[
Lists Home |
Date Index |
Thread Index
]
John Cowan wrote:
> "Jonathan Borden" <jonathan@openhealth.org> writes:
>
> > Number of living parents,
> > grandparents,
>
> This works only if we confine ourselves to either legal or biological
> parents. Arguably, there are living persons with five parents: an
adoptive
> mother, an adoptive father, a sperm donor, an egg donor, and a womb donor.
Ah yes. Let's assume "legal parents" in which case Ari's insistence that
this would be better modelled as (mother,father) blows up in the case of an
adoptee of a same sex marriage :-) The point is that the requirement that
the individuals (e.g. elements) be observably/nameably distinct imposes
certain restrictions on what one might want to model, or to say.
>
> > Digits on left hand
>
> But then the temptation would be to set the max to five, and blow up on
> polydactyly.
>
> > Human chromosomes...
>
> Bare X, XXY, XYY.
>
All these "blow up" *only* when we make incorrect assumptions about the
values of min/max -- e.g. digits on the left hand might be {0,8} and
chromosomes might be {45,50} -- or whatever limits *one wishes to set*, the
point being that there are real world modelling tasks where cardinalities of
(0,1,many) is not the most crisp way of modelling the problem.
I am merely pointing out that there are indeed real world situations where a
cardinality of other than (0,1,many) is hardly evidence of "poor design".
Jonathan
|
