The Ethereum saga continues this month with the “third hard fork in the last four months.” This one includes a mechanism to let “developers” delete empty records that have been injected to slow the system. That’s right, this “decentralized” system is yet again modifying code to enable “centralzed” sysadmins to fix things.
As I have pointed out before, these kinds of operations are necessary and routine in most software systems, but they are difficult, politically controversial, and messy in decentralized blockchain systems. (Alyssa Hertig recounts that one group accidentally executed the patch—I mean, “the hard fork”—at the wrong time, effectively dividing the network fo ra time by accident.)
This story reveals some of the peculiarity imposed by the Nakamotoan ideology. The common and necessary practice called “issuing a patch” is a traumatic and newsworthy “hard fork”, which can go horribly wrong. The completely ordinary role of sysadmin cannot exist, and so must be simulated by a cadre of “developers” whose authority may or may not be recognized. This is no way to run a railroad!
As for Ethereum itself, by now, it is almost silly to consider it a “decentralized” system. The developers are clearly taking charge and actively managing the health of the code and the network. That’s a good thing engineeringwise, but it isn’t really a proper “blockchain” thing, is it?