I agree that data modeling is underestimated but it hardly can be considered a solved problem. It is very hard because there are numerous alternative understand

I agree that data modeling is underestimated but it hardly can be considered a solved problem. It is very hard because there are numerous alternative understandings and formal defs of what we mean by data (RM, 00, OR, MD etc.) In additiin, there are several levels of representation (physical, logical, semantic). In real projects, they all are mixed.