W3C home > Mailing lists > Public > www-voice@w3.org > April to June 2005

Re: midnight in ISO 8601

From: Max Froumentin <mf@w3.org>
Date: Mon, 13 Jun 2005 17:59:52 +0100
To: "Arnold Daniels" <info@adaniels.nl>
Cc: <www-voice@w3.org>
Message-ID: <87aclup4zb.fsf@w3.org>

Hi Arnold,

"Arnold Daniels" <info@adaniels.nl> writes:

> Each language has it own format in numbers, dates, time, etc. Taking in
> concern that the text is also displayed on screen, voiceXML should not force
> a certain format. Instead the TTS system should use the current language to
> decide how to pronounce the number, date or time correctly.

Even if SSML's say-as element helps the TTS system pronounce 

<say-as interpret-as="date">4/5</say-as>

as "may fifth" (in an en-US locale), that doesn't mean that say-as
_forces_ formats. A TTS system is always allowed to pronounce the
unmarked string "4/5" as "may fifth" if it finds that the context
indicates that it's a date. Indeed the Note says: 

[[ for "date" strings that a processor would correctly interpret
without a say-as hint, the addition of the appropriate say-as
information should not alter the way in which the date is pronounced. ]]

The format attribute shouldn't be seen as meaning that the content
format inside the say-as element must be in that format. Like the 
interpret-as attribute and the say-as element in general, it's just
a hint to help the speech engine. In my opinion, the SSML spec is
clear on that point.

Max.
(my opinion, not the working group's)
Received on Monday, 13 June 2005 17:00:01 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 30 October 2006 12:49:01 GMT