Quote from: kokojie on August 22, 2017, 07:37:14 PM

Just throwing it out there, would anyone be interested in a Bitcoin fork that has faster blocks (Let's say 10 second blocks). Block size 0.2MB. Effectively increasing block size to 12MB per 10 minute

Advantages:
* Time to 1st confirmation will be a few seconds on average, good enough for even buying coffee and actually getting a confirmation, instead of relying on completely insecure 0-conf or 3rd party service
* Much better tx/sec capacity obviously
* Better user experience overall, it will just feel fluid for all tx, anyone that has used a sub-10 sec altcoin should know the feeling. Everything is just fluid

Disadvantages:
* Potentially bigger blockchain size obviously, but that's not really a big problem unless we are doing VISA level transaction volume
* Orphans for everyone, but that's fair

I agree that a increase in the block creation rate would be useful. It would allow for much quicker settlement of transactions, which I consider very important if we want that Bitcoin becomes more mainstream. However I see a couple of problems:

The problems associated with much higher block creation rates have been thoroughly studied. It causes, like mentioned before, an increase in the orphan block rate and so also a decrease in the efficiency of the network, which makes the network more vulnerable to attacks. More computing power is wasted on orphans, so the attacker needs less computing power to have a reasonable chance of success. However, solutions have been suggested for this problem. One of it is GHOST, which chooses the chain to mine on top of differently. An interesting article is Sompolinsky, Zohar "Fast money grows on trees, not chains". A more accessible article is of Vitalik Buterin "Toward a 12-second block time". These changes allow for very high block frequencies.

A different problem is of course that the higher block frequency will make it more resource intensive to run a node in my opinion. Not only will the block chain increase rapidly in size, also the bandwidth usage and probably CPU usage will increase. This will result in fewer people running nodes, which will lead to more centralization. This is in my opinion something that we really need to avoid. So i think we would also need some kind of compression method, so that this burden is decreased, and kept within the range of ordinary users. Even an increase to 0.2 MB per 10 seconds or 12MB per 10 minutes will in my opinion be a problem.

Next to that there is in my opinion also an implementation issue. It would require a massive rewriting of the Bitcoin code and will be a hard fork. It would in my opinion be much more difficult to implement this than a "simple" increase of the block size like segwit2x, etc... suggest. This makes it in my opinion practically impossible to do.

A simpler solution would be to increase the transaction processing by using sidechains or something like that. For example there would be a chain that starts with a few relevant bitcoin UTXO's which serve as the basis for the sidechain. Transactions are done quickly on the sidechain, which has a much higher block creation rate (for example every 10 seconds). This can be done by using GHOST and small blocks. Once every few hours a settlement is done on the Bitcoin blockchain on the basis of the end result of the balances on the sidechain. A bit like the Lightning network but using a blockchain.