Re: Genealogical evidence and data model

Wes Groleau <groleau+news@[EMAIL PROTECTED]
> writes:

> Haines Brown wrote:

>> relation. True, but is that the issue? Can't I convert the
>> grandfather-grandchild relation into a series of two father-child
>> relations, each of which does have existence in genealogical format?
>
> Yes, but if the intermediate father is potentially one of
> two or more in your database, what then?

Wes, sorry to be so slow about this. When there is a father-child
relation (in any database), aren't we always identifying specific
persons? That is, child is Child A because we always uniquely
distinguished that child from others. Now, Child A has a father, which
we also always specifally identify as Father A. So if we trace back from
child to parent, it is always from Child A to Father A. Are we not
always identifying specific people?

Your example of ambiguity, if I recall, involved children having the
same name and other common identifiers. Suppose we have two people with
the same name and share all the features that are used to distinguish
that child from others. Despite all that similarity, when the child is
assigned a record in the database, the fact that there may be another
record that is entirely similar should be irrelevant, for the record is
somehow indexed uniquely. For example, in GEDCOM, each INDI is assigned
a number automatically, and so two identical records can't be mixed up
(or so it seems to me).

-- 
 
       Haines Brown, KB1GRM