I’ve written about the return of the client-server model over on the SproutCore blog. When the web started to get big, the Great Enemy was the client-server model. This was the model that dominated in 80’s when every software vendor your purchased from required you to install some special software (the client) to talk to their server.

In many ways, client-server computing was superior to the central-server approach of the web. Software running on your PC is faster, often more friendly, and easier to use. Despite these strengths, the web has largely won out over the last decade because client-server models required you to download and install software while web-sites do not. Only a few client-server apps still make it today, such as Skype.

That is all about to change however. Client-server is poised to make a comeback, surprisingly, perhaps, because of advancements on the web itself. The truth is, web browsers are now powerful enough that we can realistically write entire applications in JavaScript and run them there. This means we can create true “thick” clients with the rich, speedy interfaces of the desktop, but in the browser where they can be run without having to install them first.