W3C home > Mailing lists > Public > www-voice@w3.org > April to June 2006

Comment on the SRGS spec

From: Shane Smith <safarishane@gmail.com>
Date: Tue, 25 Apr 2006 17:02:36 -0500
Message-ID: <8fc15e140604251502r24467f2di59ab52d79e635c77@mail.gmail.com>
To: www-voice@w3.org

Folks,

The SRGS spec currently allows the GARBAGE rule to be implemented as NULL:

----- clip -----
GARBAGE
Defines a rule that may match any speech up until the next rule match,
the next token or until the end of spoken input. A grammar processor
must accept grammars that contain special references to GARBAGE. The
behavior GARBAGE rule is implementation-specific. A user agent should
be capable of matching arbitrary spoken input up to the next token but
may treat GARBAGE as equivalent to NULL (match no spoken input).
------------------

It seems that many vendors are taking this allowed shortcut of
implementing garbage as if it were null, without the ability to match
the arbitrary speech.  Is this really the intended behavior?  If not,
the statement needs fine tuning then so it's not possible to
misconstrue it's meaning. Otherwise, leaving this alone basically
means that srgs grammars that "fully comply" to the spec are not
necessarily portable between vendors that "fully comply" to the spec.

Any grammar I've written that makes use of GARBAGE, will have to be
rewritten for vendor implementations that just treat GARBAGE as if it
were null.

What I personally think is a better way of saying this is:

A user agent MUST be capable of matching arbitrary spoken input up to
the next token, including no spoken input at all.   GARBAGE is treated
as equivalent to NULL when no spoken input occurs.

I'm very forgivable when a vendor doesn't fully comply to the spec, as
long as they are up front about it, and are working on it.  But right
now, a vendor can simply say they do comply, and were not going to
ever implement garbage.  (I'm not naming names, but it ryhmes with
fooance)

Any thoughts?

Thanks,
Shane Smith
Received on Tuesday, 25 April 2006 22:03:46 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 30 October 2006 12:49:02 GMT