Of course nobody "wants" that. Yet adding more complexity is not necessarily the strategy to push transactions forward; often the best we can do is to help people find issues before the production phase.

All Hibernate libraries have always made extensive use of runtime exceptions for this purpose, it's a profound design choice on how to deal with software defects.

You wouldn't wrap all business operations with a try / catch / log&ignore block, no? We need to prevent issues from happening; in the backend code it having to deal with I/O and fledging things such as disk space, disk health you can't really prevent all problems.

I'm not strictly negative about the proposal: just pointing out the need to think carefully about which kind of errors you plan to catch, as most of the time this is a bad idea but I'm happy to accept that there are exceptions.