Comment 1 by briang1 on 2011-10-23 19:28
Is this a shortcoming of particular synths, or of the system used for language selection. I saw the interchange over this. if the synth cannot be changed to work, then I think firstly, some people need to look at this not just for Arabic, but maybe other languages where this might occur on certain synths.
Perhaps two interrelated options. IE a selection to override to a particular language, and a toggle to allow it to be switched in or out of action might be the way to go here.

Comment 2 by jteh on 2011-10-23 22:25
The problem is that there is no difference between most English and Arabic punctuation, so NVDA doesn't know which language to speak it in. Therefore, it uses the language reported by the synthesiser.

If we implement this option, it will always use NVDA's user interface language to pronounce symbols. If you have NVDA set to use English, it will alway speak symbols in English. If you have NVDA set to use Arabic, it will always speak symbols in Arabic. I suspect this is not what you want.

How did you configure NVDA with version 2011.2 and earlier? What did you set as your user interface language (in General Settings)? What symbol level did you use?

Comment 3 by jteh on 2011-10-23 22:27
Btw, if the document you are reading correctly reports its language (currently only works on web pages and Adobe Reader documents), you will get symbols in the correct language. NVDA cannot detect the text language any other way at present.

Comment 4 by fatma.mehanna (in reply to comment 2) on 2011-10-24 10:23
Replying to jteh:

The problem is that there is no difference between most English and Arabic punctuation, so NVDA doesn't know which language to speak it in. Therefore, it uses the language reported by the synthesiser.

If we implement this option, it will always use NVDA's user interface language to pronounce symbols. If you have NVDA set to use English, it will alway speak symbols in English. If you have NVDA set to use Arabic, it will always speak symbols in Arabic. I suspect this is not what you want.

How did you configure NVDA with version 2011.2 and earlier?

as for 2011.2, the user who wants the english interface was using it and the same for the user who wants the arabic interface. so the punctuations were being reported according to the interface as you mentioned previously.
as for the earlier versions, the matter was being set to the synth to report the punctuation, so it was ok as the synth was detecting the language.
What did you set as your user interface language (in General Settings)?
englishand arabic for the others who don't know english.
What symbol level did you use?
all and in some cases most.
if this option won't achieve what we want, so i suggest to have 2 options to switch between the old system (the synth) and the new system (nvda).
thanks.

Comment 5 by jteh on 2011-10-24 11:12
Allowing the synth to handle punctuation was discussed and rejected in #1597. Read that ticket for more details, but in short, this will cause problems when reading by character (including typing), as the synth probably won't read some symbols at all.

Comment 6 by mdcurran on 2011-10-26 06:37
8a35966 forces NVDA to always use the interface language instead of the synth language for symbols if autoLanguageSwitching is off. Does this sufficiently fix the issue?

Comment 7 by fatma.mehanna on 2011-10-26 14:35
yes this is what we want.
thanks.
Changes:
State: closed