This might be an equally effective but less centralized approach. What do you think?
How does one detect a fork? Some guy in his basement can create a fork with his own new code but be useless due to most transactions going with the main fork, but it would still be a fork.
Run several versions of Bitcoin and if they start to mutually deviate over three blocks then that is either a fork or is, a least, sufficient reason for merchants to go into safe mode. Sounds to me like a business opportunity.
Cool idea! And yes, this _is_ a business opportunity of, some size even.
The clients could not only be of different codebase (differnt versions and different clients), but also try to be connected to different "parts" of the network in order to also be able to detect a network split. Now add different configuration options and you have a whole bunch of instances you need to run and maintain.
Hmm, that begs the question: why run the instances yourself in the first place? Why not simply ask a bunch of clients about their opinion about the latest block (height and hash or whatever) via the protocol itself? Would that also be workable or does that approach have a problem?