Framework for describing characters

[allan]

Read phonetic representation and ASCII value is quite important and should add into NVDA future release.

[jteh]

The problem with reading phonetic representation is that it presents localisation issues. While we can certainly allow for phonetic reading of the English alphabet, providing similar functionality for all languages is extremely difficult. Furthermore, it is dependent on the synth language in use, not necessarily the NVDA language. We have very similar problems with the current punctuation support.

[aleksey_s]

at least reading ascii representation of the character was done by me and you can press numpad2 three times to hear this one as well as hex.

[bernd Dorer]

Alexey_s wrote:

at least reading ascii representation of the character was done by me and you can >press numpad2
three times to hear this one as well as hex.

You mean:
You can press numpad2 twice to hear this one as well as hex.

I have tryed it on two systems and you don't have to press numpad2 three times.
I think this fixes this ticket.

[jteh]

No, this ticket is only partially implemented. The reporter requested both reading of phonetic representation (alpha, bravo, etc.) as well as ASCII value. The latter has been implemented, but the former has not. (These should probably have been filed as separate tickets, but no matter.)

[Bernd]

I think for now, it would be better if we could implement it for each language. If we find a method to bind it to the synthesizers voice we can do this then.

[jteh]

We've determined that this will be done separately from symbol/punctuation handling, though probably in the same module (characterProcessing).

[mdcurran]

Work is underway in branch:  http://bzr.nvaccess.org/nvda/characterDescriptions/
In this branch it is now possible to get the description (examples) for a character by using particular review cursor commands:
review current character (numpad2) pressed twice will speak the description for the current character.
Review current word (numpad5) and review current line (numpad8) pressed 3 times, will spell the word/line using these descriptions.
A locale should provide a characterDescriptions.dic file inside their locale directory. English and traditional Chinese (Taiwan) have already been provided.
The english one uses the standard English phonetic alphabet (alfa, bravo, charley etc). The traditional Chinese one provides examples from the libchewingdata project:  http://libchewingdata.openfoundry.org/

This means that for Traditional Chinese, it is now possible for an NVDA user to identify the meaning of a given traditional chinese character.

At this point in time, the characterDescriptions.dic that is used is based on NVDA's configured locale. It would be nice to use the synth's reported voice language, but it seems that a lot of synths incorrectly report this.

[m11chen]

Hi,

I am trying to call the character description function with the laptop keyboard layout, but pressing NVDA+, twice does not speak anything. Does the gesture map need to be updated to reflect the recent implementation?

Thanks

[m11chen]

Replying to jteh:

The problem with reading phonetic representation is that it presents localisation issues. While we can certainly allow for phonetic reading of the English alphabet, providing similar functionality for all languages is extremely difficult. Furthermore, it is dependent on the synth language in use, not necessarily the NVDA language. We have very similar problems with the current punctuation support.

Hi,

Sorry, didn't mean to change the type to task.

[mdcurran]

At this point in time your NVDA locale must be either en or zh_TW.
Is it one of these?
What characters were you trying this on?
Do you get any errors, debugwarnings in your log?
In Python console, what output do you get for:
import characterProcessing; characterProcessing.getCharacterDescription('en','d')

You should get the word 'delta'.

Do characterDescriptions.dic files appear in your locale/en and locale/zh_TW directories?

[m11chen]

Oops, I was testing by running the distribution build from the source. Everything works fine when running the original source tree.

[jteh]

Merged in changeset:main,4150 with some subsequent fixes.