Re: [Kde-accessibility] Annoucing libbraille


Roger Butenuth wrote:

And don't forget the braille drivers in brltty. I used them as a base for
brass and I am thinking to change brass to use them again. I did the code
fork because at the time I did it they were not loadable at run time,
something I wanted to have. But know they are. With minor changes, they
should be useable as a common base. They support a lot of different
braille devices.
It would be great if all Braille open-source projects could agree on a standard API for drivers of Braille devices. We could discuss the API in details; maybe is the right place for that?

As far as I am concerned I also forked my drivers from brltty for different reasons (it was quite a long time ago, so it may have changed since then - at least I know that the modularity has improved a lot):
- there was not a clear separation with the screen reader part of brltty
- the GPL didn't seem to me the best for drivers, if we want some kind of "standard" usable by as many applications as possible (I use LGPL) - some things were not consistent, like you would call the initialisation function of a driver and it would loop forever waiting for a display to be connected, while another driver would come back after a few tries - the drivers did too many things which should have been done in another higher layer; like some drivers do propose a configuration menu to change the configuration of keys of to show some help, and this is coded inside the driver itself - not very efficient serial programming : all reading functions (when waiting for some keys input for example) are returning immediately so you have to put some delays and loop to read some keys. When there is a need to wait for the terminal to be ready, the driver will wait a fixed amount of time, then read data. There are, in my opinion, more efficient ways: the serial communication system on unix (termios) or other platforms provides some ways to block while waiting for events and you can even configure a time-out.

I was also concerned about portability to win32 but that may not be a concern for all open source people.

So maybe if we could discuss all those issues together, we could define a common API and a _common behaviour_ (I think it is even more important than the API since the API is rather basic anyway) for Braille drivers so that drivers working with one application could work with all the others.


Sébastien Sablé

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]