With regards to game distribution, I don't see any particular advantage that a browser has over something like Steam. You don't need to download multiple gigabytes either. You could design a game to stream levels/assets/etc on-demand. In any case, even with webGL, a game would need to do that.
Absolutely not; the sandbox of Javascript apps is entirely within the browser. The CPU architecture may theoretically prevent a userspace app from breaking into ring 0, but that's not the interesting thing; you're trying to prevent random apps from gaining access to all the tasty data in userland.
"Please do not download and run random executables" is a lesson that security people have been trying to get across for a while. Running a native game executable involves trusting it not to be malware. Running a webgl or HTML5 game does not.
(The sandboxing of webgl may be dodgy, which is a weakness here.)
I don't know why you felt compelled to state a categorical denial when the facts dispute it. Please read up on how security tokens work in Windows, for e.g.
http://msdn.microsoft.com/en-us/library/bb625957.aspx
> Running a native game executable involves trusting it not to be malware. Running a webgl or HTML5 game does not.
There is no difference between those two executions in modern operating systems that implement mandatory access control.
It is convenient. What is it not understandable about that?
I already excluded this part from my criticism. Please read my original comment.
>You can't say the same for Steam.
So, the solution to that is to create something like Steam, but better. Choosing an inferior technology is the opposite of a solution.
I predict that advertisers are going to overuse webGL just like flash. We're going to see webGL blockers crop up the minute it gets any kind of traction. Certainly, I am not going to subject my machine to another malware vector.
If your application is inherently networked, it works just fine in the browser.