The thing that bothers me about web development is that in order to get all that glitz and glam, you end up writing an application that practically mirrors a native application. At that stage, what have you really gained? A relatively slow and limited VM, often lesser development tools, but a really great distribution model.

I am constantly impressed by the boundaries people are pushing with the browser. Just the other day I read about someone releasing a cell phone tethering application that runs in the browser. That is amazing, but also ridiculous, in my mind, that it is useful beyond anything more than an academic exercise.

The web browser should be another network application, not the platform for network applications. I feel we could do so much better.

We've been hearing this lament from GUI developers for a long time. App development on the web is an abuse of HTML and HTTP and that it's a huge step backwards even from decades-old GUI toolkits.

The problem with that line of thinking is that GUI toolkits will never be cross-platform and cross-medium in the way the web is. It's not like there haven't been attempts, but they are all half-baked because there is no incentive for platforms to support them. The web on the other hand is must-have for every device. This momentum is what gives web app technology real legs.

It's really a textbook example of how worse-is-better can be disruptive, and how what we end up with has no correlation to what any one individual thinks is best. It doesn't really how much better an idea someone can come up with, the question is what can the industry be mobilized to get behind.

What you can gain is a central point of deployment, delivery and updates reachable by all platforms.