Simon’s Blog

Scoble points to a recent Ray Ozzie speech, highlighting that Ozzie doesn’t think the Web is the be all and end all of software applications. Ozzie says,

Given the assumption of low bandwidth, it seemed very natural that the limited desktop terminal model, like that of a browser, would be required if centralized services were to broadly succeed. But today for many of us… the availability of inexpensive, high-bandwidth, always-on communications has become the norm. Our world has evolved into one with amazingly powerful “edge” devices, amazingly powerful centralized services, and high-bandwidth pipes connecting the two. And so rather than having to have a limited client and needing to put all the intelligence onto the service, we can for the first time consider how to intentionally balance where to put the application, where to put the data, and how rich to make the user experience based on factors such as mobility or the nature of the device, the nature of the Internet connection to that device and so on. Some applications or data are best kept on the centralized service—projecting their user experience through a browser or through software that’s temporarily downloaded onto a client. Other applications or data are best kept on a PC or mobile device, projecting subsets of that data to other users by temporarily uploading it onto a centralized service as a cache. That architectural choice is now ours, and I can’t sufficiently emphasize the importance and significance of this architectural choice that we now have.

Ozzie and Robert are quite correct; and actually, the network has been at this point even since the web browser first came into existence in the early 1990s. In other words, the choice of where to build the “intelligence”, and where to put the data in applications has been there for years.

However, the world went through a period in the late 1990s, and early 2000s where most people couldn’t see this. The conventional wisdom was that, if you weren’t building browser-based applications there was something badly wrong with your software, and your IT strategy. I’ve been on several panels at conferences where my co-panelists pushed their ideas of “everything in a browser” as though they were somehow great visionaries. In fact, it wasn’t vision. It was merely following the crowd - everyone was demanding their applications to be “web-enabled”. The truth is, however, browser-based applications are rarely compelling; and it has always been only a matter of time before people would come to realise this.

Why do I think that? Well, a little history will explain…

During the mid-to-late 1990s, we developed a software platform that you might call an “Application Browser”, as opposed to an “Web Browser”. It enabled us to develop rich, inter-linked, networked, robust, massively compelling, self-deploying software applications, at a time when everyone else was building what was frankly uninspiring software that ran in web-browsers (hell, there wasn’t even any AJAX back then!) that were rather flaky, and kept crashing.

The software was a massive success for us; although, the funny thing is, the idea for it came about completely by mistake. Or rather, from our incomplete understanding of what Java Application Servers were. We had just come back from a meeting where some software vendors were presenting early versions of ther brand new, super-expensive, software platforms - called App Servers. We simply couldn’t afford licensing costs of tens of thousands of dollars per CPU that these things were going to cost, but they sounded kinda neat, so we decided to build our own. However, at the meeting, we had gone around to each vendor asking, “Can you tell us exactly what an Application Server is, and what it does?” And the truth is, amazing as it may seem now, not one of the reps there had any idea what their App Servers did, so they couldn’t explain them to us! So, we headed back to the office, and brain-stormed the features we thought an App Server should have; and the end result was the above-mentioned App Browser technology.

Anyway, during the next several years, we demo’ed applications built on this platform to thousands of people at potential corporate partners all over the world. Without exception, there were gasps from the audiences as they saw how compelling the software was, and how completely different it was to the relatively low-quality experiences they were having from the web browser based software they were using.

So, back to the present. What has changed, and why is Ozzie’s speech timely? Well, I think there’s a growing understanding among users of applications that everything really doesn’t have to run in a web browser. And the reason for that is, in the same was as our App Browser applications touched thousands of people, software such as Second Life and iTunes is today touching tens of millions of people; and I believe the effect is the same. People are starting to demand more compelling user experiences from software providors than ever before - on their PCs and their mobile phones; and the fact is - web browsers aren’t delivering. This doesn’t mean that HTML/Javascript/XML web applications aren’t still hugely important. They are. It does mean, however, that they’re no longer the only game in town…