Archive for February 2005
VBSALT, a Rapid Application Development (RAD) environment from Pronexus (Microsoft Speech Partner), is a Visual Studio .NET Add-in that allows rapid development of Speech applications for the Microsoft Speech Server (MSS), making full use of components provided by the Speech Application SDK (SASDK). The name "VBSALT" may be a bit misleading at times because application code can be generated for both, VB.NET and C# applications. The company probably used "VB" in the name since VBSALT is the .NET version of "VBVoice", a Speech RAD tool for Visual Basic (classic).
VBSALT lets IT Managers, programmers, and Call Center developers define call-flow and swiftly generate SALT-side portion of their Speech apps, thus enabling them to concentrate on the business logic. The steps involved in developing an app using VBSALT are:
- Planning the System
- Creating the Call Flow Diagram and Setting Properties
- Creating Grammars
- Creating and Configuring Greetings
- Testing the System
VBSALT also comes with a couple of pre-recorded greeting prompts. Call-flows can easily be defined using drag-and-drop components, which are subsequently linked together to build a visual flow-chart of call-flow logic. Grammars can be developed either using the Grammar builder provided by the SASDK or using the VBSALT VoiceForm wizard, although I think the former one is much better. Come to think of it, people had been developing Grammars by simply typing in notepad before the SASDK arrived. For now, Life is Good !
KOMaL (Knowledge Object Modeling Language) is something I have been trying to develop for the past 2 years (but my schedule wouldn't let me). Anyway, I finally came with the following small description for KOMaL which I would refine further as I discover new tools and technologies.
The Internet is an almost-limitless resource of Information which exists in the form of web pages, followed by images, pdf files etc. However, web pages provide the only means for hosting information, each one having a URL, and acessible only through a web browser which renders the HTML in the page to display the content. Looking up or searching for information on a particular topic involves knowing the exact URL or by using a search engine where searches can only be made for a keyword.
Currently, the global directory of domain names maintained by InterNIC has no means of indentifying the type of content hosted at a particular domain name. In most cases, a user does not necessarily get to the exact information she requires since a search usually turns up thousands of results and it is impossible to look each one up. Since web pages are composed of textual information contained within markup, the only way the content of that page can be characterized is by making a search on the terms and keywords in it. Usually, the meta tag is used to explicitly add keywords to a web page so they may be picked up by a spider or web-crawler.
With the advent of computers decades ago, the word "data" was adopted and used in connection with the textual and numerical facts and figues used in electronic data processing. With the proliferation of Internet a little more than a decade ago, the word "Information Technology" with emphasis on "Information" became widely used. The focus now has shifted to the word "knowledge" which primarily refers to the Information that is of particular use to an individual or organization. True, vast information is available on the internet; however, the internet failed to become a source of knowledge i.e. the most relevant and meaningful information for an individual or enterprise. In order to allow information to be represented as knowledge, changes in the existing internet architecture or infrastructure is not a possible option. The need is to come up with a new architecture that can run parallel to the existing system, is sharable, and can enable a more thorough representation of knowledge and its search mechanisms.
Finally, I have also found a couple of energetic students at SSUET to undertake this project. Wish 'em all the best.
I mentioned the real-sounding SpeechWorks/ScanSoft TTS in one of my previous blogs. But one TTS I came across with actually took my breath away. It not only gives voice output (as most TTS engines do) but also provides a human character developed in Flash to speak the voice output. Brilliant work, and can prove very effective in teaching applications. Check it out at http://www.sitepal.com