updating (Re: Catalog vs curate)

[Doug Yanega]

Peter Rauch writes:
>One of the troubling issues in creating systems to manage online
>catalogs is how those systems are designed (or not) to deal with
>updates/corrections/alternative opinions. In particular, how do the
>systems journal the history of corrections/etc? Most, I believe,
>don't.  Most, I believe, simply replace an erroneous/outdated/earlier
>opinion with the current thought, and irretrievably discard the older
>data (except for system backup archives, which are not the proper place
>to journal transactions). (I'd very much like to be contradicted in
>my opinion that most collections databases don't do very extensive, and
>in many cases do almost no, transaction journalling.)

Well, perhaps a cataloguing database I worked on will serve as a nice
contradictory example: the INHS Plecoptera database is set up in FileMaker,
which is very good for things with a lot of text. We have one field set
aside for the VERBATIM entry of the labels contained in a vial, but also
have completely separate fields for all of the possible bits of data, such
as date, nearest town, stream, state, county, etc. - and it turned out to
be quite common when one or more of the things written on the label was in
error, or missing (especially county listings). For example, a really bad
one might be "Kentucky: 3 mi W Peterville, Oconnee River, 31 June 1953,
T.H.F." when it turns out the actual locality was across the border in
Jackson county, Tennessee, the town's name is Petersville, the river is
listed as the "Little Oconee", there is no June 31st, and T. H. Frison
collected there in 1935 (believe me, we had errors like those and worse to
filter out). What we'd do then was leave all that erroneous data in the
"label" field, and in the individual fields, we'd enter the correct data,
putting anything that represents added/corrected information in
parentheses. That way, searches will turn up the records when they're
supposed to, but all of the original information is also still intact, and
the changes are made obvious. Similarly, we have separate fields for
"species" and "identifications", with the latter including ALL of the names
that have been put on that particular material, when, and by whom. It is
therefore easy to track the "history of corrections" - and for anything
complex, there is a field set aside for "remarks" as a catch-all. For an
inventory, you *have* to keep track of such things, though if all you were
after was a biogeographic database, for example, one might leave all the
faulty data out. A lot will depend on whether you might ever need to
retrieve that material...but in any event, you should get all the data you
might need entered the FIRST time around. Databases that do not include ALL
the data, like gender, detailed locality, host, and so forth, will have to
be completely reworked somewhere down the road, and that is also wasteful.
Sincerely,

Doug Yanega       Illinois Natural History Survey, 607 E. Peabody Dr.
Champaign, IL 61820 USA      phone (217) 244-6817, fax (217) 333-4949
 affiliate, Univ. of Illinois Urbana-Champaign, Dept. of Entomology
  "There are some enterprises in which a careful disorderliness
        is the true method" - Herman Melville, Moby Dick, Chap. 82