Michael Stonebraker recently challenged the NoSQL community's interpretation of the CAP theorem in his CACM blog post, 'Errors in Database Systems, Eventual Consistency, and the CAP Theorem.' Stonebraker argues that many common database errors, such as application bugs and data loss, are not mitigated by eventual consistency. He points out that eventual consistency does not protect against implementation errors or large-scale disasters where data clusters are lost. Stonebraker's critique suggests that relying solely on eventual consistency may not be the best approach for all workloads.

Stonebraker's article sparks debate over the suitability of eventual consistency in database systems. Many NoSQL databases have embraced eventual consistency as a solution to scaling issues, but Stonebraker questions whether this is necessary. He suggests that full consistency models can be implemented at scale and might be more intuitive for many applications. This perspective aligns with recent developments, like Amazon SimpleDB introducing support for full consistency, which offers developers more flexibility in choosing the most appropriate consistency model.

The discussion around CAP theorem and consistency models remains relevant as cloud computing and distributed systems evolve. As more applications require high availability and consistency, the debate between eventual and full consistency continues. Developers must weigh the trade-offs, considering factors like system complexity, error handling, and user experience. Stonebraker's insights encourage a closer examination of these trade-offs and suggest that full consistency might be a more robust default choice for many applications.