My head's spinning ;-( First, Cheryl pointed out that I should have said lineage-linked vs event-linked, rather than relation-based vs event based. If she was only insisting that I use conventional terminology, I would understand and accept the correction. But she may imply that these are not synomyms, but different concepts. If so, I'm still in the dark, such as the difference between a lineage link and a relation. Now Dennis says makes quite clear the difference between lineage-link and event-link database. Or so I thought until: > The alternative, used by TMG, is to create relationship events -- > one event for father relationship, one for mother relationship. In TMG, > every event ("tag") has room for two "principles", all the "other > people" are added into a witness list in the event. A "relationship event" seems to conflate the categories. An "event" (in the sciences, anyway) refers to a change of state. Yet the child-mother relation is the _result_ of a birth event, but in itself does not refer to a change of state. Also, who are the principles in the case of a person's death? To presume that the death was witnessed seems unwarranted. At some earlier point in this thread, there was a suggestion that event-link relates genealogy to historiography. I don't think that stands up well as a valid generalization. While there is a particular branch of history (narrative) that strings events together, that approach is rather in a shadow these days. Historians often lay emphasis on relations (such as class relations). There's also a possibility of representing history in terms of process in which there is no "state", but instead evenescent structures that constrain causal potencies, which are primary. All I'm suggesting is that I don't believe one can assume that historiography is simply or always based on events. -- Haines Brown, KB1GRM