SYSML21 — astronomical unit defined as 1.5e22 m (factor 1e11 too large); electronvolt uses CODATA 2006 value

In the Quantities and Units Domain Library, file SI.sysml (SysML-v2-Release commit fb97b754, 2026-09), line 102:

attribute <ua> 'astronomical unit' : LengthUnit { :>> unitConversion: ConversionByConvention { :>> referenceUnit = m; :>> conversionFactor = 149597870691e11; :>> isExact = false; } }

The literal 149597870691e11 is 1.49597870691e22 m. The digits are the pre-2012 value 1.495 978 706 91 x 10^11 m with the decimal point lost, so the unit is a factor 10^11 too large. The direction of conversionFactor is fixed by MeasurementReferences.sysml line 346 ("kilometre for conversion factor 1000 with reference unit metre") and by the neighbouring declarations (minute: referenceUnit = s, conversionFactor = 60). The same value is already present in the 2022-01 release.

Proposed resolution: conversionFactor = 149597870700; isExact = true. The SI Brochure, 9th edition (V4.01, June 2026, https://doi.org/10.59161/AUEZ1291), Table 8, gives "1 au = 149 597 870 700 m" with the note "Defined by Resolution B2 of the XXVIII General Assembly of the International Astronomical Union in 2012". It also gives the unit symbol as "au"; the library uses "ua".

Related, same block of SI.sysml ("Units recognized in SI as specified in ISO 80000-1:2009"), line 95:

attribute <eV> electronvolt : EnergyUnit { ... referenceUnit = J; :>> conversionFactor = 1.602176487e-19; :>> isExact = false; }

This is the CODATA 2006 value 1.602 176 487(40) e-19. The same Table 8 gives "1 eV = 1.602 176 634 x 10^-19 J" and notes that since the elementary charge was fixed (Resolution 1 of the 26th CGPM, 2018) "this conversion factor has been exact". Proposed: conversionFactor = 1.602176634e-19; isExact = true. The adjacent dalton (line 94) already carries its CODATA 2018 value.

See also SYSML21-405 (tonne, line 99 of the same block, factor 1.0e-3 instead of 1.0e3) and SYSML21-320 (degree); lines 105 and 107, 'minute (angle)' and 'second (angle)', have the same 7-digit precision problem as degree and could be resolved together with SYSML21-320 as pi/10800 and pi/648000.