AFAIK, there isn't any braille screen input support for Linux, so I've been
considering trying to implement some. To be clear, I'm thinking about a
program that detects and characterizes touch events and then recognizes
gestures, and finally reports on these as Unicode sequences.
Approach
My approach is based on the Elixir programming language. For the curious, I'm
planning to construct a directed acyclic graph (DAG) of lightweight processes,
flowing from the touch screen to a back end application which would handle
braille input and feed it to various Linux CLI tools. The graph would be
created and used by means of an Elixir pipeline:
start_reporter()
get_action() |> add_gesture()
start_reporter() starts up a "reporter" process which listens for touch and
gesture events, boils them down a bit, and reports the resulting Unicode to the
client application. add_gesture() is the front end for a set of gesture
recognition processes. It starts these up as needed, broadcasts messages about
actions to them, then mostly gets out of the way.
Issues
One major issue is that I'm not at all clear on how to deal with the output.
What back-end programs should I target and what interface(s) do they generally
want to deal with from input devices? For example, is there an easy way in
Linux for a user mode process to emulate a keyboard device?
More generally, I'd like to get some feedback on a11y, system and user
interfaces, etc. For example:
- What back-end programs should I target?
- What kinds of gestures would folks want?
- What sorts of interfaces should I present?
Advice, caveats, clues, and pointers would all be welcome...
-r
