[orca-list] orca and speech-dispatcher

[Michael Whapples <mwhapples aim com>]

Hello,
In trying to find the cause of a problem I am having with orca, 
speech-dispatcher and ibmtts, I discovered a potential problem with 
orca's use of speech-dispatcher.

The problem is that orca does not set any priority levels and relies on 
the default priority level of speech-dispatcher. You may ask what is the 
problem with this? Well there are possibly two issues:
1. It is possible for the user to change the default priority level in 
speechd.conf which alters how each call to speak is handled and some 
levels can actually lead to output which is not correct from orca (eg. 
the text priority level would mean each call will stop the last speech 
output and this means that where orca sends output in a few calls like 
when there is a line with a link half way across a line in firefox, you 
will only here the last call to speak).
2. This is related to the first, I suspect that the default for when 
priority is not set in speechd.conf is not correct and if 
speech-dispatcher developers feel this way and fix it then this will 
lead to the priority level being text with the undesired impact as above.

Now the question is how to resolve it. I see two ways:
1. When orca creates its speech-dispatcher client it could explicitly 
set its priority level so making it unaffected by changes in 
speech-dispatcher or speechd.conf. This should be very simple to do and 
I could easily create a patch in minutes (I guess). If going down this 
route what level would be suggested? I think from reading 
speech-dispatcher documentation it would probably be message although 
important seems to be the current default orca has been designed around. 
Setting the level to message would allow only output with priority 
important to interrupt orca and would maintain the current behaviour 
that multiple outputs from orca would be queued rather than the later 
ones interrupting the earlier ones. Speech-dispatcher documentation 
indicates important should be used rarely as nothing can interrupt it.
2. Orca could make greater use of the priority levels of 
speech-dispatcher. This might be a better solution as it would mean that 
its output would be considered in importance to other applications 
producing speech output. The difficulty would be, is orca designed in 
such a way that priorities could be set for each call of speak? How much 
work would it be for this to be supported? I imagine it might be quite a 
bit of work.

Michael Whapples