Should we use the emacspeak speech back-end system-wide in Vinux?

[Tomas Cerha]

Hi Halim,

I think the discussion is unnecessarily nervous.  I understand your concerns, 
but if you
try to understand ours, you will see that everyone here tries to make Speech 
Dispatcher
as neutral to all possible system configurations as possible.  If there is also 
a chance
to make these configurations simpler, we are interested as well, but that must 
not
hinder the others.

Dne 9.10.2010 21:16, Halim Sahin napsal(a):
> If you, and others ignore my mails like the one making a suggestion of
> implementing an audioagent which could fix all these issues that's not
> my fault.

You know that many your constructive suggestions are not ignored on this list.  
I know
that this one didn't receive any feedback.  I personally thought about it, I 
also talked
about it with some other people.  I appreciate this suggestion as constructive, 
but
technically I am not convinced that it is a good idea.  I only didn't feel the 
need to
reply as I don't have strong arguments either way.  As I see that nobody else 
replied,
maybe others have similar problems, I don't know.  But believe me, there is no 
internal
policy that would not permit replying to Halim's suggestions :-).

> An easy way to show you the problem is:
> Install a debian system with speakup/speechd-up and orca and try to get
> it working with newest speechd.
> Make sure that the textlogin should be accessible as well (not only gdm
> login).
> If you can do that please tell us if this is a common easy task which
> can be done by endusers as well.

We rely on packagers to do this kind of things and report any problems they 
encounter.
If anyone knows of a particular problem with this setup, or even has a 
suggestion how to
solve it without breaking other setups, we are interested to hear that.  It 
just needs
to be solved step by step pragmatically without flaming about unrelated topics. 
 If you
pollute a good suggestion with a few irrational attacks, it may very well remain
unnoticed despite the fact, that it contains a good idea.

> Plain Text-login can only work if speechd runs as system-service and the
> screenreader starts in with an init-skript as well.
> Please try and show us howto do it if you know an useful way.

Text login is quite a special case and it might deserve special handling, if it 
doesn't
fit into the concept which works for everything else.  Maybe you don't actually 
need an
all featured generic screen reader to make login accessible?  Just my two 
cents...

Best regards,

Tomas