Thomas Howe

Taken to its essence, what is the difference between voice mashups and voice applications? I joke to others that I feel like Jane Goodall, and I’ve noticed that some of the apes are using different tools than the others. This set of apes are working on a different set of problems than the others, and their tools are different. I’m sure if they could speak, they’d claim to be solving the same problem, maybe even use the exact same words, but I know that they aren’t. These apes are working on an entirely different problem, so they use new tools and, I’m quite sure, need new tools, as the old tools simply wouldn’t do.

What is the difference between voice mashups and voice applications? From a non-technical perspective, the difference is pretty simple. Voice mashups are applications that happen to use voice. Voice applications are applications that are centered around using voice.

A voice mashup will use a telephone so that you can order a pizza, but the real problem that’s being solved is that you are hungry. The phone is there to help get that job done. If you didn’t have the phone, there would be other ways of ordering a pizza, perhaps some better (telepathy?) A voice application is something like unified communications, where the voice is used as the content for voice mails, etc. If you took voice out of a voice application, the essential reason for the application to exist would be gone. A voice application without voice would seem silly. A mashup application without voice would still be an application, just perhaps not as compelling.

If we were to leave it at this, you might conclude that the difference between voice applications and voice mashups is simply with intent. For instance, nothing differentiates the pen and paper used to write a novel versus a non-fiction report. This, however, is a case where using certain tools and approaches to writing voice mashups have significant advantages over other choices. Until very recently, writing voice applications meant that you would take a programmable platform, such as those provided by Excel and Summa4 in the old days, and more modern offerings from the Tier 1 NEPS today, and use their API to accept an inbound call - maybe play some audio - maybe record some audio - and then send it off somewhere else. As an application designer, you were worried about call state - is it ringing, what happened when somebody picked up - etc. The route and path of the voice were important to your application. I would call this a voice application. A mashup application uses voice almost as an afterthought. If you are writing a logistics application for UPS, and you wanted to notify the recipient that a package was about to arrive, you would not be worried about call state. You wouldn’t be thinking about what to do when the phone picked up. You would make a call to a hosted provider that would send a message for you. There would be no reason, as an application developer, to get into the guts of the switch in any fashion. The tool you would use would be a Web Services call; the architecture would be WITA. The UPS application is enhanced by the phone call, but the problem is not voice, but getting a package to the intended recipient.

In essence, this selection of tools and architectures determines the differences between voice mashups and voice applications. Yes, there are simply times when you need an AIN sort of architecture so that you can catch the call right before the far end picks up. But, in the world of light weight applications, and particularly in the way voice can extend and enhance the business process, you really don’t. Far better to make a simple web service call to deliver a message, connect two parties or invite three to a conference call. This is the essential difference between the two, and why they have two different approaches.

Technorati Tags: telco mashups, thomas howe