Re: [orca-list] Capital, Capital, Capital



Dnia 2008-03-19, Åro o godzinie 11:24 +0100, Peter VÃgner pisze:
[...]
This looks as you haven't applied any patches to orca, right?

Exactly - I only upgraded Orca to last SVN version. This patch really
fixes problem with speaking characters as strings, but there are still
problems with accented letters and capitals announcement.

While typing, I have only pitch change. It is not very big problem,
because I know when I press shift key, but I think it's intended. While
reading characters (in both normal and flat review mode) I have both
pitch change and 'capital' word.

Accented letters are now spoken properly (in Polish) while reading
characters, but still is problem with typing. To insert 'Ã' character I
must press right-alt_semicolon, then 'a'. In gnome-terminal Orca is
silent. In Evolution, gedit and Firefox edit fields Orca speaks 'a'
instead of 'a acute' (or something informative).

Military spelling is still unusable with accented letters. Probably it's
similar bug ('mÃj' is spelled as "mike u juliet"), but IMO it needs more
work to be usable with other languages than English. My proposition:

a) provide mechanism allowing individual letter names for particular
language (for example 'Å' in Polish should be spelled 'Åucja')

b) for other letters send string like _("alpha")+" "+_("acute") for 'Ã'.

And I still think it should be done with synthesizer, not screenreader.
[...]
oh BTW I have got capletrecognition commented out in speech-dispatcher 
config and I had to convert characters to lowercase in 
speechdispatcherfactory.py to force the word "capital" to go away.

...and I prefer the word 'capital' (really not very long in Polish) :)
Also - I'm still playing with Ivona module for speech-dispatcher
(pre-alpha stage yet), and Ivona does not support pitch control.



I think speaking of keys is somewhat limited in the speech-dispatcher 
backend of orca so I am not going to wonder about that now. It's 
supposed to process printable characters only but e.g. double quotes 
(ascii char 34) is not processed and orca processing is taking place.

This is limitation of "key" command - double quote, space and underscore
must not be send as-is, but must be defined 'by name'. There is
synthesizer problem and probably will be fixed in eSpeak (Jonathan
promised to think about it).

Bohdan





[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]