RIL Summary storing seed counts as stockprop vs phenotypes

[carolyncaron]

During a meeting about cvterms, we discussed the practicality of storing RIL seed counts for each generation, seed colour + cotyledon colour, and specifically the "RIL_complete" stock property that is used to store total seeds for a "completed" RIL population (which we believe is always in flux so not sure on its feasibility). It feels more intuitive to store seed counts as phenotypes rather than properties, but some benefits were mentioned if we maintain them as properties:

• The ability to add this information by other people than the curator, eg. it could be a breeder, or someone who manages seed at the field lab and thus doesn't need to be trained on how to use any of the phenotypes importers
• Less work is involved to change the cvterms used rather than migrating all existing data to phenotype storage
• Every new import as a phenotype would not overwrite existing values and instead be added to a list of values, which can cause confusion

For properties such as seed colour and cotyledon colour, we also don't want to emphasize the variability that likely exists in the dataset, as often this property is regarded as the expected phenotype of the RIL or germplasm/population in question.

For RIL_complete, is it necessary to keep this at all? It was originally intended to be used as a boolean. Do we want to return it to its original intent, or is there argument that we don't ultimately know if a RIL is even complete? Perhaps a name such as RIL_under_active_development is more meaningful.