>>1-3
He asked if it's possible for the "webmasta" to record that he clicked on a link, not the browser. That question is so obviously retarded that I refuse to answer it, but your answers are incorrect.
But of course you don't know if the browser implements onhashchange, so what you need to do is rewrite every link to end with !hash instead of #hash. This causes an (easily trackable) GET request to be sent to the server. The server can use mod_rewrite to respond with 302 to the original file with a #hash.
>>6
This got me thinking — disregarding the actual GC "debate", running code written by ``web developers'' without automatic memory management would be suicide.
Even if GC killed your parents, you'd be stupid not to want a GC'd language for your browser.
>>16
You'd be stupid to want to run any untrusted code in your browser. Web code should be native code run in a restricted subtask. Note that the kernel "collects garbage" by reference counting VM objects.