Identification

All instances of Item Types are uniquely identified using the ordered compound key in baseDirectory/Settings/Identification.csv. The file is required and must contain at least one row. Each property is included in every root item type and inherited by its descendants.

The columns of the Identification.csv file are the same as the columns in the Item Types properties CSV.

An optional baseDirectory/Settings/Identification.Mixin.csv file can be used to add additional identification properties. Those mixin properties are also part of the compound key and every generated reference.

COGS 2 identification fields are deliberately narrow. Every field must use string or anyURI, must be required and singular (1..1), must not be ordered or allow subtypes, and must have a unique name in the effective item property set. Every item and reference must supply a nonempty lexical value for every field. The logical key is the concrete item type followed by all values from both files in declaration order. URI strings are compared as serialized; COGS does not normalize them before reference lookup.

JSON references contain only $type and all identification properties. XML references contain all identification elements followed by TypeOfObject. Forward and repeated references to the same key resolve to the same object inside an item container.

The DcTerms macro is not permitted in either identification file.