Part B - Control error and computational cost

§	ID & Title	Status	Keywords & Search Queries	Dependencies
1	CMP.8 - Construct an Approximate Computation with Controlled Error	Usable, evolving	approximation; scaling; discretization; finite stopping; rounding; conditioning. How can a permitted error reduce computation while retaining the answer quality needed next?	MATH.20/.21 for bounds and convergence; CMP.3/.10 for shared computation and representation.
2	CMP.9 - Construct a Randomized Estimator or Sampling Procedure	Usable, evolving	randomized algorithm; sampling; estimator; proposal; dependence; stopping time. Which random procedure supplies the required law or finite-run estimate?	A supplied probability target, with MMP.7 where modeled; CMP.10 for access and storage.
3	CMP.10 - Choose a Computational Representation for Its Access and Update Operations	Usable, evolving	data structures; representation; queries; updates; conversion; arithmetic; memory. Which representation makes the required operations affordable?	CMP.2/.3 for compositional summaries and shared work; MATH for preserved structure.
4	CMP.11 - Derive a Computational Lower Bound from Indistinguishable Inputs	Usable, evolving	lower bound; adversary; indistinguishable inputs; decision tree; communication; error. What must every algorithm in this model observe or communicate?	MATH.19/.20 for argument and bound; CMP.1 for reduction; C.29.2 for the cost model.