Taxaonomic database structure

[Stan Blum]

At 11:02 AM 8/29/97 -0500, Jim Bennett wrote:
>I would like to ask the readership how they deal with binomials and
>trinomials in their taxonomic databases.

It is difficult to recommend a good "middle ground structure" without
knowing more about your situation.  For example:

What is this database intended to capture, and how is it going to be used?
Are you capturing literature references and supporting your own revisionary
work, or developing an authority (look-up) file for a collection (or
observation) database, or what?

Is the current list of uses likely to grow?

Which DBMS are you using?

How much effort can you expend in programming versus data entry and proofing?

How many people are doing the data entry?

>Up till now I have entered the
>entire binomial or trinomial in one field, with separate fields for family
>and genus.

What does a record in this table represent; a taxon, a species-level taxon,
a valid (accepted) species-level taxon, a specimen, an identification of a
specimen?  (There are a lot of possibilities!)  How many tables will taxa
appear in?

>In some more advanced databases
>it appears that each taxonomic rank is made into a separate record with a
>field for what the rank is. I don't need this level of sophistication for my
>databases, so am looking for a middle ground structure. Any thoughts on this
>would be most appreciated.

As Norm Johnson noted, a recursive (self joining) table for taxa is very
flexible.  It accepts taxa of any rank and status, unbalanced
classifications (e.g., incertae sedis species), etc., without requiring a
lot "work-arounds".  Unfortunately, it can also require a lot of
programming to do data retrieval and reporting (in some systems).

You should note, however, that the question "how to represent a
species-level name?" is a separate from "how to represent a
classification?" (= relationships among taxa).  It is possible to use a
single field to capture the full name of a taxon (of any rank and status)
in a single long list, and to create a classification from this list in one
or more separate tables.  Whether you add fields for rank, status,
abbreviation, parent taxon, TaxonID (numeric surrogate key), etc., will all
depend on your situation and objectives.

Finally, please note that the correct URL for the new "ASC model" is:

  http://gizmo.lbl.gov/DM_TOOLS/OPM/BCSL/BCSL.html

And, yes, it's still under revision (being simplified in several ways).


Stan Blum

- - - - - - - - - - - - - - - - - - - - - - - -
Co-Chair, ASC Computerization & Networking Comm.
  http://www.bishop.hawaii.org/asc-cnc/
Bishop Museum
Honolulu, HI
Tel. (808) 848-4173
Fax. (808) 847-8252