Lineage

Formats like XML provide an automatic way of depicting a top-down hierarchical relationship. Unfortunately, genealogical lineage is more of a 'network' than a pure 'hierarchy', as many researchers will appreciate, so a simple nesting of "offspring" under their associated "parents" is insufficient. In fact, to be strictly accurate, genealogical lineage is really a “directed acyclic graph” (DAG) since there cannot be any loops, e.g. a child being their own grandparent etc. In contrast to lineage-based genealogy, micro-history associations between Persons, Places, Animals, Groups, and Events are a lot more general and do form an arbitrary network.

There's also a problem with any top-down approach (where biological parents point to children) unless a specific union between two people has a single representation in the data, but that then causes further problems with the nature and the lifetime of that union. In fact, the only universal events for any person are their birth and death. All the others, such as baptism, christening, marriage (civil & religious), divorce, burial, cremation, etc., are culturally dependent and so cannot be expected to be present. Multiple concurrent marriages may even be legal in some cultures. The alternative bottom-up approach (where children point to their respective parents) is the only practical representation.

Each person has just one progenitive father and one progenitive mother[1] and so the associated Person entity can have upward links to its appropriate two parents (where known) using <FatherPersonLnk> and <MotherPersonLnk> elements. These links are permanently valid (i.e. have no date-limited applicability) and form the basis of simple genealogical lineage. They do not define any sort of ‘family unit’ and this is where representations that quietly infer the concept of a family unit from the genealogical data would fall down. See Happy Families.

STEMMA can represent additional types of parentage such as guardians, foster parents, adopted parents, etc., that are designed to support the identification of family units. These are valid over specific periods of time since family circumstances can change and so those time periods are determined by Event entities. Furthermore, because micro-history can sometimes be even more complicated than that, the narrative feature associated with all top-level entities will allow arbitrary person-to-person associations.

A typically difficult situation to handle would be a death notice that mentions a mourning grandson by name but gives no indication of which side of the family he is associated with. Such a Person could still be represented in STEMMA and associated with the relevant grandparent through the narrative support.