Measurement, Approximation & Numerical Error

← Back to Domain-Specific Families

Abstractions about quantifying and approximating physical or computational quantities, spanning error and precision concepts (truncation error, arithmetic underflow, significant figures), numerical approximation methods (iterative methods, residuals, log-log plots), and measurement conventions (physical quantity, metrication, coefficient).

17 abstractions in this family — domain-specific abstractions that sit near one another in structural-signature space (k-means over structural-signature embeddings). Each is shown with its short description.

• ABC Software Metric — A software-size metric reported as the triplet of assignment, branch and condition counts for a code unit, with an optional Euclidean magnitude summary.
• Arithmetic underflow — A finite-precision arithmetic condition in which a nonzero result's magnitude is too small for normal representation in the destination format, causing a subnormal result, rounded zero or signaled exception.
• Asymptotic analysis — The study of approximations and expansions that become accurate as a variable approaches a specified limit, emphasizing dominant scales and controlled remainder behavior.
• Circular convolution — A convolution on periodic sequences in which indices wrap modulo a fixed length, corresponding to pointwise multiplication under the discrete Fourier transform.
• Coefficient — A multiplicative factor attached to a term in an algebraic expression, series, equation or linear combination, determining that term's scale under a stated basis or representation.
• Differential of a function — The linear map giving the first-order change of a differentiable function at a point, written df and represented in one variable by dy=f′(x)dx.
• Iterative method — A numerical procedure that repeatedly updates an approximation from prior iterates toward a solution under a convergence rule.
• Log–log plot — Plot positive x and y values on logarithmic axes so multiplicative ratios become equal distances and a power law y=ax^k becomes a straight line with slope k and intercept log a.
• Metrication — The institutional transition from customary or local measurement systems to the metric system through legal standards, technical conversion, education and changed everyday practice.
• Phase response — The frequency-dependent phase difference between a system's sinusoidal output and input, forming the angular component of its complex transfer function.
• Physical quantity — A measurable property of a phenomenon, body or substance represented by a numerical value together with a measurement unit and quantity kind.
• Process window index — A normalized manufacturing metric equal to the largest absolute deviation of a process-profile statistic from its specification center, expressed relative to the allowed half-window.
• Pseudo-differential operator — An operator defined through a position- and frequency-dependent symbol, extending differential operators to non-polynomial multipliers and supporting microlocal analysis of PDE.
• Residual (numerical analysis) — The discrepancy obtained by substituting an approximate solution into the original equation, usually b−f(x₀) or its signed convention.
• Running total — A cumulative sum updated incrementally by adding each new observation to the previous total, yielding every prefix sum without recomputing the full sequence.
• Significant figures — Digits retained in a measured or calculated quantity to communicate precision under a stated measurement and rounding convention.
• Truncation error — The discrepancy introduced when an exact infinite, limiting or continuous mathematical process is replaced by a finite approximation, distinct from finite-precision roundoff and analyzable through omitted terms or local discretization expansions.