This is not entirely true. If the reward from my work isn't enough to provide me profit then I'll stop mining. But, if many people do that then the difficulty will decrease and this will result on making mining more profitable for me. Thus, it turns out that the total units of bitcoins you're rewarded don't have to do with your profit. We may see huge fluctuations of difficulty in the future. Recently, if I'm not mistaken, around 20% of the total power consumption for mining was lost. Cheers to the miners that will work within the next difficulty adjustment's period.

As I said, the fee is based on the amount of transactions in the mempool. Whether the reward gave 50 BTC or 0 BTC, the current transaction median fee would be the same.

Of course and they will. The cryptocurrencies you mentioned (BCH & LTC) are much more useful if you want to broadcast enormous transactions with low fees. You should consider, though, that if many people think same-like, their usefulness will be ruined.

This opens another discussion, again. There are surely altcoins that process transactions faster, such as Litecoin with 2.5 minutes per block. The fact that someone chose 10 minutes back in 2009 isn't a non-sense randomly-made decision.

If they invested all that fortune on performing such powerful attack, then they could stop Bitcoin. It's pretty simple if they just extend the chain with empty blocks. But it's not beneficial for them! Despite the cost of the miners that should work consecutively, they should do that for every altcoin existing! As said by Wind_FURY it's a game theory and they ought to play by the rules. It'd be better for them if they helped the network and got rewarded in bitcoins.