# Sleela Script Science and Engineering Turing 5 gives scientific and engineering work first-class language facilities. ## Numerical domains The standard profile reserves APIs for integers, arbitrary-precision integers, floating-point values, rational values, complex values, vectors, matrices, intervals, uncertainty bounds, quantities with units, symbolic expressions and observed/derived values. ## Calculus The language provides a calculus vocabulary including derivative, partial derivative, gradient, divergence, curl, integral, numerical quadrature, differential-equation solving, finite differences, limits, Taylor/series expansion, root finding and optimization. Example: let f = fn(x) { x^3 + 2*x } let slope = calculus.derivative(f, 2) let area = calculus.integrate(f, 0, 2) Symbolic and numerical results are distinct. A numerical approximation is never silently represented as an exact identity. ## Engineering quantities Where the unit system is available: let force = 120 N let distance = 4 m let work = force * distance assert units(work) == J Domain libraries can provide mechanics, electricity, signals, thermodynamics, materials, controls, structures and related engineering operations. ## Definitions and lookups Scientific lookup is explicit: let definition = lookup.definition("entropy") let constant = lookup.constant("speed_of_light") let material = lookup.material("aluminum") Lookup results can carry provenance and evidence status. External information is not silently promoted to an observed fact. ## Performance Pure mathematical expressions should be eligible for native optimization, vectorization, constant caching and specialized numerical kernels. Observable host operations remain explicit barriers so optimization cannot reorder them incorrectly. ## Scientific constants registry Demesresmes™ provides an explicit constants lookup surface. Canonical names are namespaced to avoid collisions between symbols such as mathematical `e` and the elementary charge `e`. ```sleela-script let c = constant.get("physics.speed_of_light") let h = constant.get("physics.planck") let kB = constant.get("physics.boltzmann") let NA = constant.get("physics.avogadro") let pi = constant.get("math.pi") ``` The initial registry is published as [constants/math-science.constants.json](constants/math-science.constants.json) and [constants/math-science.constants.xml](constants/math-science.constants.xml). Its physical reference values use the NIST/CODATA 2022 recommended dataset; NIST currently identifies the 2022 values as the latest CODATA values available and describes the next scheduled adjustment as 2026. citeturn0search2turn0search3 Scientific scripts should use the registry rather than hard-coding physical constants when provenance, reproducibility, or unit metadata matters.