From a consensus persective, each different version of a codebase is a node that could behave differently; the same even applies to hardware and configuration parameters, although if the code is written properly, those aren't supposed to have consensus-critical effects.

As you can see from BIP0050, the decision in that case was actually to downgrade the update, despite the fact that one could argue that the new version actually fixed an undocumented bug in the old version.

adlai

bitcoin

Core & Knots aren't the only choices

siggy47

>  each difference in the node's behavior could lead to forking away from the main network

> happened just between different versions of core, is the March 2013 Fork

I think this illustrates the case for competing implementations, if any one of them has a bad update the damage is relatively confined... whereas if 99% of the network is running one implementation and it has a bad upgrade, causing a substantial chunk of the network to upgrade with it, that could be much more damaging.

All eggs in one basket theory