Hi,

When I looked at the notification stuff I noticed that speech should be a 
possible output. And there are some text messages already... :-)

My idea was to view the speech synthesizer as - a synthesizer!
Compare with reading an mp3 encoded file with a text file.
* You have to decode it to waveform
* You have to output it to a multimedia device
* The data is preferably transfered over the interfaces in its most compressed 
format - in this case text
=> the text to speech part could be an arts plugin...

But then it hit me:
* The spoken messages need to be localized!
=> Languages do not use the letters in the same way.
=> You will need a speech engine per language.
=> And sometimes that won't be enough - you will need additional hints.
=> The hints has to be standardized to be able to use an alternative engine
	 for a language.
=> So - how will you encode the sentences?

My suggestion - use the already translated i18n sentences as much as possible.
* Button names, menu names, tool tips, ...
With a way of adding extra hints and/or more text.

It is possible that what you really want is a "qt-speech" - one that only says 
the item with focus.

In addition to this you will need the proposed interfaces - to handle users 
modifiable fields... (like when writing or reading this mail...)

And one more thing: I often read English text with Swedish localization, and 
vice versa...

/RogerL

-- 
Roger Larsson
Skellefteċ
Sweden

 
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<