We're abusing CombinatorialDerivation somewhat in its usage within a ProtocolInterface.
This is problematic when it comes to expressing things like allowed ranges of numerical values, since CombinatorialDerivation expresses only discrete finite sets, but regions of allowed values will often be continuous.
Likely, we should shift to sort of FactorSpace class that is similar, but allows us to specifically express allowed ranges. This sort of typing would be valuable for use with parameter values as well, and likely on into PAML.
We're abusing CombinatorialDerivation somewhat in its usage within a ProtocolInterface.
This is problematic when it comes to expressing things like allowed ranges of numerical values, since CombinatorialDerivation expresses only discrete finite sets, but regions of allowed values will often be continuous.
Likely, we should shift to sort of FactorSpace class that is similar, but allows us to specifically express allowed ranges. This sort of typing would be valuable for use with parameter values as well, and likely on into PAML.