PublicVoiceXML- European radio stations evaluating
VoiceXML
PublicVoiceXML (www.publicVoiceXML.org) is both, the name
of a VoiceXML 2.0 compliant browser and the name of a trial project with
European community radio stations, who are evaluating VoiceXML technology for
their day-to-day business. VoiceXML is a browser for users, who call an
interactive service via a (mobile) phone. It follows the W3C VoiceXML
specifications, which are supported by the VoiceXML forum, an organisation,
which has been set up by IEEE. Standardisation efforts have been now taken over
by W3C in the Voice Browser Working group.
The project consortium consists of:
� PUBLIC VOICE Lab as the project
initiator and co-ordinator together with
TEAM TEICHENBERG, both located in Vienna
� a software team in Chandigarh,
India
� MTA SZTAKI in Budapest, also
hosting the W3C office for Eastern Europe
� Vienna's community radio station
Orange 94.0, who has contributed to the user requirements phase and has been
integrating PublicVoiceXML into their content management and remote journalists
management system.
PublicVoiceXML is designed to be used by SMEs, and it works nicely
on low cost telephony hardware. The protocol handling is making use of
ISDN-CAPI, which works fine e.g. with the ISDN B1 active card from AVM. Analogue
lines are currently not supported, as too many country/telco specific cases for
DDI or disconnect handling would need to be taken into account. The VoiceXML
browser is written in C++ and comes with a lot of comment lines (explaining each
function) and nice logging features.PublicVoiceXML supports all "must-have" tags
of the VoiceXML 2.0 specs (speech recognition is not yet implemented). A
pre-release running on Windows platforms has been published on SourceForge in
Dec. 2002 and a Linux version was made available recently (https://sourceforge.net/projects/publicvoicexml/).
The consortium has performed the W3C
implementation report test (http://www.w3.org/Voice/2003/ir/voicexml20-ir.html) and is using the results for further improvements. Most of the trial
applications were built on a basic framework developed within the project. This
framework uses PHP as a middleware, the Postgres database (through the PEAR
database abstraction layer), and the Smarty template engine for the generation
of HTML and VoiceXML pages. This results in clean PHP code, and the separation
of presentation from business logic. HTML templates can be edited by designers
(not necessarily having programming skills), and VoiceXML scripts can be
fine-tuned without touching the PHP code.
Trials at community radio stations have been carried out and are
well documented on the project web site. Embedding basic speech recognition is
already scheduled and there is provision for telco-grade HW to be supported in
Software & Service Technologies � Issue no.
4, October 2003, page 4 the near future.
Exploitation of the project results will be effected through providing support
for customisation, integration and double licensing to telecommunication
companies.