CMP.Preface:7 - Shared sources, alternatives and relations

Erickson’s algorithm construction material and Morin’s Open Data Structures support the connection between a problem, its representations, a constructed procedure and its argument. The bodies retain explicit conversions, recurrences and operation costs. Direct use of a known algorithm remains preferable when constructing another one adds no needed capability.

The approximation and learning sources cited in CMP.5-.9 connect a deliberately weakened or data-dependent answer to the procedure that obtains it. They also expose limits: a recovered relaxed solution needs feasibility and quality arguments; an optimized training criterion leaves generalization conditions to be established. CMP.8 compares its transparent scaling construction with a stronger recent knapsack scheme. CMP.7 compares historical assessment with adaptive assessment under temporal change. Those comparisons justify reopening the corresponding method when another construction changes cost or supported use.

SICP’s evaluator and compiler construction supplies an explicit historical demonstration of programs that interpret and transform programs. CMP.12 combines that construction with current behavior-preservation distinctions. Cousot’s Principles of Abstract Interpretation develops computable summaries from the properties they must support; CMP.13 compares direct exploration, coarser abstractions and targeted refinement. These approaches answer different questions: a useful overapproximation can deliberately introduce behaviors that a meaning-preserving translation would have to treat differently.

Lamport’s A Science of Concurrent Programs and the composition, memory and crash studies discussed in CMP.14 support its separation of observable behavior, permitted interaction and progress. Sequential composition is cheaper when interference is absent. More detailed composition arguments earn their place when shared observations or failures change what the whole computation can do. Each body gives the adopted contribution, serious alternative and conditions that would reopen its selection.

The Suite Reference locates the related publications and explains their shared architecture. C.29.2 supplies computational formulation; C.29.3 physical realization; C.29.1 the transfer between mathematical accounts; C.29 the model-to-subject correspondence. B.5’s inquiry methods support the next question when a limit or failure makes the old one insufficient. Notational Engineering supplies methods of expression design as those contributions become available; the operative rules stated in CMP remain usable without an unwritten supplier.

The three Parts group the presentation. The fourteen bodies form a repertoire of related Methods, usable individually or in combinations selected for the question. A particular connected use can describe a composite way of working, but membership in this publication alone does not make every method a mandatory step of that work.