Data Entry => Question of Priorities and Resources

[christian thompson]

Yes, I can't dis-agree except on basis of limited resources.

Unfortunately, in Entomology, at least, the resources are very limited
and the question become whether to waste them on RETROSPECTIVE specimen
data capture of dubiously identified material or spend those resources
on research to produce HIGH QUALITY information.

Unfortunately I have worked too long at the Smithsonian and seen many
resources lost to Automation projects that eventually failure to deliver
any useful information. Yet ADP people continue to spend more on
software, etc., than, for example, the whole research budget of
Entomology.

If we had ALL the resources we needed, I too would like to have one
retrospective capture the data on ALL our specimens so that when the
researcher* got to them they would not have to do the data entry work
themselves. But the experience of the last thirty years is that we have
lost about half of our research positions and have seen a great
expansion of ADP which didn't exist 30 years ago. Add that to the fact
that in Entomology we know less than half the species (or if you accept
Terry Erwin's estimate, less than one percent), the priorities should be
reverse. Unfortunately the lure of automation continues to distract
Administrators from facing the high cost** of basic research. Just
Automate our specimen data and you will have all the biodiversity
information you ever needed! Sounds good to me. And just think, if the
trend continues for another 30 years, you won't have any specialists
left to tell you whether things are correctly identified or not.

*It should also be noted that sometimes, ONLY the researcher can
understand what the cryptic labels on old specimens really means.

** The Administrators are also told that Automation is cheap, that, for
example, you can get volunteers for free to do your specimen label data
capture. Or even scan your type specimens! And after those are scanned,
who needs researchers any longer as any one will able to just look at
the pictures.

Oh, well ... I will say no more as I need to describe at least one
species today if I am ever going to get the Costa Rican flower flies
done!


F. Christian Thompson
Systematic Entomology Lab., USDA
c/o Smithsonian Institution
MRC-0169 NHB
PO Box 37012
Washington, DC 20013-7012
(202) 382-1800 voice
(202) 786-9422 FAX
cthompso at sel.barc.usda.gov e-mail
www.diptera.org  web site

>>> "Thiers, Barbara" <bthiers at NYBG.ORG> 01/08/03 09:17AM >>>
 I completely agree that quality taxonomic judgements are key to the
usefulness of any specimen databasing endeavor, and that the importance
of
such judgements, and the level of expertise required to render them is
often
overlooked as a component of specimen cataloging projects, and
biodiversity
inventories as a whole.

However, I don't subscribe to the idea that databasing should be
deferred
until one is certain that all specimens are determined and curated
correctly.  Specimen determinations are hypotheses, and organizing
these
hypotheses into a searchable and plottable format should be seen as a
tool
to improve taxonomy, not a definitive answer to what grows/lives
where.
Specimen databases become more authoritative over time as users of
these
data subject them to scrutiny, just as they do with data they acquire
from
other sources, and then report the results of their analyses.We receive
many
updates and corrections to our specimen data every week, and thus our
data
are improved through use.

There are some ways to target specimens whose determination is most
likely
to be suspect.  We have developed a profile for specimens most likely
to be
misdetermined:  they lack complete data (e.g., lacking an authority, or
date
or a collector name or number, or only a vague locality), they were
collected and/or determined by unknown people (i.e., those not already
represented in the database); and they have not been annotations by an
expert in the taxon or the floristic region represented.  We database
such
specimens anyway -- sometimes they turn out to be types!  However, we
can
pass along this profile to users who need help with their
interpretation of
our data.

As to the question posed in the original post, two bottlenecks that we
encounter are nomenclature (what is the correct name to use, when the
taxonomy of a group is currently in flux?) and most importantly,
georeferencing.  In a herbarium like NYBG (and BM) with many
historical
collections, the proportion of specimens with geocoordinates is small.
Other bottlenecks are deciphering handwriting and incomplete names of
people
-- e.g., when a collector or an author is given just as "Smith"
obviously
additional work has to be done to figure out which Smith is meant (or
to
determine that the data are insufficient to figure it out).  Extensive
authority files help with these bottlenecks, and as data in these
authority
files accrue, the bottlenecks are lessened.  Constant editing of
authority
files is key, however.

-----Original Message-----
From: Daniel Janzen
To: TAXACOM at USOBI.ORG
Sent: 1/8/03 10:14 AM
Subject: Fwd: Data Entry Query

Fran, the time/cost bottleneck that dwarfs all other costs so much as
to
make them almost irrelevant is that of the time/cost of quality
taxonomic
action by taxonomists revising/curating the collection to where you
KNOW
that the specimens being databased are what you say they are
(including
that when two different specimens are both called species X, whatever
is
species X, those two specimens are conspecific) - at least for
retroactive
data capture of insect collections from the tropics.  Data capture for
incoming specimens is a totally different matter and needs to be
treated
as
such.

Not meaning to be disrespectful, and a non-taxonomist's observation.

Dan Janzen



>X-Sender: frank2 at mailserver.nhm.ac.uk
>X-Spam-Score: 1.6 (+)
>X-Scanner: exiscan for exim4 (http://duncanthrax.net/exiscan/)
>           *18WECS-0005BG-00*6qheYLXxghg*
>Date:         Wed, 8 Jan 2003 11:10:33 +0000
>Reply-To: Fran Kern <F.Kern at NHM.AC.UK>
>Sender: Taxacom Discussion List <TAXACOM at USOBI.ORG>
>From: Fran Kern <F.Kern at NHM.AC.UK>
>Subject:      Data Entry Query
>To: TAXACOM at USOBI.ORG
>Status:
>
>Hello.
>
>I'm currently working on a project to identify and quantify time/cost
>bottlenecks in data-basing natural history collections.
>
>Would anyone know of similar projects happening elsewhere?  Or be
able
to
>point me in the direction of relevant articles,  journals,  or
meetings?
>
>Thank you for your help,
>Fran
>
>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>Francisca Kern
>Rapid Data Entry Assessment  Project,  Department of Botany
>The Natural History Museum,  Cromwell Road,  London SW7 5BD
>