1. 14

  2. 4

    I agree, to some degree, but I think this blog post took a really complicated topic and boiled it down too much. Not every problem should be solved with eventual consistency, some problems do need to strong consistency. I agree with the author that too many people start at strong consistency. They should really be starting with eventual consistency and figuring out which problems they cannot solve with that.

    As for the blog post, the author doesn’t link the title with the content. The content is rehashed quotes and idea from other distributed systems people, and the issue of being a UX problem is never actually brought in, just stated. I would phrase the title differently: Distributed Systems Have UX Constraints. Saying it’s a problem implies there is a solution, to me. In reality there is no solution, just different sets of constraints that all have their own UX problem. Strong consistency means the UX might not be there at all during an outage. Eventual consistency might mean the UX makes no sense for a little while.