The author critiques the trade-offs and flaws in Remix's approach to concurrent submissions and revalidation, highlighting issues such as race conditions, lag due to double round-trips, and incorrect assumptions about data versioning. Alternative solutions involving causal ordering and persistent connections, like those used in
Table of contents
Submission and revalidationHello DatabaseIntermission: Q&AWhat about single fetch mutation?In search of solutionsSolution #1: causal orderingSolution #2: persistence all the wayWe still have to talk about cancelled submissionsSort: