Discover Life in America

Christian Thompson - 9 July, 1999

Re: Barcodes on insect specimens

Date: Fri, 09 Jul 1999 11:10:51 -0400
From: "christian thompson" <cthompson@sel.barc.usda.gov>
To: ashe@falcon.cc.ukans.edu, pick@pick.uga.edu
Cc: faulzeitler@ascoll.org, scottm@bishop.bishop.hawaii.org,
        msharkey@byron.ca.uky.edu, sackley@compuserve.com,
        jugalde@euclea.inbio.ac.cr, longinoj@evergreen.edu,
        christine.deal@intermec.com, ksem@kuhub.cc.ukans.edu,
        brianb@mizar.usc.edu, FURTH.DAVID@NMNH.SI.EDU, becky_nichols@nps.gov,
        Chuck_Parker@nps.gov, Johnson.2@osu.edu, mkaspari@ou.edu,
        djanzen@sas.upenn.edu, whallwac@sas.upenn.edu, windsord@tivoli.si.edu,
        colwell@uconnvm.uconn.edu, pin93001@uconnvm.uconn.edu
Subject: Re: Barcodes on insect specimens

Steve as usual makes some critical points. Thanks.

To re-cap: What is important is the UNIQUE specimen identifier and that
should be human-readable. And this is the lasting part of the ECN standard
we established years ago. That is, the UNIQUE identifier should be an
ALPHA identifier for the collection (most follow the Ross Arnett's CODENs)
followed by numeric sequential number.  Our is USNM ENT 99999999 for
United States National Museum Entomology etc.

In handling large numbers of specimens, barcodes are great as they can be
scanned, etc. But they also can be used as if they were a plain old
printed label. And as others have noted, once the data record is finalized
with a species identification, most specimens will not be re-scanned.

My problem with my Code 128 is how little it can encode. Two characters
for a UNIQUE collection identifier is too few. Most museums use 4 or more
letters to identify themselves and then in large museums more characters
are needed for the department. These identifiers are critical if some day
we are ever to build a truly interoperative global database of specimen
label data. Yes, I do realize one can encode many more things into two
characters if one uses different bases, etc, but that departs for the
simple human understandable museum codes we have standardized on today.

The other thing that is not clear from Steve's message is whether these
Code 128 barcodes ALSO have the information printed in ALPHANUMERIC
characters so one can read them, etc.

Maybe in the future, CODE 49 will not be able to be scanned, but what is
really important is that people will 1) recognized that any specimen with
such a barcode means that there is likely a computer record somewhere with
specimen label data. AND if we have followed our ECN standards, they will
be able to identify the organization that holds those computer records
from the unique alpha identifier. It is the specimen and the computer
record associated with it that ultimately is the most important. The real
challange is not what symbology we used today, but data capturing the
specimen label data, then perserving the specimen, the computer record
and the LINK between them for the future.

F. Christian Thompson
Systematic Entomology Lab., USDA
Smithsonian Institution
Washington, D. C. 20560
(202) 382-1800 voice
(202) 786-9422 FAX
cthompso@sel.barc.usda.gov

>>> "James S. Ashe" <ashe@falcon.cc.ukans.edu> 07/09/99 09:53AM >>>
Hi John and Everyone,

    I think the series of e-mails from John and others regarding the
eventual loss of Code 49 illustrate an important point about bar codes
and their use in entomology collections - they represent a technological
solution to a persistent data entry problem, and all technology is
ephemeral.  We must recognize that ALL barcodes will become as archaic
and unreadable as the stacks of computer cards that were
state-of-the-art technology in the 1960's.  In the long run, we must
reach peace with this reality, and keep the role that barcodes play in a
collection in perspective.  We often talk about barcodes in entomology
in a kind of mystical way that gives the barcode itself some kind of
special significance.  In fact, the barcode is only a representation of
a number, or a group of letters and a number, that act as an individual
identifier for a specimen.  The barcode itself only allows us to enter
that identifier into a computer more quickly and accurately than can be
done by hand-typing.  However, because the identifier (number) is also
printed on the barcode label, that specimen identifier is still useful
(though more combersome to use) even after the technology of the barcode
has been left behind.  I mention this in particular because John
commented in a previous e-mail about the possible need to re-barcode
millions of specimens if Code 49 barcodes cannot be supported in the
future.  I think this places undue significance on the specific kind of
barcode chosen - any effort to maintain a single barcode standard will
ultimately be foiled by changing technology.

    I also tthink that it is important to respond to John's statements
about Code 128.  In particular, he mentioned that Code 128 barcodes do
not have sufficient information content to make them useful.  I do not
know what kind of information that John wants to include in his specimen
identifiers; however, our Code 128 barcodes include 7 digits and 2
letters that identify the collection - in a barcode label size that is
smaller than most of our specimen labels.  We think that we are unlikely
to computerize more than 9,999,999 specimens before barcodes (or at
least Code 128) are replaced by other technology.  Thus, we are
confident that Code 128 labels have sufficient information for our
collection needs. We experimented with numerous barcodes (including Code
49) before choosing to go with Code 128.   Ultimately, we chose Code 128
because it is single-stacked and lightning-fast to read - and always
reads correctly.  We place them face-up as the bottom label and read
them from above at a slight angle - one only needs less than a
millimeter of bar-code visible in order to read them.  Since speed and
accuracy of data entry for specimen identifiers are the only reasons
that we use barcodes, we think that Code 128 barcodes serve our purposes
very well.  However, in this regard, it is important not to get into a
kind of bar-code chauvanism - a "my code is better than your code"
contraversy.  This places much more significance on the barcode itself
than it deserves.  In fact, any good barcode reader can read any barcode
- it simply needs to be recalibrated for each one (a process that takes
only a couple of munutes).  So we san still exchange specimens and read
each other's barcodes even if we choose different barcode standards.  It
is also important to remember that the database of information about the
specimens is our goal, and the database does not contain barcodes - it
only contains the specimen identifers.

    My best wishes to all.

Steve Ashe



Discover Life in America | Science | Unique Identifiers & Barcodes | Correspondence | Christian Thompson - 9 July, 1999