Bitcoin is only divisible to 8 decimal places without a hard fork. Hard forks are very difficult to orchestrate, so although extending the number of decimal places is technical feasible, it will be difficult in practice and will never happen before it is agreed to be an absolute necessity.
When bitcoin first rolled out it was 4 decimal places. Back in 2011 they decided that 8 were needed and it came out as part of a core update. Similar thing with the view on "dust" last year. Those sorts of updates are pretty minimal and don't require a big change. Changing the hashing algorithm or moving to PoS or something would be a big deal. So if more than 8 decimal places are required, I think it will happen without causing any issues.
If required, yes. But if you try to do it before there is consensus that it is necessary you will end up with two blockchains each claiming to be Bitcoin. Such is the case any time you try to change the rules in a way that is incompatible with older versions of the software. The "dust" changes were merely parameter adjustments. Dust is still valid and is included by some miners, but most choose to filter it. So not really that similar. Also, you're full of shit because I was 0.5% of the hashpower in 2010 and at that time we had 8 decimal places. It might have been 4 in a very early version, but certainly not in 2011.
True, there's always a chance that not all people will choose to play by the new rules causing a fork - I'm just suggesting that adding a few more decimal places should the need arise is not something that I can see many (any?) people rebelling against, so I am going to guess it would not be hard to roll out.
I might very well be wrong about the date of the change - I first learned of Bitcoin back in 2011, and I remember the change to 8 places, so I assumed it was then, however if this is wrong, then my bad.
Well, just off the top of my head, how about the fact that increasing the number of decimal places means that you have to change the value field on every txout to something larger than the current 8 bytes. This increases transaction size, which means that without also increasing block size, you are decreasing transaction throughput of the network. Why would a miner want to increase storage costs, decrease throughput, all while taking in less fees (because of less transactions)? I assume if we are at the point where 8 decimals isn't enough, we are also pushing the limits on throughput, but blocksize is an entirely different can of worms.
