Specimen databases

[Matt Yoder]

some rants:

        At TAMUIC we've used MSAccess97 in the last couple of years to build
our curatorial and unique specimen data capture (barcode) systems.
We've also linked an image database and URL database to our taxonomic
authority files, all relatively easiliy done once the learning curve is
climbed.  It has allowed us to integrate collaborators data relatively
quickly, placing new names and data into a relational system with the
use of a dataset/name analyzer we have developed.  USGS gazeteers are
tracked and integrated at various levels etc, data can linked to
ArcView/Info etc... All told we manage around 50-60 tables with Access.

        Data from Access can be quickly manipulated and exported in a large
variety of ways, for example _all_ data for the TIARA project-
http://csdl.tamu.edu/TIARA/, is managed in Access. It is customized and
exported from Access to an indexing progam (MG) on a unix based platform
and from here (with much code) it is served to the web (this is the the
two part burdern Una speaks about, MG/cgi scripts combined with MS
Access).

        If you are using your own custom setup (ie Filemaker, Access, Paradox)
I don't believe the linkage/cross platform issues should be of much
concern (i.e. use whatever modern db software you want, if fields/data
structures are set up correctly you can export/link to other
platforms/software extremely easily.  Data can be imported and exported
to a wide range of formats from these software packages by advanced
users (see DeltaAccess).

        One note of warning.  When thinking about entering into a project like
this think time and $.  Once you have a system of this size in place you
will need to have someone _maintain_ it, especially if you build it
yourself (this is perhaps a good reason to use something like BIOTA or
Specify which you don't have to maintain from a programming
standpoint).  The intial developer of your custom project will have
intimate knowledge of the system and few other people will.  If
development is done by a grad student (who went through a 6 mnth
learning curve) and that student leaves you basically need to fund
another 6 mnth curve before you can continue development.  Basically I
think that in the long run you need a computer techie type that will do
nothing but manage these hardware/software systems (and teach others how
to use the system) _if_ you want to do your own custom thing.

cheers,
Matt

Matt Yoder
Dept. of Entomology
Texas A&M University
College Station, TX
77843-2475