So things like Gears, Silverlight, and Apollo look like they’re here to simplify the development of rich internet application. We’ll be able to run our rich web apps on and off line. But aren’t we still talking about HTTP, the hyper text transfer protocol? Ultimately, the web we experience through a browser was created to serve up text. So why are we trying to shoehorn in full blown applications?
For all their javascript headaches, browsers are still fairly standard. They are the write-once-run-anywhere dream. Everyone has a browser, and everyone can run your webapp. So at first glance it seems natural that these webapps evolve and become more and more complex. But remember HTTP and consequently the browser weren’t written to run applications. The ultra rich web application just doesn’t quite make sense. Why waste all the time re-implementing mostly solved problems with inadequate tools? We shouldn’t be editing movies through a browser; it’s just silly.
Ultimately we’re in this position because we’ve failed to create rich client side applications that can be easily distributed and updated. I think that Java’s webstart was a nod in the right direction but very poorly executed. Also, the appeal of web applications goes beyond the distribution; we also want to get at our stuff no matter where we are. Can we build rich apps that live on and store their data on something like S3?
I want my documents and apps, but I don’t want them through a browser.
Comments !