Read on Twitter
If you're deep in the technical weeds, it's easy to forget how much small things impact less technical users. Here is a list of the biggest issues I presently see as affecting medium-technical level users if we go live with the deposit contract now:
1. No detection of previous deposits on the ETH2 Launch pad validator deposit flow. People WILL mess this up without it (see here for why: https://twitter.com/technocrypto/status/1318465927130787840 ). There's an open issue to watch for it here: https://github.com/ethereum/eth2.0-deposit/issues/165
2. The default tool for generating keys does not allow you to regenerate your keys from your backup mnemonic and the instructions don't cover it. Open issue: https://github.com/ethereum/eth2.0-deposit-cli/issues/29 @CarlBeek says he wants this before launch but #1 above is blocking it rn https://discord.com/channels/595666850260713488/720658991763685377/764742903171252234
Note that technical users can just use ethdo for this: https://github.com/wealdtech/ethdo but it's not an official tool and is still undergoing audit. Ideally we would have both and be able to verify them against each other. But that brings me to:
3. There have been an enormous flurry of releases from the major clients (many of which are breaking changes, and they are not even done yet). Downstream projects like @DAppNode, the guides, the block explorers, etc. are still scrambling to keep up.
So what? Well, deploying the deposit contract starts a clock. Major client devs may feel confident that they can keep up with that clock, but it's not clear that they're leaving room for downstream projects to integrate their last-minute sweeping changes.
In some cases the recent changes have placed additional issues (like finessing and standardising the cross-client validator migration and slashing protection details) onto the critical path to maintain basic functionality. Once we start using real money this stuff matters.
4. Recent issues during non-finality on Medalla have revealed sync and memory problems. The most urgent are being fixed (see #3) but a lot of users who followed the "recommended specs" ended up unpleasantly surprised. We need to be much clearer about node memory requirements.
Here are the issues for Prysm: https://github.com/prysmaticlabs/prysm/issues/7585 and Lighthouse: https://github.com/sigp/lighthouse/issues/1795
5. The Medalla situation has also revealed that block explorers don't yet function properly during longer periods of non-finality, which is right when less technical users especially depend on them. See https://twitter.com/technocrypto/status/1318610920642314240
There are currently efforts to figure out whether this is a client performance issue or something explorers can fix themselves: https://discord.com/channels/694822223575384095/694822223575384099/768214404881449060 but this is a further reminder that there are many downstream projects which will matter a lot to average users.
Again, people are working on all of these things. They just need the time to get a little bit further ahead before we press the big button. Yes, we are at the stage where the most technical people would probably be okay. But decentralization happens mostly at the edges!
The question of how decentralized #Ethereum #Proofofstake actually ends up being is partly a social question about how much effort we make to lower barriers to participation. The reward for taking just a little more time to get these things solid will be more solo stakers. #ETH2
