A star schema contains Sales facts and Product and Date dimensions. Why avoid storing independent dimension descriptions repeatedly in every fact row when unnecessary?

1. Separate dimensions improve consistency and reduce repeated attributes
2. Fact rows may never contain any keys
3. Dimensions must contain only numeric measures
4. Normalization always requires removing every dimension