W3C home > Mailing lists > Public > www-voice@w3.org > January to March 2008

Inconsistencies in application.lastresult$ for <transfer>

From: Robert Keiller <robert.keiller@voxsurf.com>
Date: Wed, 20 Feb 2008 09:57:09 -0000
To: <www-voice@w3.org>
Message-ID: <DEEGIBIDCHOCDKMBLBMJAEIDDDAA.robert.keiller@voxsurf.com>


On a close reading of the VoiceXML 2.0 spec there seems to be an
inconsistency in the
was application.lastresult$ is filled in for transfer elements.

For blind transfer the spec says:
---------------------------------

"The caller can cancel the transfer attempt before the outgoing call begins
by barging in with a speech or DTMF command that matches an active grammar
during the playback of any queued audio.

...

Also, the application.lastresult$ variable will be filled as described in
Section 5.1.5."

For bridged transfer the spec says:
-----------------------------------

"If the transfer was terminated by speech recognition input, then
application.lastresult$ is assigned as usual."

The implication is that for bridged transfer application.lastresult$ will
only be filled in for
speech recognition, whereas for blind transfer it is filled for both speech
recognition and DTMF.
This seems highly anomalous to me. In particular it means that the semantic
interpretation
of DTMF input, which is not available from the form item's shadow variables,
can only be obtained for
blind transfer.


Robert Keiller
Head of Voice Technology
e: robert.keiller@voxsurf.com

t: + 44 (0) 207 240 3621
f: + 44 (0) 207 379 7573
m: + 44 (0) 774 873 6049

VoxSurf Limited
Registered & principal office:
2nd Floor
Elme House
133 Long Acre
London WC2E 9DT
United Kingdom
Company No (England):  4051448

http://www.voxsurf.com/

Email disclaimer: This can be viewed at
http://www.voxsurf.com/disclaimer.html
Received on Wednesday, 20 February 2008 09:57:25 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 20 February 2008 09:57:29 GMT