Plover development doesn’t seem very active. It does work well, but the GUI is old looking and its implemented in python. One feature I would like is automatic input selection based on available devices (COM8 -> COM7 -> keyboard). I’ll only begin implementation after learning stenography. It is assumed that plover will continue to be maintained (they do have a foundation).
High level set of features
- Json dictionary reading
- serial port reading (get keystrokes from steno machine)
- input manipulation (S-P -> Space)
- I don’t care if it has a gui, but it would be a good thing to have to attract users
- Should be entirely configurable declaratively using json or toml
Lower level implementation details
-
Need to unravel gemini protocol, though it shouldn’t be too bad since I have a keyboard that can generate gemini protocol.
- same for txbolt
-
Lookup tree for the strokes needs to be generated from json file. This needs to be performant. SHouldn’t be too hard since the json must be parsed and read into memory anyway.
-
tray icon for windows and macOS. I don’t really care about this for linux though, it can just be a service.
Packages
input handling
inputbot for input sending and listening for keyboard events. The main problem is that it doesn’t have MacOS compatibility though it seems partially implemted. Also, if this thing gets popular, someone with a mac can just update inputbot. I could even maintain my own fork.
Serialport handling
Can use serialport-rs for listening for serial ports. This is completely cross-platform, which is awesome.