I just had an idea about the client part of web apps.

What if we produce the user’s dom on the server, convert it into video and stream it to him.

I know its not very efficient for heavy graphics apps, but there are a lot of other ones who would benefit from this system.
Each user will need to have his own dom-window on the server, rendered, based on the user’s screen size. Everything will be calculated there. Once we update his server dom we will take a snapshot of it, and create a streaming video to show in his local window.

The client will show the streaming video in an HTML5 canvas and also track clicks on it, which will be sent as coordinates in the user’s server dom for process.

We can even have various, advanced, html-like markup technologies for ui, since everything in the end will be based on canvas, video and click input coordinates.

Do you know if this thing exists or any tools to make it ?

I know its not very efficient for heavy graphics apps, but there are a lot of other ones who would benefit from this system.

Like what? I don't see why you would ever want to do this.

I've never heard of such a system using those specific methods and technologies, and I'm not really seeing the potential advantages. What you're talking about is conceptually similar to the concept of a "thin client", where the user's computer takes input and displays results but does minimal (if any) of the actual work -- which is instead done on another machine, often a central server -- but I think you've perhaps gone a bit far or got your technologies mixed up with the idea of sending video frames or using the HTML DOM.

Is that the sort of thing you're thinking about at all? Where were you thinking this might be beneficial?

OnLive?

[quote name='sermi' timestamp='1346607520' post='4975779']
I know its not very efficient for heavy graphics apps, but there are a lot of other ones who would benefit from this system.

Like what? I don't see why you would ever want to do this.
[/quote]

I've never heard of such a system using those specific methods and technologies, and I'm not really seeing the potential advantages. What you're talking about is conceptually similar to the concept of a "thin client", where the user's computer takes input and displays results but does minimal (if any) of the actual work -- which is instead done on another machine, often a central server -- but I think you've perhaps gone a bit far or got your technologies mixed up with the idea of sending video frames or using the HTML DOM.
Is that the sort of thing you're thinking about at all? Where were you thinking this might be beneficial?

Any kind of app with little movement on screen.
The reason behind this technology is not to mesh with client code of javascript or anything else.
It solves the server/client programming problem once and for all, by using the same language of choice everywhere, with any kind of advanced ui library.
Also the client devices can be very thin, they just need to play video.

OnLive?

This seems to be similar but it uses a client for each platform, i need html5.