Should model.rules only be generated on the fly?

[rmtfleming]

Should model.rules only be generated on the fly within a function from model.grRules? It seems like people might get dependent on .rules, like they did with .rev as the former duplicates (although in an abbreviated way, the information in model.grRules.

[tpfau]

Id actually advocate it rather the other way around since there are no restrictions on the values of model.genes and thus a grRules field could have unparsable entries. At least to me, grRules is only a visualisation of model.rules and rules should have priority.

Model fields should never be modified manually but always through functions, as otherwise you never know, whether fields get out of sync.