[Following up on the messages we exchanged on Github, copy-pasting it bellow to keep it all here]
Jerome:
- As the Merge is getting ready, why not merging and canceling the Ice Age all together?
Tim:
Ideally, we never "reach" this difficulty bomb because we've merged before. But, if we haven't, I think it's better to have to push it back again. The alternatives are (1) remove the bomb altogether, which I don't think we should do (makes it easier for the PoW chain to keep going, also to launch scam forks) or (2) push the bomb *way* back, which is sort of equivalent to removing it in this context.
The Difficulty bomb was introduced has a mean to strongly incentivize the Ethereum community to switch to PoS. The difficulty bomb schedules a network slowdown when its deadline is missed. The network suffered from previous bombs:
- Blocktime went up to ~30s before EIP-649 came to the rescue
- Blocktime went up to ~20s before EIP-1234 saved the day
- Blocktime went up to ~17s before EIP-2387 cut the right wire
- Blocktime wasn’t impacted as EIP-3554 pushed back the bomb even started ticking
This EIP is proposing to diffuse and delay the bomb for the 5th time. I assume no one would argue that the 4 previous intervention were not justified, the readiness of the switch to PoS was miles away from what it is today. I assume that, with the state of the network today, having even 0.1 sec more delay between blocks would greatly impact the user experience (transaction fees likely going to the roof). Nevertheless, I am concerned that removing all together or delaying the bomb again would arm the credibility of the Merge.
If the Merge is close to ready, if we start to tease on social media that tests of merge from clients is going well, let’s the just not delay and do the Merge.