[Babase] AMOK Questions
Susan Alberts
alberts at duke.edu
Mon Jul 14 18:11:21 EDT 2008
Thanks for the reply Karl, see below.
Karl and I talked on the phone about the problem that we think AMO =
BIL. He said, basically, that if we think they are the same individual
then we should just have one row. If we think they might be different
individuals then we should not try to have the same row for them. This
makes sense.
Problem is, "sure" is a tricky word. We are as sure as one can be with
genetics that BET and NAC are his parents. We are not 100% sure that
he is the same individual as BIL. These are different things. He is
probably, but not certainly, the same as BIL. So, for instance, we
might feel comfortable saying that all of BIL's group-level attributes
(density and group composition at BIL's birth, for instance) apply to
AMO, and we might be happy with some individual-level attributes too,
such as birthdate. We might not really want to use BIL's parity
though, for AMO, depending on how sure we were about AMO = BIL.
GIven that we are sort of waffly on the "sure" part, I like Karl's
suggestion:
>> ......
>> I could see a table called something like
> ALTERNATE_SNAMES. It'd have, say, 5 rows:
>
> Asnameid Integer used for a row id.
> Sname The real sname of the individual
> Nick The alternate sname of the individual
> Date The date the alternate sname was assigned
> Notes (text) As much comment as you'd like about why the
> alternate sname exists.
>
> That way when you've got data outside Babase with the "bad" sname
> in it you can still work with the data and link it to the "good"
> sname.
>
> "Inside" babase I'd still prefer that 1 baboon means 1 row in
> BIOGRAPH. IMO moving away from that is going to cause real
> headaches.
>
Jeanne, thoughts? Others?
Susan
--------------------------------------------------------
Susan Alberts, Dept of Biology, Duke University, Box 90338, Durham NC
27708, 919-660-7272 (Ph), 919-660-7293 (Fax)
More information about the Babase
mailing list