{gerber, svggeometrymodel, svgstructuremodel, svgpresentationmodel,}
{gerberplot, outmostsvgelement,}
{gerberpoint, point, point3d,}
{xcoordinate, x,}
{ycoordinate, y,}
{zcoordinate, z,}
{coordinatevalue,}
{millimeter, svglengthinmillimetre,}
{inch, svglengthininch,}
{linearinterpolation, line,}
{fromposition, startpoint,}
{toposition, endpoint, drawtopoint,}
{circularinterpolation, ellipticalarccurve,}
{parabolicinterpolation, quadraticbeziersegment,}
{cubicinterpolation, cubicbeziersegment,}
{throughpoints, controlledparameters,}
The CCUS suffers gigantic quadratic explosion due to the abundance of type_hlcs, domain_rules_hlcs and non_nulls. So I decided to run this experiment with these three HLC-classes disabled (i.e. with type_hlc-optional.filc, type_hlc+non_null-attr.filc, non_null-exists.filc, domain_rules_hlc.filc and propagate_from_defined_type.filc excluded from the case-library).
| Mode | E | -T | -U | -S | -H | -R | M | X |
| HLCs extracted from first model | 92 | 92 | 92 | 92 | 0 | 92 | 92 | 92 |
| HLCs extracted from second model | 135 | 135 | 135 | 135 | 0 | 135 | 135 | 135 |
| Comparisons | 90755 | 71909 | 90401 | 75003 | 18844 | 31276 | 31976 | 90755 |
| HLC-comparisons | 6742 | 6742 | 6742 | 6742 | 0 | 1688 | 1732 | 6742 |
| Correspondences | 33 | 6 | 33 | 31 | 27 | 30 | 33 | 33 |
| HLC-correspondences | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Differences | 3541 | 3451 | 3523 | 3522 | 86 | 974 | 1036 | 3541 |
| Number of FC-thesaurus lookups | 43144 | 0 | 42891 | 31298 | 2 | 5212 | 5648 | 43144 |
| Number of name-thesaurus lookups | 79343 | 0 | 79339 | 64040 | 18721 | 28782 | 28808 | 79343 |
Detailed results suppressed because the models are private.