Thanks for the info on that, but then how to avoid address resuse? If i have 10 coins and I send you 2 coins, what I do with the 8 left in my wallet? Are you saying I have to immediately move them to another address?
Your wallet already does that. Bitcoin doesn't work on the concept of balances it works on the concept of creating and destroying outputs.
So if you have an output worth 10 BTC and you want to send me two your client creates a tx which destroys the 10 BTC output and creates two new outputs valued at 2 BTC (to me) and 8 BTC (to a new address in your wallet).
while we're on the topic of "can wallets be bruteforce cracked"...
when we talk about supercomputer speed (petaflops, etc) (floating point operations) -- how many
floating point operations actually go into trying 1 private key ?
Zero. We are interested in integer math when brute forcing private keys. Flops refers to floating point math. There is no conversion factor which would work for all systems. Generally speaking computer science doesn't look at the individual implementations to determine if something is infeasible.
As an example say a given computer would require 1000 steps to make one attempt to brute force a key and it will take more energy than 20 of our suns and 10 billion years. Now lets say you could reduce that to a single step. Ok now it only takes 1 billion years and more energy than 2 of our suns. Either way it is infeasible.
Any classical computing problem which is more complex then O (2^128) is generally viewed as infeasible. Some people use the word improbable but infeasible is a stronger word. It is improbable you will win the lottery however it is infeasible that you will brute force a 128 bit symmetric key (simply requires material and energy on a scale the human race is utterly incapable of). In comparison to the lottery it would be like you win seven lotteries in a row by purchasing just a single ticket.