W3C home > Mailing lists > Public > www-archive@w3.org > September 2003

Minutes of telecon of Sept 19

From: Jim Larson <jim@larson-tech.com>
Date: Fri, 19 Sep 2003 09:16:29 -0700
Message-ID: <3F6B2BDD.1030403@larson-tech.com>
To: Daniel Burnett <burnett@nuance.com>
Cc: Richard Ishida <ishida@w3.org>, Martin Duerst <duerst@w3.org>, luc.vantichelen@scansoft.com, scott.mcglashan@pipebeach.com, jerry.carter@scansoft.com, paolo.baggia@loquendo.com, jk@us.ibm.com, www-archive@w3.org, w3c-i18n-ig@w3.org

Attendees at Sept 19 2003 telecon

Dan Barnet
Jim Larson
Richard Ishida
Martin Durst
Luc VanTichlin


Richard:  Bidi is a small change with attribute with values 
left-to-right, right-to-left, right-left-override, and 
left-right-override  Only Mongolian currently uses top to down.

Jim:  Martin should write white paper for Philip Hoscka's group 
outlining (1) the problem, (2) current preferred solution.  Hoscka's 
group should review this, and figure out next steps.

Martin:  And add this as an issue for V3.

Jim:  We will add this as an issue for V3.

Martin:  I will write paper and forward to Hoscka's group for a broader 

Dan:  item #58:  We often split text strings across several lines to 
improve readability of the code.  We don't understand your comment that 
the line break will have negative effects on visual rendering. 

Martin:  There are lots of little potential trouble points here.  The 
visual rendering doesn't need to be perfect. 

Richard:  We don't see any big problems.

Dan:  I will officially reject this one

Richard:  We formally withdraw item 58.

Martin:  Issue 1 is of concern

Dan:  we will respond to this later

Dan: item 4:  Both SSML and SRGS are used outside of VoiceXML. 

Dan:  change to "SSML is one of these standards and is designed to ...."

Dan: item 11:  Pronunciation dictionary may be language specific

Martin:  OK

Dan:  We will need e-mail responses to officially acknowledge agreement 
on issues.

Martin: We can all approve the mintues and that will be fine for 
officially agreeing to comments.

Richard: Xml:lang will should change the visual display of the language.

Dan:  We want to be clear that changing xml:lang will change other 
characteristics for voice applications.  We permit xml:lang on <p> and 
<sentence> elements. 

Dan: We can provide a description (if not a transcription) of audio.  
The content of the <audio> element is text that is rendered if the audio 
file is not available.  xdml:lang is not available on <audio> element. 

Martin: <audio> contains content which may use ssml elements.

Dan:  correct

Luc: disc is designed to tell you what the audio is if the audio can't 

Luc:  When you change a language for vocal rendering, that will have an 
effect on various other parameters (including gender, speed, age, pitch, 
etc.) which will be disruptive to the listener.    There may be breaks 
between language shifts.  Thus we discourage frequently use of xml:lang.

Dan:  Think of it as a voice change, not language change.

Martin:  This should be documented. 

Luc:  we have, in section 2.2.1, a description of what happens.  After 
the second example, is additional description. 

Martin:  We have made progress in understanding.  We will want to 
discuss this in the group.  You have made a strong case for an exception 
to using xml:lang everywhere. 

Dan: I will be back week after next.  Let's continue then.

Martin:  Yes.  We have a lot to learn on our side. 

Richard:  we can prepare our thoughts in advance of a telecom.

Dan:  Please let us know which items you want to discuss.

Dan will start process of scheduling another telecom.

Jim will schedule the next meeting.
Received on Friday, 19 September 2003 12:25:15 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:42:31 UTC