For example, dogecoin has 1 minute blocktimes and they seem to be doing fine for the past 5 years. What are some issues they might run into with 1 minute blocktimes and increased adoption? What other downsides does something like dogecoin have vs bitcoin, in relation to a shortened blocktime?
One of the great challenges there is that the security vs time tradeoff for a given block interval depends on network parameters that are hard to observe and change over time. If you use too long an interblock interval, nothing bad happens-- security grows slightly slower. If you use a somewhat too short interval security is significantly damaged.
The tradeoff in security gain per time vs block interval relative to propagation time looks something like this-- http://people.xiph.org/~greg/fn_secure_shape_napkin.png (iignore the actual specifics of the graph-- I'm citing it just to show the general shape of the tradeoff that I'm describing) so ideally you'd want to be at the peak, but since we can only get it approximately right, we should strongly prefer being slightly too long vs slightly too short, because too short can result in a pretty rapid drop off..
We've certainly had periods where blocks were taking a very long time to propagate-- until various improvements, like caching and the relay network solved the issues... but the conservative interblock interval in Bitcoin kept things from going off the rails. I find that hard to argue with, but it's easy to argue with if you forget that we don't just need to work on average, we need to work in adversarial conditions, all the time.
In alpha software prior to Bitcoin's release, the target interblock interval was 15 minutes, and it was reduced to 10 minutes for release.
Miner centralization, because the effects of latency would be stronger. That would give miners more of an incentive to be located close to each other.
For a longer explanation of these issues, read: https://petertodd.org/2016/block-publication-incentives-for-miners
If anything the block interval should get longer.
Shorter block times hurt decentralization
- because they hurt smaller mining pools
- because smaller mining pools depend on orphaning rate more
- because orphaning rate depends on the number of competing blocks in unit time
- since network connection quality becomes more important with higher orphaning rate
- and of competing blocks, the larger miner with better connection is more likely to win
- because the number of competing blocks in unit time depends on the number of blocks in unit time
Bitcoin has 6 blocks per hour. Doge has 60. Doge would have more competing blocks. In the end, larger miners are more likely to unfairly (i.e. not proportional to hashrate) win with the competing block. Hence longer block times means fairer block reward distribution to pools.
Unfair: Larger pools get larger more easily, smaller pools can't achieve their fair potential and would lose some % of their blocks hence their rewards and might not be able to compete.