Re: Request for clarification on I18N speech synthesis question

It would make mor sense to me to have soetihng that does as Masayasu is
suggesting - if it can handle language changes to make them (and only
signpost them if asked) but if it cannot handle the language, then it should
say something like "Korean text - unable to read".

Charles McCN

On Wed, 10 May 2000, Masayasu Ishikawa wrote:

  Jon Gunderson <jongund@ux1.cso.uiuc.edu> wrote:
  
  > I think that the user agent should allow for configuration of the following 
  > options.  But option one in each case should be used if the user agent does 
  > not support configuration.
  > 
  > When the author specified language is supported:
  > 
  > 1. Announce that a language change is taking place and speak the 
  > word/phrase in the supported language
  
  MI:
  Hmm ... if language change doesn't occur frequently, that would make
  sense, but when it occurs frequently, I think announcing language
  changes would be rather annoying than helpful.
  
  Suppose that a user's primary language is Japanese, and a speech user
  agent supports Japanese and English.  These days many English words
  appear in Japanese documents, especially in technical documents.
  For example, when I'm reading a guidebook for HTML, I don't want to
  hear "English Phrase" "HTML" "end English Phrase" each time a word
  "HTML" appears if a UA doesn't support configuration - in that case
  I have no way to stop it!  Anyway, I strongly recommend that UAs
  should allow configuration so that users can choose not to announce
  language changes.
  
  JG:
  > When the author specified language is non supported:
  > 
  > 1. Announce that a language change has taken place and that it is not supported
  > 
  > 2. Allow the user to get the numerical character information of the 
  > unsupported language
  
  MI:
  BTW, I'm wondering what a UA should do when most part of a document
  is written in unsupported language(s).  I tried IBM's Home Page
  Reader 2.5 (English version), and it does change speech engine
  when a document's language is specified via the `lang' attribute
  of the `html' element.  It speaks languages like French and German
  pretty well, in addition to English.
  
  However, when it encountered a Japanese document, which is not
  supported by the English version of HPR, the result was quite
  terrible.  It got all Japanese characters as garbage (i.e.
  misinterpreted them as Latin-1 characters), and just read those
  garbage literally.  Just too bad.
  (Note: the Japanese version of HPR speaks Japanese pretty well, though.)
  
  In this case, is it appropriate that just announcing the document's
  language is not supported and don't read the document?
  
  Regards,
  -- 
  Masayasu Ishikawa / mimasa@w3.org
  W3C - World Wide Web Consortium
  

--
Charles McCathieNevile    mailto:charles@w3.org    phone: +61 (0) 409 134 136
W3C Web Accessibility Initiative                      http://www.w3.org/WAI
Location: I-cubed, 110 Victoria Street, Carlton VIC 3053
Postal: GPO Box 2476V, Melbourne 3001,  Australia

Received on Tuesday, 9 May 2000 14:36:39 UTC