Classes: motivations

1. Problem analysis is based on data and algorithm break-down structuring $\Rightarrow$ hierarchical design for data and algorithms
2. Fewer bugs if data inter-dependency is low $\Rightarrow$ design data structure first, then associate algorithms to data (not the reverse)
3. Data structures are usually complex entities $\Rightarrow$ need for sufficiently rich expressive powers for data design
4. Different data objects may share some properties $\Rightarrow$ exploit this fact in hierarchical design