Just once I'd like to see someone write an opinion about ORMs without resorting to sweeping generalizations ("Vietnam of computer science") or making assumption

Just once I'd like to see someone write an opinion about ORMs without resorting to sweeping generalizations ("Vietnam of computer science") or making assumptions about how they are being used ("it wasn't a good fit for my use case so therefore it must never be"). Maybe, just maybe, it's possible that ORMs are useful for solving certain types of problems and less suitable for other types. If you work on web stuff or applications that are report-oriented where most of the time you're just fetching data (possibly with complex queries) and rendering it to display, then maybe ORMs aren't a good fit. On the other hand if you work on client-side apps where your objects are backed by a database but are otherwise long-lived, then sometimes the other features (beyond SQL generation) that ORMs provide come in handy (tracking units of work over the object graph, maintaining an identity map for consistent access to objects, and providing change notifications when an object or collection of objects is manipulated). If you've never needed any of these features that's perfectly fine. I've never to needed to use a bulldozer either. But I'm not going around declaring bulldozers useless just because I've only ever needed to use a shovel.