i have to tell you that despite the fact that i understand the theory and mechanics behind bitcoin and other derivatives, the hard fork is the hardest overall technique for me to sink my brain into.
it has always been my thought that a hard forking could be done by anyone and that it would take if more than 51% of the full nodes and/or nodes were to adopt it. i may be wrong.
I think you're wrong because the word "hard fork" has been misused up to the ETH split. Up to the ETH split, it has always been assumed that a change in protocol which renders it incompatible with the past, would simply MODIFY the future of a SINGLE block chain. That comes down to ALL nodes in the end accepting the "winning" chain, in the same way as competing miners finally accept the "longest chain" (or the chain with most PoW), even though during a certain time, there may be two competing forks (usually 2 or 3 blocks, rarely more).
In other words, the idea was that there would only be one single block chain in the end, the one that "won" (in length, PoW, or whatever the standard criterion is to prefer one over the other). In order to "win", you need 51% of the miners to be on the side of the winning chain for the whole time the competition is on.
But this is in fact an assumption that is not necessarily verified, because it is based upon a logical flaw: the competition between chains is only valid for chains with the same protocol. A hard fork changes the protocol. There is no "chain competition" between, say, the bitcoin chain, and the litecoin chain. Miners do not build on one, or the other as a function of the "longest one".
When you do a hard fork, you have, by definition, two different protocols (even though a lot is similar of course). If it weren't, it would be a soft fork. As such, you really actually created two different coins, with two different block chains.
The only difference between two different coins, and a hard forked coin, is that they have a common piece of past chain.
So yes, ANYBODY CAN PULL A HARD FORK, even on bitcoin, and that person doesn't need to own 51% of the mining capacity on his side. You can pull a hard fork of any coin in your basement, and then you've created a new coin. If people get interested in that coin, then this coin can live as a separate coin. The difference with a totally new coin, is that it has a common history up to the fork, and that all owners of the old coin are also holders of the new one. Your ICO is identical to the coin distribution at the moment of the fork of the old coin.
The only thing that is needed, is that people get interested in your coin, and that SOME miners get interested in your coin.
But given the nature of the ICO, the bulk of the coin possession is in the hands of people that were holding another coin. So in as much as they like THEIR coin, they may not be interested at all in yours, as yours may do competition to theirs. On the other hand, they all got "free money" so why wouldn't they use your coin, after all ?
With bitcoin however, there's an added difficulty. Bitcoin has a very conservative way of handling difficulty. So unless you have a lot of mining power on your hands, you won't even be able to mine a single block on your own in your basement, because bitcoin's difficulty is so high - unless you change that in your protocol of course.
I think that ETH/ETC is very interesting to observe, because it is the first REAL hard fork that ever happened, where the two prongs survive and get independent market existence.
Up to now, "hard fork" was simply meant "change of code and protocol", but the assumption was that the chain would not split. That assumption is now proven false.
This has a very interesting double consequence. The first is that the "power of the devs and the miners" has been shown actually not to be so great, which is good. If they decide to pull a hard fork, people are not obliged to follow. In the "single chain" assumption, the devs and the miners have the ultimate power, because "the longest chain wins". Not any more.
But the second consequence is much more obscure: probably
every coin, even bitcoin, will split one day. There is absolutely no reason that at no point in time, someone will not pull a bitcoin fork that will live on. And "forking" is in fact a way of saying "inflation". The number of coins doubles. They remain distinguishable, and hence can have other than 50/50 % market value, but the market cap is diluted over them.
The probability of bitcoin splitting one day is probably similar to the probability of a new chain starting (altcoin) in the first days.
Up to now, people thought that the only way of making a new coin was to make a new genesis block, and then the issue of ICO or other initial seigniorage indicated itself. But now we see that forking from a known coin is just as well a possibility, and there is much less discussion over ICO or premining: the actual state of coin distribution of, say, bitcoin is probably much better than an ICO or the core devs getting a lot of coins.
So the ETC/ETH fork may open an entirely new era in crypto, that of forking, and the associated inflation that goes with it, rendering the sound money doctrine somewhat superfluous. The ETH hard fork did much more to crypto than was ever thought.
