[josh2600]

This is canon and dates to the original DAO. A quick google on DAO crisis should give you a bunch of reading material.

Compare that to the time Binance lost $500M in BTC and contemplated doing a chain revert because it would be way cheaper than $500M but didn't because it would break the illusion that BTC tx are irreversible.

BTC tx are reversible, it's just going to become prohibitively expensive for almost anyone to do soon.

[ryebit]

That is rather misleading to use as an example... It's widely regarded as a one time thing that was agreed upon partly because the system (and testing tools) were so new at the time.

Bitcoin had a bug that was required an organized chain reorg back when it was early as well, but it's not regarded as fatal to putting trust in the chain now (search for "bitcoin value overflow bug", was kinda interesting)

More recently, Parity, one of the main developers of one the Ethereum clients, had millions locked up due to mistake in their smart contract code. They complained loudly, tried to get another roll back, and finally gave up... The overwhelming concensus was that Ethereum is no longer alpha grade, there will be no take backs ever again.

[josh2600]

To be clear, I wasn't asserting that Eth was less valuable as a result of the DAO reorg (quite the contrary actually). I was merely pointing out that a reorg occurred.

Thank you for the education on the bitcoin reorg, I didn't actually know that and agree that it's orthogonal to the trust narrative now.

My broad point was that the trust is not in the code, it's in the community and their policies.

[ryebit]

> the trust is not in the code, it's in the community and their policies

I can heartily agree with that.

That's why I think having the "layer 0" social consensus of a network aim for maximum clarity, so participants have as few points where there's three potential for surprise disagreement later.

Ethereum is very strongly "no abnormal state changes" starting with the Parity issue. I think there have been a few more similar cases of contract bugs, some even involving client devs, and any suggestion of a new fork has met with strong opposition from all ends of the community.

Another thing I think is useful for a blockchain is to have multiple independent clients... It helps prevent devs from having outsize voice in discussion (though the users and the node runners are always the final vote).

I think working out the meta structure of how to work these social level contracts is definitely something the whole cryptocurrency industry needs to work on.