Erik’s wonderful post today jogged my memory about Ribbit and the edge of the network. In Erik’s post, he’s asked to describe the difference between his service, Lypp, and Ribbit, and gives his honest appraisal of the environment in which edge applications live:

The edge of the network is a nasty place. Bandwidth issues, carrier packet shaping, lack of end user control and costly redundancy solutions make it nearly impossible to deliver a predictable and reliable telephony service.

Indeed, Erik’s right on, as you expect, given his experience. He’s a true authority on the subject. Not to disagree, but there are other differences as well that should be mentioned:

• Ribbit’s API is an Adobe Flex API, which at first gave me pause, but upon reflection, I like. Rich Internet Applications (RIAs) are the future of desktop applications, and although there are still many choices for which technology will win (Flex, Ajax, etc), there’s no doubt it’s going to be one of them. Given that, there’s lots of programmers who will know and understand this programming approach – good to see they now have an option. Lypp’s API boasts a beautiful REST interface, which a Ruby-ist like me appreciates, but is truly cross platform and scalable on any language. If you want to program a Ribbit application, start learning Flex. I’m giving this one to Lypp.
• I firmly believe that applications of the future have the network, and in particular the Internet, as the platform. Ribbit’s API is a good step towards that, and enables many web developers to blend voice into their applications. I think Lypp’s API is a better step towards that, as it doesn’t limit your application to a single technology and vendor. Lypp gets this one too.
• I’m going to get slammed for this, but since there are more Web script guys in the world, an easier to learn technology is beneficial in and of itself. Not that the Ribbit developers are less intelligent, but to make Lypp sing, you probably need to be a programmer. (Actually, I think that the Lypp API is so beautiful that you shouldn’t be allowed to use it unless you have the training appreciate it. Sort of like owning a BMW.) Ribbit is more appropriate to the general population of web developers, so they get that one.
• Erik is right about not trusting the voice quality at the edge of the network, but it stretches reason to say that we need to abandon all soft clients. I got one word for you: Skype. There are many, many reasons to have soft clients, and even though I think the computer makes a lousy phone, there are enough reasons to want to scream at your computer to make sure this technology will be required. Simply, Ribbit needs to exist, I just hope they drop that damnable “First Phone Company” slogan.

My real opinion? Most of the Ribbit applications STILL are putting voice onto your desktop, instead of using voice to improve other applications. It’s nice to see a soft client that looks like an iPhone, but it would be beautiful to see a soft client that increases the closing rates of my sales people. APIs are nice and all, but you only need an API to solve a real world problem you can monetize.