Well, I am always reminded of
this post by satoshi:
I don't believe a second, compatible implementation of Bitcoin will ever be a good idea. So much of the design depends on all nodes getting exactly identical results in lockstep that a second implementation would be a menace to the network. The MIT license is compatible with all other licenses and commercial uses, so there is no need to rewrite it from a licensing standpoint.
Although there are so many scenarios Satoshi could not have been expected to plan for. I've always been a fan of alternative implementations although they must be maintained and tested with utmost care and precision.
I think this is an artifact of Satoshi being stronger in crypto than in software engineering.