Maintaining speechd-up

[Hynek Hanke]

On 14.7.2010 16:53, Bill Cox wrote:
> Hi.  When I updated the speech-dispatcher package in the Vinux testing
> repository to 0.7

Dear Bill Cox,

unfortunatelly, we no longer maintain speechd-up, as is
stated on the project webpage.

The problem is in Speakup. If something didn't change
since we last checked, it is a kernel patch which does
a lot of things in kernel space which are inapporpriate
and should be done in user space. As such, it was not
accepted into the mainstream kernels. It has no notion
of users, so it suffers security issues and is impossible
to cleanly integrate with user-session technologies, such as
Speech Dispatcher, Pulse Audio, at-spi, DBUS etc.

We have talked to the Speakup developers of the issues that exist
and that will make it technically hard to integrate with other
technologies in accessibility, but there was no will on their side
to fix things. The most often stated reason was that hardware
synthesizers will always stay most important and that for them,
it's better if Speakup is purely in the kernel.

I think it is still possible to setup the current version of speechd-up
with Speech Dispatcher 0.7 using inet sockets and running it
as a system service (which is already being deprecated however).
This however means going through a lot of new hacks and depreacted
solutions, which is frustrating, time-consuming and eventually leads
nowhere. Given that our resources are limited and there is other
very important work to be done, we decided to discontinue the
development and the support of speechd-up.

I know the OpenTTS developers maintain ospeakup, which
I think is very good for some users, and I think it will still
be able to live for some time, but unless Speakup is changed
significantly, it will cease to work with the other technologies
sooner or later.

Of course I understand that this is bad news for many
users and I'm sad the situation came to this point.

Best regards,
Hynek Hanke