So imagine a tiny cluster of top 20 witnesses decided to get together and run code that claimed a block was irreversible, but then were like "just kidding" a few blocks later and voted to remove the block?
This is something new for me. I thought that every block is irreversible, and that this makes this platform resistant to censorship, because nothing is editable on the blockchain level (every post/comment edit/delete is visible on the blockchain level). But even if this (block removal) would be technically possible, I still do not see the reason, nor the goal behind it. So why would they do such a thing? For what? The only reason I see is deleting some kind of illegal content. But would not that (removing a block) break the integrity of the blockchain?
Every block becomes irreversible after a certain period of time, to account for temporary inconsistencies that pop up during processing. It takes a while for all of the nodes operating around the world on the not-always-reliable internet to get onto the same page, so to speak. Currently this happens after about a minute. The new system is to speed that up to normally a few seconds.
Crim is kinda explaining how and when it happens, why blocks get removed, with an extreme example that happened with Eclipse Hard Fork.
The blockchain is just a database and any database can be changed.
The only question is will the people who run the database allow it to be changed.
Blockchain makes it much more difficult to manipulate data because it's a public ledger.
But anyone can spin off a fork and decide something did or did not happen.