The Boost document reads again, I feel that "libraries listed by category" section of the Boost is not suitable for understanding the composition of Boost, according to their own understanding, the class:
1, Standalone Library for Applications: Graph, Thread, Python, MPL, Preprocessor, Math ...
Put MATH here is a bit of difficult
2, Complement Library for Stl: Container: Array, Multi_Array, Dynamic_bitSet, Property_Map ... Iterator: Iterator_Adapters ... functor: Bind, Compose, Function, Functional, Lambda ...
3, Generic Concept for Developer:
Call_Traits, Type_Traits, Concept_Check, Static_assert, Base from Member, Chain of Base ... These libraries or concepts are greater than the meaning of the code itself than the meaning of the code.
4, Utility: All Kind of Cast, All Kind of String Parser, Regex, FileSystem, DateTime, Any, Smart_ptr, Tuple, Timer, Test, Operators, IO_STATE_SAVER, RANDOM, SIGNALS ...
This part is part of the dragon snake mixed, but there is no shortage of standard library candidates such as Smart_Ptr et al.
