It seems that in some fonts the Braille space (code point 0x2800) is not the same width as the other Braille characters. On this Windows computer it seems that "Segoe UI Symbol", a Microsoft font, is being used by default, and has that problem.
How do braille displays deal with pages containing these unicode characters? Could something along these lines offer a way to make diagrams accessible to blind users? (a few lines at a time, anyway...)
Whether written, spoken, or extruded as Braille, words are generally used for a serialized conveyance of information. Images on the other hand are meaningful because of the spatial relationships of their content.
You're probably better off describing the visuals.
I can’t imagine trying to interpret a diagram on that would make a lot of sense.
Also, keep in mind that any labels or other text in the diagram would also need to be communicated using the Braille symbols.
A physical diagram that you could touch and feel on the other hand, that used Braille for text and lines for the lines and other shapes for the other shapes, on the other hand is something else I think. I have seen such diagrams in reality. But those you can touch and feel freely in any direction, anywhere at the same time.
Right. What does one of those do if it encounters the Unicode braille codepoints?
One could say it just echoes them to the device verbatim. But to me, this seems like a movie that's originally in English, but part way through the movie, someone says something in German, for effect. If you then translate (specifically, dub) the movie to German, how do you translate the German? If you just leave it as-is, you lose the "they switched languages here" information that the original contained.
Not sure what else you could do. You could inject some text that says "Oh hey user, this was originally Braille, even on visual displays", as an escape of sorts.
But for all I know these devices do something like "unrecognized character" and draw the Braille equivalent of boxes.
And if you're into robots, I also wrote a ROS topic visualizer: https://github.com/dheera/rosshow