Thanks for the comments and great examples.
You raise good points and I should first point out
that this module is badly in need of a complete rewrite.
Both the code and the data model are seriously outdated.
The original scope of this was to try to represent
and visualize some subset of networks of familial
relations. One intentional limitation was to
focus on handling a current point in time -- no
notion of ex-anything -- as a way to manage the scope
and complexity and also to distinguish this effort
from that of e.g. genealogy (which I see as primarily
concerned with history).
The data model basically accommodates:
1. pairs of people being in relationships
2. each person having up to two parents
The two items above are independent, and I used "spouses"/
"marriage" for the first, and "mom" and "dad" for the second
(without defining those terms or assigning genders or even
defining "gender").
Regardless, the data model above is definitely limiting and
I'd love to see it expanded.
Show quoted text> There are 2 types of spouses: current and ex.
Only "current" is currently represented. I think adding
"ex" would involve a notion of 'life events' (e.g. the GEDCOM
format supports this)
Show quoted text> There are at least 4 types of parents and siblings: biological, step, adoptive, and foster.
> There is the additional singling type: half.
For the derived relationships, it would be great to have the
ability to infer them (not just siblings and half-siblings but
also things like "third cousin twice removed").
For the ones that aren't derived, it would be cool to add new types
of parental relationships. This would involve removing the second
aspect of the data model.
Show quoted text> Also, how do you differentiate between the relationship types like married, life partner, affair, etc?
Only one committed relationship can be represented at a time.
Yes, it would be nice to expand that (i.e. remove the the first
constraining aspect of the data model).
Show quoted text> Examples I made up off the top of my head just now:
>
>
> Marge is married to Anne. Anne gets artificially inseminated by Bill. The resulting child is Steve. Steve is the biological son of Anne and Bill who are not married, the adoptive child of Marge who is married to Anne.
>
Right, representing this case would currently be limited until there's a way
to have more than two parents per kid.
Show quoted text> There could also be poor Linda who is in the system with a string of foster placements. She has her biological parents but also a string of foster parents. She does not find a forever home before turning 18 when she is now out of the system.
>
Yes, representing this would involve a notion of 'events' and
taking on the issue of history.
Show quoted text> I am trying to build a fictional family tree where I only know X is a descendant of Y and Z, and where there are alternate reality relationships with alternate reality children. It is a mess.
>
Yes, sure is!
The test directory has a number of situations I wanted to accommodate,
maybe you could add some tests there illustratingi how the interface
should look, and how they should be visualized? Even if this module is
rewritten, I think parts of the test suite may be salvageable.
Show quoted text> Would you think about adding other types of relationships?
Sure! I'd love to see this module expanded and modernized.
Brian
