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

Re: Braille support for Debian



On Sun, 28 Jan 1996, Bruce Perens wrote:

> > I'm still waiting for a couple of guys in Montreal to update their 
> > drivers ... And the documentation isn't finished ...
> 
> OK. A package that drove only your device and wasn't documented would
> allow me to install it with some confidence that I am placing things
> where you want them, and would allow you to test the boot disk. You may
> then elaborate the package as time permits.

That should be fairly straightforward.  I personally use /dev/ttyS2, but 
I'll make /dev/ttyS0 the default since that would be more general (after 
all, not all machines have a /dev/ttyS2 :-) ).

> > Even so, there is the small problem of the number of possibilities: there 
> > are a minimum of three major manufacturers (Tieman B.V., Netherlands; 
> > TeleSensory Inc., USA; Alva, Japan?), and then there's no telling which 
> > serial port the user will want to use ...
> 
> Hm. How to self-bootstrap a blind user without sighted assistance...
> If they have access to an MS-DOS system with a braille interface, they
> could run a DOS program on the boot floppy that would help them write a
> configuration file which is then read by the installation program.

Does that mean your root floppy uses umsdosfs rather than ext2fs?  If so, 
the easiest way would be to write the correct executable on to the floppy 
(there is a different executable for each make ...).  The ports question 
can be done easily by some sort of config file.  Have I got the 
filesystems bit right, though?

> The installation program I'm currently working on is more graphical (using
> ncurses forms and menus) than previous. For the long term you might want to
> work on a more braille-friendly installation.

Perhaps.  Having said that, as you seem to realise, ncurses is manageable 
- it's svgalib and X that are the main problems. :-)

> How do you like "rsynth"? When it has access to the OED database it's
> pretty good, but its implementation using Klatt's software vocal tract
> makes it too slow to be practical on some computers. A synthesizer using
> pre-recorded phonemes would be able to run in real time.

To be honest, I've never tried it.  It may seem odd, but I don't have a 
sound card.

In general, I think, blind users of synthesized speech tend to go for 
external purpose-built synthesizers; DECTalk is a popular one, but there 
are others which, while the speech is lower quality, can go much faster.  
A friend of mine uses one so fast that I can't understand a word it says! 
:-)  I think he's considering moving to Linux, in which case he may write 
some software for it ...

Nikhil.



Reply to: