I wonder if Sir Tim Berners-Lee envisioned the concept of people communicating on the World Wide Web when he invented it in 1992 whilst at Cern. At the time of course, the browser, linked to the internet, was based largely on an academic requirement to share information and text-based content.
Given the communication revolution we’ve experienced since then the concept of real-time media over the web is more obvious than it was 15 years ago. But there are a number of substantial problems to overcome to realise this vision.
The browser fundamentally was not designed to cater to real-time audio between multiple users. Technologies have emerged which enable us to contemplate this alternative use – specifically Flash and JAVA. Blabbelon uses both, and here’s why.
Flash makes for a wonderful interface. It gives you a lot of flexibility over graphical and interaction design. But the voice component of Flash gave us some dilemmas:-
- It doesn’t let you use your own audio codecs (a codec is what transforms your voice into bits and bytes that can be sent over the internet)
- It doesn’t let you intercept the keyboard for interrupts which enable a “hotkey” push to talk mechanism (in Blabbelon’s own research, a feature critical to gamers)
- Distributing your audio to multiple users requires using Adobe’s own media server engine, which limits possibilities to those that Adobe decides you need.
We didn’t like those restrictions, and gamers and potential users that we spoke to didn’t either. So we examined JAVA. JAVA allows us to do the “hotkey” intercept that we need, utilise our own audio codec and our own server to do the audio distribution. But JAVA also gave us a dilemma:-
- It isn’t as flexible when it comes to designing User-Interfaces
So we decided that we only had one course of action – put both into an atomic accelerator (also based in Cern, which we think is at least of historic value given it’s where the web itself was invented) and smash the two together at high speed.
OK, so I’m exaggerating about the atomic accelerator part, we didn’t need one in the event, but we did smash the two together at high speed and the resulting output was Blabbelon.
Blabbelon represents a convergence of these two key web technologies. It gives us the best of both Worlds – a UI that our users tell us they want as well as the functionality that they require.
In using JAVA for the key voice components (and a custom patent-pending technology on our servers), Blabbelon has been able to design and implement a very high scaling architecture that is not dependent on any third parties or particular voice codecs. Blabbelon can run anything; we actually started out in ALPHA using GSM, the same audio codec that your mobile phone uses (we quickly realised that you deserve better than that). It can scale this to any number of users, simply and cleanly and without major architectural changes and without you, the user, even knowing that it’s happened. We think that’s good.
In the process of doing this, we’re pushing the limits of Tim Berners-Lee’s original design for the web-browser. We’re covering new ground and doing things that have never been done before. That means that many of the issues we face are difficult and new. We are bound to make occasional mistakes, but we do learn from them and we strive to continually improve the way that Blabbelon works and to stay true to the original vision of putting voice on the web.
