I would like to allow installation of Slint (and Slackware itself and other derivatives if wished by their maintainers) by blind people using a software synthesizer.

Rationale: not all blind folks can afford a hardware synthesizer or a Braille terminal aka a screen reader, and I would like that all be able to install Slint without the help of a sighted person, as they already can if they own a Braille terminal.

So far, I have built the needed packages to have this working on an installed system on startup (the machine speaks just before log in), I will make a request for testing this stuff on Slackware later.

Now begins the hard work, which is to ship in the initramfs all that's needed to have the machine speak during installation.

Among this stuff are the sound kernel modules.

But in the Slackware installer there is no sound module (to save some space, I assume).

So my question is: how do I go about adding these modules?

Is it just a matter of adding the directory /lib/modules/sound and its content?

In addition, should I modify something so that the regular files in /lib/modules/<kernel version> "know about" the added modules; and how do I do that?

Or should I edit the config file, and rebuild the modules?

Sorry for this newbie question, I haven't built a kernel since years.

I ask in this forum as maybe there is something specific to building /lib/modules in the installer, that only a Slackware insider would now.

maybe I'm missing something obvious but have you had a look at the speakup system?

http://www.linux-speakup.org/
http://www.linux-speakup.org/distros.html
http://www.linux-speakup.org/spkguide.txt

Indeed I had, Matteo

Actually I have packaged speakup and espaekup:
http://slackware.uk/slint/x86_64/sli..._64-1slint.txz
http://slackware.uk/slint/x86_64/sli..._64-3slint.txz
Sources:
http://slackware.uk/slint/x86_64/sli...source/espeak/
http://slackware.uk/slint/x86_64/sli...urce/espeakup/
There is no dependencies beyond a full Slackware installation: as I have set pulseaudio as backend, portaudio is not needed (it's available in the same repo, though).

So, just install the two packages in Slackware64-14.2, chmod 755 /etc/rc.d/espeakup, reboot at run level 3 and the computer will begin to speak what's on the screen and spell what you type when you are asked to log in. Adapt the language, rate (speech speed), pitch and volume editing /etc/espeakup.conf as you see fit.

You just need a sound card, I think that all recent machines have one.

I will add a README with the keyboard shortcuts for the screen review and some guidance on usage, hopefully later to day, then post a request for testing. There is already a documentation available, of course, but I want to help the users who don't read long and technical documents.

But my real aim is put all the stuff needed in the installer, to allow blind people to install without the help of a sighted person, even if they don't own a specific hardware.

Fortunately I am helped by Samuel Thibault, aka the Linux Accessibility Guru.

I will also credit Aiyumi Moriya (see her blog) whose posts and SlackBuilds got me started.

I will also provide Orca with her help. That shouldn't be very difficult, but as we say here "à chaque jour suffit sa peine".

You could look at how I create a modified initrd for liveslak, with a regular initrd to start with, lines 2067 to 2143: http://bear.alienbase.nl/cgit/livesl...=1.1.8.2#n2067

Basically: extract the installer's initrd, add the modules you think you need, use my script as an example on how to find the dependencies for these modules as well as any firmware files these drivers need. Then run depmod.
FInally, add the speech binaries and any libraries they depend on, and wrap the initrd-tree into a new installer initrd image.

1 members found this post helpful.

Thanks Eric this will save me a lot of time!

Meanwhile, I have posted a request for testing.