John Hibel of Vocalocity
Q. When will you update OpenVXI to support: a. VoiceXML 2.1
b. Final W3C recommendation for SRGS (both XML and ABNF?) c. Final W3C recommendation for SSML?
d. Latest draft of W3C Semantic Interpretation Language?
e. Latest draft of W3C CCXML?
f. Latest draft of W3C EMMA?
g. Extension for N-gram grammars?
h. An object for speaker verification/authentication? A. That's quite a list of improvements and certainly a lot of development work, but I think you've hit on precisely why we think Vocalocity has such a great opportunity to add value to OpenVXI. I want to make it clear that Vocalocity is going to invest in OpenVXI and the OpenVXI community. This is absolutely core to our business strategy and one of our most important initiatives. Our first priority is to upgrade OpenVXI to full compliance with the latest standards and to ensure that it remains fully compliant and current as the relevant standards evolve. This clearly puts VoiceXML 2.1, SRGS, and SSML high on the list. We also see considerable demand for a CCXML component that's compatible with OpenVXI. We will be discussing more specific product plans and the roadmap with OpenVXI customers in the coming weeks. Q. Will you provide MRCP to enable OpenVXI to access speech recognition, speech synthesis, DTMF recognition, and speaker verification/authentication engines from multiple technology providers? A. Yes we will. Vocalocity provides software components and tools that enable standards. We view MRCP as a component that can be combined with OpenVXI to provide access to various media technologies. These and other components allow platform vendors to quickly adopt and maintain standards compliance. Q. IBM, Motorola, and Opera have introduced XHTML plus Voice Modules (often simply referred to as X+V); will OpenVXI be modularized so it can also be used to develop multimodal applications? A. As platform vendors consider how they will support multimodal applications, we do expect a market to emerge for an X+V component that is compatible with OpenVXI. Vocalocity's mission is to provide a full suite of components and tools that enable standards for our platform customers. Q. Why did you select OpenVXI rather than OpenSALT? Will you do a similar deal with OpenSALT? A. It's all about demand and market presence. Many of the world's leading telecom platforms today have VoiceXML offerings based on OpenVXI. A large contingent of smaller but highly innovative platform companies have already adopted OpenVXI. We believe we have a lot to offer this community, not only through our plans for OpenVXI but also through our portfolio of additional OEM software components and tools, which currently includes a SALT offering. Q. Will you create an OpenVXI News group, so developers can post questions and answers to OpenVXI technical questions and problems? Will you create a site like "sourceforge.net" for posting of audio files, grammars, reusable dialogs, etc. from contributors for use with OpenVXI?
A. Vocalocity OpenVXI will remain on sourceforge.net, and we're currently looking at a number of ways we can energize and enhance the OpenVXI project and community. In addition, there are existing mailing lists for OpenVXI. Visit the OpenVXI project on sourceforge.net for more information on how to sign up for those mailing lists. Q. Does this action have the blessing of the VoiceXML Forum, the industry consortium for promoting VoiceXML?
A. Vocalocity is a VoiceXML Forum Sponsor and board member, and in that capacity we are actively engaged in promoting VoiceXML. OpenVXI is broadly adopted among major telephony platform providers worldwide and has been instrumental in driving adoption of VoiceXML since SpeechWorks introduced it several years ago. Our commitment to invest in OpenVXI will not only energize the OpenVXI community, but also will help promote VoiceXML in general by ensuring that many major platform vendors will have rapid access to the latest versions of the standard. Q. Will you make some VoiceXML vendors unhappy because OpenVXI will undercut their pricing model? A. As an OEM software component that enables telephony platforms to support the VoiceXML standard, Vocalocity's OpenVXI doesn't compete with VoiceXML platform vendors who sell primarily to enterprise and telco customers. Most adopters of OpenVXI already had built their own platforms, and as such did not represent a viable market for vendors who offer complete VoiceXML platforms. Any VoiceXML platform vendor who may be concerned that OpenVXI will undercut their pricing model has nothing to worry about because OpenVXI is appropriate for a completely different market. Q. Will implementers extend OpenVXI to support non-standard features? Do you think this will destroy the advantage of an implementation with no non-standard features? A. Application developers benefit from broad adoption of a standard like VoiceXML because the potential market for their applications is much larger when their applications can run on many platforms. End users benefit as well because this can lower the cost of applications and increase innovation among application developers who can pursue larger markets than if they were tied to specific platforms. Proprietary extensions to VoiceXML undermine these benefits when they lock applications to a single platform. While implementers can extend OpenVXI to support proprietary extensions, any benefit they might gain through differentiation from their competitors may very well be far outweighed by a disincentive for application developers to utilize those extensions because they would lock their applications to one platform and constrain their market opportunity. We see this effect in practice today when application developers specifically avoid proprietary extensions available on some VoiceXML platforms because they want to maintain application portability for the benefit of their customers as well as themselves. Q. Is this the kickoff of the "new speech wars" in which platform vendors will no longer compete by adding extensions to VoiceXML, but will compete at the tool kit level? A. Why have a standard if every implementation treats applications differently? In just a few short years we've seen an astounding adoption rate for VoiceXML, which speaks volumes about the market's desire for standards-based applications and platforms that avoid proprietary lock-in. VoiceXML does not define a new category of platform, but is instead a standard feature that has come to be expected in every telephony platform. A telephony platform is a highly complex product offering and there are many axes on which to differentiate and compete. I think the winners of the "new speech wars" will be those platform vendors who can support and evolve with new standards the fastest and with the least impact to internal resources so that development efforts can be concentrated on the features and capabilities that result in true differentiation. Q. Will OpenVXI work under Microsoft's .NET? A. Absolutely. There are many deployments today that generate VoiceXML from a .NET framework using OpenVXI as well as other interpreters. One of the benefits of having open standards is that customers have choices at every layer of the solution. Q. Will OpenVXI compete with SALT for telephony application under .NET? A. Some vendors are positioning the SALT specification to compete with the VoiceXML standard. This is a problem for platform vendors who must decide which standards to support and when. Whether it's VoiceXML, SALT, X+V, or whatever comes next, we believe that by dramatically lowering the cost, time, and risk associated with adopting standards Vocalocity can eliminate this problem for platform vendors. They can support standards more quickly and application developers can gain experience, provide feedback, and form intelligent preferences sooner. This accelerated cycle of innovation will create more value for customers and benefit the market as a whole. Q. How will OpenVXI affect the unbundling of the VoiceXML processor from media technologies (ASR, TTS, Speaker authentication, etc.) and the licensing fees for media technologies? A. VoiceXML is a user interaction standard that enables the unbundling of application and platform. Now we're seeing a burgeoning market for packaged applications in which application developers compete across platforms solely on the value and merits of their applications. This drives cost down and innovation up, and we fully expect tremendous growth for applications as a result. OpenVXI is a software component that enables platforms to support the VoiceXML standard. Vocalocity sees demand for similar components to enable other emerging standards like CCXML and MRCP that in turn enable unbundling of platform, switching, and media technologies. Proprietary lock-ins that fostered market inefficiency will melt away. While competition will increase, a more efficient standards-based solution stack and robust application growth will drive substantially increased demand for media technologies.