing will be capitalized
  • for something like ©ab, the a will be capitalized: ©Ab
• suppress_space_before is the same as a {^} before command in Plover
• retrospective add space will add a space in the stroke buffer
  • this means that undo will "undo" the space stroke that was added
  • retrospective add space itself cannot be undone either
• retrospective add space stroke is set in the config (not with {*?})
• document the overall architecture of plojo
• prefix + suffix doesn't apply orthography (this may differ from Plover)
  • ex: TPAUR/OUS makes forous, not forrous
• force same case overrides force capitalize (when it applies to the next word)
• disable with stroke and re-enable keyboard shortcut (uses rdev keycode format)