C.a.T.S. has successfully presented VoiceXML training to nine delegates - two from a GSM network abroad. The training was hosted at the new C.a.T.S. offices.
Spanning five days of intensive training, the course comprised two days of theory, which included VXML, CCML, SSML and SRGS-XML. The remaining three days were constituted by workshop-like practical sessions, covering topics like DTMF key input, speech synthesis, voice recording, host interfacing, server side scripting, variable handling and sub-dialogues.
The delegates attending the training hailed from various backgrounds, which constituted a good representation in terms of the voice fraternity. MD Johan Grobler stated: "As a company specialising in voice in its myriad applications and platforms, we see these training events not as a threat, but rather as an opportunity to grow the open standards community and get as many people as possible involved. I believe the fundamentals underpinning a language like VoiceXML and its associated languages allows for the creation of a user group that is truly effective to all parties involved."
Apart from the VoiceXML training, attention was given to call control in terms of CCML, showing the attendees how CCML can be used to build functionalities like call conferences, call transfers and screening of calls. Grobler continued: "When one looks at the structure of CCML, it is all about programming and execution efficiency and, in a sense, moving call control from the switching domain to the programmer."
The basic IVR architecture used for training was the Tecnomen test platform with Vodacom in Midrand for PSTN calls, the C.a.T.S. lab IVR for PSTN and VOIP/SIP calls. Also roped in were the C.a.T.S. speech server with voice recognition and speech synthesis capabilities, and a primary and secondary application server.
The rationale behind the construction of the architecture was so delegates could grasp and work with a distributed multi-platform environment, as would most probably be the case in a production environment. Grobler also noted: "The architecture used for the training goes a long way in illustrating how distributed an IVR configuration can be, integrating the Tecnomen platform in Midrand with our speech and application servers in Pretoria/Tshwane."
Grobler believes the true challenges, however, will be in terms of managing the applications and also the application server and speech servers employed. "The openness and simplicity of VoiceXML brings along possible threats with the obvious advantages.
"On an enterprise level, careful consideration will have to be given to the development, maintenance, deployment and management of applications. This is a field C.a.T.S. has concentrated on and we believe we have a refined process in terms of the IDE (Integrated Development Environment), the uploading and activation of applications." He continued to caution enterprises not to rush into VoiceXML orientated architecture for voice if, the process has not been thought through and documented.
Half a day was spent on the design philosophy of voice applications and also on test procedures and strategies. Enough information and building blocks were passed on to each delegate, empowering them to go back to their respective organisations and implement workflows and procedures in terms of application development, touching on every aspect it encompasses.
In preparation for the training, Grobler added: "I requested that specific attention should be given to the prototype phase of development, as many applications being launched by the industry today are not up to standard, and primarily due to the prototype phase not being thorough, and testing not as rigorous as needed."
Thus, the training emphasised this phase of application development, and focus was placed on how the prototype phase can be expedited and facilitated, making use of speech synthesis. Developers can build a prototype IVR application within a very short period of time, which is natural sounding, making use of speech synthesis mark-up language. Within a brainstorm session, the application can be changed and adapted according to suggestions on the fly, and demonstration can take place via a developer's IDE.
Lastly, the whole scenario of IVR farms and voice portals being implemented in line with Web portals were visited in detail. Grobler concluded: "We believe this training will go a long way in demystifying voice, and IVR in particular, and empower developers in general to add voice development to their skills complement. The response was overwhelmingly positive, and we are looking at presenting the training on a regular basis to accommodate the requests."
The training was conducted by two members of the C.a.T.S. team, with Grobler concluding the course with a final word on the past, present and future of voice, especially within the South African context.
Share
Editorial contacts