In the course of a conversation with Zooko, I started thinking about the transient nature of “expertise”. Both he and I might be considered experts in certain areas, centered around distributed data replication/distribution/etc. What I wonder is whether this expertise really represents our having climbed some particularly high pinnacle of knowledge, or whether it’s more like we’ve begun to open a gate that everyone will soon be walking through. Maybe ten years from now we’ll try to explain our contributions to some young punk who’ll respond, “What’s the big deal? Every first-term CS major knows that’s how it’s done.”

Sadly, that seems to be the general fate of innovators. People like Turing, von Neumann, or Dijkstra developed ideas so useful and so general that people just take them for granted now and can’t imagine that anyone ever considered alternatives. Nobody thinks twice about cache-coherent SMP machines now, but even within my own career I remember when that was new territory and nobody could quite agree on how to solve the basic problems. That stuff “just works” now, but only because hundreds of people tore their hair out developing methods and protocols that actually worked. Ditto for high-availability clusters. Maybe it’ll be the same for wide-area data access. Zooko says that will be a good thing, because at least it’ll reduce the number of people repeating the same basic mistakes. I guess he’s right, but it still saddens me that even the greatest innovators’ contributions can be forgotten so quickly.