[Kde-accessibility] Names of characters and keys

[Gary Cramblitt]

On Thursday 06 July 2006 14:21, Jonathan Duddington wrote:
> A text-to-speech facility needs to include the ability to speak the
> names of individual characters and also the names of keys on a computer
> keyboard.
>
> These are specified with the
>   interpret-as="tts:char"
>   interpret-as="tts:key"
> attributes in http://www.freebsoft.org/doc/tts-api/tts-api.html
>
> 1.  Does a list of these names already exist on the computer, perhaps
> as part of internationalization data, or does every speech synthesizer
> need to produce a list for every language which it supports?
>
> 2.  Should characters be announced with their proper unicode name, such
> as:
>  "LATIN SMALL LETTER E WITH CIRCUMFLEX"
>  "RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK"
>
> or would more abbreviated names be expected?

For brevity, I'd make it "lowercase e circumflex" and "end quote".  I guess it 
depends upon the intended usage.  If one is talking in a mixed language 
environment, or talking the interface of a word processor, the difference 
between English quote characters and German quote characters would matter, 
but when speaking text of a single language, just "begin quote" and "end 
quote" are all the user wants to hear.

> 3.  When spelling a word (from  interpret-as="glyphs" ) what should the
> TTS do if it finds a letter which is not part of the language's normal
> alphabet?  For example in English, "cafe" with an e-acute, or
> "angstrom" with "Latin letter A with ring above", or a German letter
> "sharp S" for "ss"?

I think you should interpret the silence to your email as "well that's an 
issue that will need to be addressed".

BTW, I don't think this is an issue just for the TTS API.  I believe the W3C 
is struggling with some of these issues too for SSML.

-- 
Gary Cramblitt (aka PhantomsDad)