- From: Jeff Kusnitz <jk@us.ibm.com>
- Date: Tue, 25 Nov 2003 13:26:48 -0800
- To: "Wolfeld,Jeff" <jwolfeld@cisco.com>
- Cc: www-voice@w3.org
A browser wouldn't have to generate six different grammars, it would have
to generate one grammar with six or so rules (or possibly one rule with a
number of optional parts).
$command = $stargazer | $astrophysics | $news |
( $stargazer $astrophysics ) | ( $stargazer
$news ) | ( $stargazer $astrophysics $news ) |
...
$stargazer = Stargazer;
$astrophysics = Astrophysics;
$news = News;
Something to that effect anyway..
Jeff
> How are voice browser developers expected to implement
ACCEPT="APPROXIMATE"
> in grammars? It appears to be defined for VXML, but there is no such
> provision in SRGS+XML. The only way I can see is to create multiple
SRGS
> grammars, one for each legal combination of words. For example,
"Stargazer
> Astrophysics News" would require six grammars:
>
> Stargazer Astrophysics News
> Stargazer Astrophysics
> Astrophysics News
> Stargazer
> Astrophysics
> News
"Wolfeld,Jeff" <jwolfeld@cisco.com>
Sent by: www-voice-request@w3.org
11/25/2003 12:20 PM
To: www-voice@w3.org
cc:
Subject: ACCEPT attribute for grammars
How are voice browser developers expected to implement
ACCEPT="APPROXIMATE"
in grammars? It appears to be defined for VXML, but there is no such
provision in SRGS+XML. The only way I can see is to create multiple SRGS
grammars, one for each legal combination of words. For example,
"Stargazer
Astrophysics News" would require six grammars:
Stargazer Astrophysics News
Stargazer Astrophysics
Astrophysics News
Stargazer
Astrophysics
News
This can quickly become unwieldy.
Thanks,
Jeff Wolfeld
Received on Tuesday, 25 November 2003 16:28:32 UTC