Preprints of the Metadiversity Conference Proceedings

  Session 6: The Metadata Challenge for Museums

Museum Informatics: Where We Have Been and Where We Need to Go

JULIAN HUMPHRIES, Associate Professor/Research Department of Biological Sciences, University of New Orleans

Today I would like to address issues about the expectations of a directory service community—especially the issue of what those expectations ought to be from the viewpoint of the people who actually have to provide data and metadata about their collections. I am also going to be talking about who we are, what we do (and why what we do is important), the priorities of natural history museums from the viewpoint of content providers, and the nature of the problem for biological collection and the corresponding biodiversity data providers.

The Role of Natural History Collections

The traditional and primary role of natural history collections is as archivers and catalogers of specimens and proxies for specimens. The result is a curated collection of objects and data about those objects. The product that comes out of natural history collections is dynamic knowledge about those specimens in a lot of contexts— taxonomic, ecological, and functional descriptions of species, higher level taxa, and communities.

The people in museums who make these products are also frequently the same people who are responsible for the curation and cataloging of these collections. As a consequence, their role is twofold. First, they have the responsibility for seeing that the collections are well maintained and curated. But they also have the responsibility for taking the information from those objects and doing something with it in terms of publication.

Why Biodiversity Data are Important

The importance of biodiversity data is probably fairly obvious, but I want to make a few points about the role these three billion objects play in this scheme of discovery about the Earth, the history of the Earth, and what is going to happen to the Earth with time. Virtually most of what we know about the natural world in some way or another works its way back to the truth that is represented in natural history collections. So that original piece of information that someone got about a bird, fish, or insect—as well as where the specimen was collected—is the start of a long pathway of knowledge and dissemination of knowledge about our natural world.

It is also true that those same data, once used as an indication of the past, are also valuable for prediction about the future in terms of telling us about parts of the world with which we are unfamiliar, about things that we can predict, and about the location of taxa that we have not yet collected. The data also help measure the impact of particular human activities on the natural world, as well as the likely impact of a human activity on flora and fauna.

Aspects of a Museum Curator’s Job

I am speaking from the viewpoint of a former curator (and I actually remain curator of a small collection). If you look at the things that are important to me as a curator, the first thing is conservation of the specimens under my charge. It has been interesting to hear others relate the value of data to its age as they discuss the 20-year-rule. In contrast, our data don't even start to get interesting until they are more than 20 years old, and the older they are, the better they are. We think, literally, in terms of centuries—centuries past and centuries future. We are interested in data and specimens from a century or more ago. And we are interested in having those specimens that are under our charge be just as useful 100 or 200 years down the road.

That scope of time puts a very heavy load on us curators in terms of determining what to do first when we begin working with a new specimen. For the most part, the first thing we have to determine is that that specimen is going to be there tomorrow in the same condition in which we found it today.

The next aspect of being a content-provider curator is to understand that the specimens are to be used. Therefore, there must be in place tracking mechanisms that track the transactions associated with a specimen. Increasing specimen usage, contacting the researchers (both internally and externally) who might make use of the collections’ specimens, and satisfying the funding requirements of the parent organization or agency also are part of a curator’s responsibilities (in fact, they comprise a part for which the curators are not necessarily trained). At different times, this one position might be a service function, a training function, or even a research function.

Because of all this responsibility, data and data management—despite their importance—almost always comes at the end of the curator’s priority list. In addition, since data management was simply paper up until 15 years ago, the idea of devoting additional resources to data management meant something had to be taken away from the other responsibilities.

The Internet Impacts Curators

What has happened since the advent of the Internet has been just amazing and has signaled a truly revolutionary change for natural history collections. Before the Internet, paper itself was very, very important to us. We cared a lot about physical paper. We made labels out of it, we wrote our catalogs on it. It was something that we studied and investigated and used for research.

Then the digital age came upon us. Now we had to take the records that we had carefully curated and managed for a century or more and transfer them to computer technology. This raised many questions about how we manage our collections and what we were going to do with those data. These concerns became even more acute when microcomputers moved into the collections, and it became possible not only to digitize our data but also to actually put the data on machines that sat right next to the specimens themselves!

Originally what we saw was the creation of a large number of disparate and heterogeneous information systems that did not do a very good job of talking to each other. Remember, at the time there really was no organization that dealt with issues of standardization in our museums. In fact, one of the first efforts in that regard took place just 10 years ago, in 1988, at a workshop organized and funded by the Association of Systematic Collections. At that workshop, one of our charges was to deal with issues related to standardization of hardware and software, necessary primarily because the era of the PC was anything but a standard at that time. Intel was certainly the primary vendor, but there were other, competing operating systems. There also were many software choices, and these software choices varied dramatically in their capabilities and in their interchange with other software choices.

A number of workshops followed. For example, in 1993, I helped organize a workshop on standards and the exchange of data. And Berkeley personnel organized a workshop on interoperability of mechanical databases. And all the while we talked about federation, federation, federation.

During this time, the data management systems that were being created got increasingly sophisticated. We moved from a written set of applications to a more sophisticated set of applications. But within the community, we still lacked any kind of high-level effort that would make it possible for all of these databases to speak to each other. So, the challenge was—and remains—to determine what is necessary for all of these biological data sets and the data associated with them to be readily accessible and available, and for the information about these data sets to be readily accessible and available in all of the variety of clearing houses and search engines and directory services that exist today.

Curators Consider Collection-Level and Object-Level Metadata

Earlier a distinction was made between collection-level metadata and object-level metadata. Collection-level metadata might involve the creation of 50,000 to 100,000 record objects to capture the data about all the places where museum information is stored. However, when we are talking about object-level metadata, we are looking at a number at least five or six orders of magnitude larger!

There is no easy way to move from good descriptions of our collection data sets to good descriptions of our collection objects. So if we take the numbers discussed earlier in terms of the cost per record—up to $60 for each record—we realize that more detailed metadata-descriptor creation would cost tens of billions of dollars. This is not a likely investment in the near future. On the other hand, when we are talking about collection-level metadata, the cost falls to the millions of dollars. I believe that this is a feasible goal for our community, and I believe that it is one that—with the proper incentives and resources—is attainable.

Previous

Copyright © 2003 NFAIS. All rights reserved. No part of this product or service may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written consent.

Privacy Policy