CODATA fundamental physical constants as typed, dimensioned quantities in
pure Standard ML. Every constant is a Units.quantity — a real magnitude in SI
base units tagged with its physical dimension — so the relations between
constants (hbar = h/(2 pi), R = N_A k, c = 1/sqrt(mu0 eps0)) check out
dimensionally as well as numerically, and adding a length to a time is a
runtime error rather than silent nonsense.
This library builds on, and vendors, sml-units
(see Relationship to sml-units below). No
dependencies to fetch, no FFI, no threads, no clock, no randomness: the same
inputs always produce the same outputs under MLton and Poly/ML.
- The seven constants fixed exactly by the 2019 SI redefinition
(
speedOfLight,planck,elementaryCharge,boltzmann,avogadro, plus the conventionalstandardGravity) carry their exact defining values. - The measured constants (
gravitational,electronMass,protonMass,vacuumPermeability,vacuumPermittivity,gasConstant,stefanBoltzmann,fineStructure) use the CODATA 2018 recommended values. - Derived helpers:
photonEnergy(E = h f),photonMomentum(p = h f / c),deBroglie(lambda = h / m v).
sml-constants does not reimplement dimensional analysis. It vendors
sml-units verbatim under lib/github.com/sjqtentacles/sml-units/ (the copy
is byte-for-byte identical to the upstream library, verified with
diff -rq) and loads those sources before constants.sml in
sources.mlb. Each constant is then defined as a genuine Units.quantity,
built compositionally from the sml-units catalogue of named SI units (so the
dimensions are computed by sml-units, not asserted by hand). The vendored
structure is re-exported as Constants.Units, so callers get the full
dimensional-arithmetic API for free.
This is the true integration path described in the build plan (not the
real-valued fallback): constants are quantities, dimensions are tracked at
runtime, and unit conversions go through Units. CI uses Variant A — the
apt Poly/ML builds the vendored sml-units + sml-constants without trouble.
structure Constants : sig
structure Units : UNITS (* the vendored dimensional library *)
(* exact 2019-SI defining constants *)
val speedOfLight : Units.quantity (* c = 299792458 m/s *)
val planck : Units.quantity (* h = 6.62607015e-34 J s *)
val elementaryCharge : Units.quantity (* e = 1.602176634e-19 C *)
val boltzmann : Units.quantity (* k = 1.380649e-23 J/K *)
val avogadro : Units.quantity (* N_A = 6.02214076e23 /mol *)
(* derived / CODATA-2018 measured constants *)
val reducedPlanck : Units.quantity (* hbar = h/(2 pi) *)
val gravitational : Units.quantity (* G = 6.67430e-11 *)
val electronMass : Units.quantity
val protonMass : Units.quantity
val vacuumPermeability : Units.quantity (* mu_0 *)
val vacuumPermittivity : Units.quantity (* eps_0 *)
val gasConstant : Units.quantity (* R = N_A k *)
val stefanBoltzmann : Units.quantity
val fineStructure : Units.quantity (* alpha (dimensionless) *)
val standardGravity : Units.quantity (* g_0 = 9.80665 m/s^2 *)
val value : Units.quantity -> real (* SI magnitude *)
val valueIn : Units.quantity * Units.quantity -> real (* magnitude in u *)
val photonEnergy : real -> Units.quantity (* E = h f *)
val photonMomentum : real -> Units.quantity (* p = h f / c *)
val deBroglie : real * real -> Units.quantity (* h / (m v) *)
endval c = Constants.value Constants.speedOfLight (* 299792458.0 *)
val () = print (Units.Dim.toString (Units.dim Constants.planck)) (* "kg m^2 s^-1" *)
(* derived relations, checked within epsilon *)
val hbar = Constants.value Constants.reducedPlanck
val ok = Real.abs (hbar - Constants.value Constants.planck / 6.283185307) < 1e~15
(* photon energy of green light *)
val e = Constants.value (Constants.photonEnergy 5.45e14) (* ~3.61e-19 J *)Running examples/demo.sml with make example prints:
CODATA fundamental constants (SI base units):
c (speed of light) = 2.99792458E8 m s^-1
h (Planck) = 6.62607015E-34 kg m^2 s^-1
hbar(reduced Planck) = 1.05457182E-34 kg m^2 s^-1
e (elementary charge) = 1.60217663E-19 s A
k (Boltzmann) = 1.38064900E-23 kg m^2 s^-2 K^-1
NA (Avogadro) = 6.02214076E23 mol^-1
G (gravitational) = 6.67430000E-11 kg^-1 m^3 s^-2
R (molar gas) = 8.31446262E0 kg m^2 s^-2 K^-1 mol^-1
Derived relations hold:
h / (2 pi) = 1.05457182E-34 vs hbar = 1.05457182E-34
NA * k = 8.31446262E0 vs R = 8.31446262E0
1/sqrt(mu0 eps0)= 2.99792458E8 vs c = 2.99792458E8
Applications:
photon energy at 5.45E14 Hz = 3.611208E-19 J
electron de Broglie wavelength at 1.0e6 m/s = 7.273895E-10 m
speed of light = 299792.458 km/s
Requires MLton and/or Poly/ML.
make test # build + run the suite under MLton
make test-poly # run the suite under Poly/ML
make all-tests # both
make example # build + run the demo
make cleansmlpkg add github.com/sjqtentacles/sml-constants
smlpkg syncReference lib/github.com/sjqtentacles/sml-constants/constants.mlb from your
own .mlb (it pulls in the vendored sml-units first), or feed sources.mlb
to tools/polybuild (Poly/ML).
sml.pkg smlpkg manifest
Makefile MLton + Poly/ML targets
.github/workflows/ci.yml CI: MLton + Poly/ML (Variant A)
lib/github.com/sjqtentacles/
sml-units/ VENDORED verbatim from sml-units
units.sig units.sml units.mlb
sml-constants/
constants.sig CONSTANTS signature
constants.sml Constants implementation (built on Units)
sources.mlb loads ../sml-units/units.mlb, then constants
constants.mlb public basis
examples/
demo.sml constant tour + derived relations + conversions
test/
harness.sml shared assertion harness
test.sml exact values, derived relations, dimensions (44 checks)
entry.sml / main.sml
tools/polybuild Poly/ML build wrapper
44 deterministic checks: the five 2019-SI defining constants matched to full
Real.== precision; CODATA-2018 measured constants to a tight relative
tolerance; the derived relations hbar = h/(2 pi), R = N_A k, and
c = 1/sqrt(mu0 eps0) within epsilon; the runtime dimension of every constant
(e.g. c : m s^-1, eps_0 : kg^-1 m^-3 s^4 A^2); photonEnergy,
photonMomentum, and deBroglie values and dimensions; unit conversions; and
dimensional-safety checks (adding mismatched dimensions raises
Units.Dimension). Reals are compared with epsilon/relative tolerances and
printed with a forced-decimal/--sign formatter, so output is byte-identical
under both compilers. Run make all-tests to verify.
MIT. See LICENSE.