Robert Scoble: "...at Microsoft we call Internet apps that aren't in the browser 'Smart Clients'".

The web browser is at a crossroads. Microsoft announced in 2003 that it would not release any further "standalone" versions of Internet Explorer - instead it will be embedded in the Operating System (codenamed Longhorn). But along with obsoleting the web browser as a standalone product, Microsoft is also transforming its definition of a web browser.

IE will no longer be used as a central application from which to run other web applications. Applications will increasingly be non-browser, they will be "Smart Clients" requiring separate installation on a user's PC or other Internet-able device.

The other day I came across a new RSS Aggregator product that is currently in beta, called Lucmo. It is unusual for an RSS Aggregator because it is a browser-based application. Most other products of its kind are software apps that require installation on your PC - eg NetNewsWire and Feedreader.

Lucmo decided to build a browser-based application for these reasons:

"We believe that there are significant benefits to a centralized model. Users will not have to install any software. We can build collaborative filtering more easily. Like webmail, users can access it anywhere without sitting in front of their own computer."

The first and third reasons - no installation and access anywhere - are standard and powerful benefits of the browser-based model. But it's the second reason that interests me most. Lucmo believes that they can achieve "collaborative filtering" more easily in the browser than in a standalone app. Lucmo are hunting the Great White, the Semantic Web. The Lucmo blog explains:

"The killer feature, though, will be the implementation of what you might call a reverse bayesian filter: the user's interests are determined by analyzing every incoming news bit and at the other end recently changed weblogs are aggregated and feeds that comply with the user's interest are suggested. Next step is the implementation of watches: the user creates a watch by specifying a list of keywords, which are then used to filter either news bits from subscribed feeds or all recently changed weblogs."

Simon Carstensen, one of the Lucmo developers, has further information on his weblog. From what I understand, Lucmo aims to personalise RSS feeds by 1) guessing a user's interests and delivering matching feeds to her; and 2) allowing the user to manually specify her interests (what Lucmo calls a "watch") and the system automatically collects and delivers feeds that match those interests. This level of automation is what will drive the Semantic Web to reality. But it's interesting that Lucmo plans to do this in the browser...rather than a 'Smart Client'. 

Are Microsoft selling the web browser short? The browser is still a powerful tool that, as Lucmo will hopefully demonstrate, is a more than capable platform for personalised, automated Web applications.

No Microsoft aren't selling the browser short - but they are sweeping it under the carpet. You see Microsoft is in a privileged position. It doesn't need to build its web applications for the browser anymore, it can tack them onto its OS or otherwise add it into one of its ubiquitous software packages (e.g. Microsoft Office).

Let's return to the 3 reasons Lucmo has for building its app in the browser rather than as a standalone product, and see how Microsoft's non-browser Smart Clients compare:

1) Users don't have to install a new piece of software. Microsoft's Smart Clients won't have to be installed either. They'll be pre-installed, either in the OS or as part of a package like Office. The user doesn't need to worry about installing something new, it's all part of the Microsoft experience.

2) Collaborative filtering. Microsoft has the best of both worlds. Smart Clients will have the increased functionality of a desktop app, while at the same time they can 'hook into' the browser (conveniently sitting on the same OS) in order to network with the World Wide Web. Smart Clients may even share some of their components with IE.

3) Access anywhere. The browser still holds the advantage here, because it (mostly) adheres to universal Web standards. But Microsoft has such a large worldwide market share of desktops, that access from anywhere is an attainable goal for them.

In summary, Smart Clients (non-browser Internet applications) will have most if not all of the functionality enjoyed by browser-based applications. At least they will for Microsoft - because their Smart Clients won't need to be installed, they will have access to browser components on the OS, and they may even have universal access due to Microsoft's market position. Independent companies and individuals will have to work harder to get their Smart Clients installed on user desktops. 

Meanwhile maybe Lucmo can prove that there is life in the browser-based model of web applications. I've already signed up and will be following its development with interest.