add model-specific tests to comparisons? #6
Description
The new Simple_Lorenzen model discussed in #5 includes quantities related to M at age and Dynamic B0 under derived quantities. A proper test of these optional outputs would look for changes to the uncertainty in these quantities across model versions. These things are only calculated as output so even if they break, there should be no change in the other model results (unless the model crashes).
If the current comparison test still depends on a small set of quantities to compare across all models (I haven't looked lately), it doesn't make sense to test quantities which aren't output by some models. But making the test more flexible to allow model-specific comparisons could be a good way to ensure the features of interest for each model are getting tested appropriately.