GeoDIL

Another Node On the internet

GeoDIL

Dexter Perkins and Joseph Hartman

Department of Geology

The University of North Dakota

dexter_perkins@und.nodak.edu

The geosciences are visually based. Geoscience instructors have found that students are excited by, and learn well from, fieldtrips, movies, and slides - because they allow them to see, and maybe feel, a science in a way special to geology. Without imagery, geoscience topics are abstract. With imagery, geology comes alive, reflecting the world we live in. Volcanoes, minerals, fossils, planets, and many other subjects are especially photogenic. The best-selling college-level introductory geology text contains more than 500 beautiful color images (not counting diagrams and line drawings). Authors and text book publishers know that „cool‰ pictures interest students.

We have for many years used slides and overheads in our geology classes. Recently we have prepared Power Point presentations and tried putting slide collections on the Web, and we have prepared Power Point presentations. These approaches have been valuable, but they are limited. They are laborious to prepare, and once prepared they are not easily modified. The Web offers many new opportunities for bringing technology to the classroom, and we have thought about developing a Web-based geoscience photo collection for some time. Our idea is to have a collection that covers all the main topics in introductory geoscience courses, that is easily accessible, and that can be mined to produce virtual carousels for instructional use. The technical problems are many, but after much dithering, we have begun work on GeoDIL, the Geoscience Digital Image Library.

The Internet and Web-based teaching resources may soon become integral parts of teaching and learning. Today, there seems to be a rush to develop high-tech replacements for traditional teaching approaches, some of which have been successful. Yet, we still have much to work out and some widespread problems persist. One problem is that finding resources can be difficult because there are no comprehensive digital libraries (although some pioneering efforts have been begun). There are, however, two other problems that are specific to specific Web sites. First, developing good digital resources is time consuming and complicated. Consequently, much of what is available is superficial or limited. Second, some web resources are too specialized, or too complicated to use. They lack necessary "front ends" to make them accessible, and they are too much for most students (and faculty) to figure out and use.

Our concerns about inadequate coverage and ease of use guided us as we developed our plans for GeoDIL. We have, on the one hand, tried to keep our comprehensive focus and on the other tried to keep GeoDIL simple. It only includes photographs (and supporting data) - no other media ˆ and all photos are relevant to the Earth Sciences. At the same time, we have tried to make the search and retrieval techniques as holistic and flexible as possible. Our goal is to build GeoDIL for use by K-16 educators, students, and researchers. To be successful, our library must contain a large number of useful and exciting photos; it must also be easy to use. Our long term goal is to have 10,000 images in our library after four years.

We are not aware of an existing Web-based geoscience image library of the scale or scope of GeoDIL. A smaller collection, and one that in some ways started us thinking about this project, can be found at the University of British Columbia (UBC Earth and Ocean Sciences Image Collections; http://www.science.ubc.ca/~eoswr/image_db/image_db.html .The entire UBC collection contains nearly 2000 images; but it is not growing at present, and the associated supporting data are limited and not as robust or as multipurpose in function as what we propose here.

The GeoDIL project takes place in two main arenas: software development and library development. Designing the database and Web front end is not a trivial problem. For software development, we rely on Henry Borysewicz, Director of the AeroSpace Network (ASN), a division of the John D. Odegard School of Aerospace Sciences at UND. ASN has an extensive history and experience in the application of technology to education and provides instructor support, research, and development for Internet-based educational technology. Borysewicz and his programming staff have done the initial GeoDIL development and will continue to work on the project during the next few years.

Most of our initial effort has been focused on software development, but we are slowly building the image collection. Today, it only contains a few hundred images, but it is growing rapidly. When completed, the GeoDIL library holdings will encompass the vastness of the geosciences and will reflect an earth system science approach to the integration of topics.

High-quality images are essential to GeoDIL‚s usefulness and attractiveness. We are especially interested in having images of very high resolution. Our software can convert them to different image resolutions in response to user needs. As beautiful as the images are, however, they will only be useful if they are accompanied by supporting data (metadata). Pretty pictures without documentation or provenance are just that, artifacts for calendar use, and will have no place in our collection. The supporting data are the most essential and ultimately unique part of our database. They also make this an especially appropriate project for a digital format, allowing the database to be searched in many dimensions.

The more information, the more useful an image will be to a greater variety of users. More information makes the library robust, but if we require too much information, it makes the database unwieldy and adding images becomes a time-consuming, burdensome chore. A key decision, then, was deciding how much information to store about each image. Today, our database includes about two dozen fields for each record (Table 1). These encompass the information that we think will be of greatest importance to geoscience educators and others. We structured our database so that we can add fields after the library is functioning without affecting database integrity.

Scanning and cataloging images is slow work. We have five students working for GeoDIL, and will continue to employ students in the future. Scanning images can be slow, but evaluating them and making sure they are accompanied by the necessary supporting data is even slower. We spend 20%ˆ30% of our time during the school year on this project, and will spend one to two months during the summer. Our goal of 10,000 images in four years is ambitious and is based on estimates of the time it will take us to review and evaluate potential library holdings.

Table 1. Database Fields

image file name

geological category

short title

long title

caption

contributor

photographer

keywords

collection ID

photograph date

camera, film, etc.

planet

country

province/state

county

nearest town

longitude

latitude

other location information

geological province

chronostratigraphy

lithostratigraphy

åbiostratigraphy

We intend GeoDIL to be a community library. Users can retrieve as well as add images to our collection. They can submit comments about images, which we will archive in a user forum. We have developed a „submission‰ page that guides donors through the process of submitting an image and the necessary supporting information. Users enter data using standard Web-based input fields (drop-down menus, radio buttons, check boxes, text fields, etc.) and upload the data and image to our server via the Internet. Our software automatically converts images to appropriate format. We then evaluate the contributions before they become part of our public holdings. During the evaluation period, donors have the opportunity to edit the image or data, but after a contribution becomes public only library administrators can make such changes. Those who provide images to GeoDIL still maintain the copyright and contributors‚ names, unless they request otherwise, are given whenever someone loads an image.

Although the library is up and running, we still have some complicated programming before us. We have a number of aspirations, the most significant is to mesh the browsing and searching routines. Right now, users can browse library holdings by categories, such as minerals, structural geology, fossils, etc. They can also find images by number or by title. More complicated searches involve words, or word strings, in any of the data fields. Users can also do searches involving terms in several fields. We will soon add Boulean „and‰ and other functions to the search routines, and we will develop geospatial search capabilities. Search results are returned as lists of titles or thumbnail images with titles, depending on the user's specifications. Users then click on titles or thumbnails to see images, captions, and other information displayed in a larger format.

Our current search routines work. Unfortunately, without careful choice of search parameters, they commonly yield too many matches. We need to develop simple ways for users to narrow searches without having to start over again. For example, if a search yields too many matches, we want the user to be prompted with keywords that will narrow the search. Additionally, we need to develop ways for users to „wander‰ through the library. If they come across an image they find interesting, we need to have an easy way for them to find similar images based on any criteria they choose. For example, if they are looking at a photo of Mount Saint Helens, they may want to see more photos of stratovolcanoes or they may want to see more photos of Washington. We are working to find ways to facilitate such excursions.

This spring and summer we will be developing virtual carousels for our library. Virtual carousels will be a key feature of GeoDIL because they will allow library users to select and store images for later viewing. They will be able to maintain multiple carousels, grouping photos any way they wish. This means that they will be able to gather images about a given lecture or topic for later showing without searching the database. We will store their carousels in our library, so users can share carousels with others. Users will also be able to download carousel data and images if necessary. Thus users may access GeoDIL at their office or another central location, but use them for display at a remote location where Web access does not exist.

GeoDIL is still a work in progress, but you can access it by going to http://www.und.nodak.edu/instruct/mineral/ and following appropriate links. We are looking for photos to add to GeoDIL, and welcome user feedback..

John C. Butler

Associate Editor

jbutler@uh.edu