Annual BioACT Report 2004/05

During this early phase of the project we focused on:

Field Notebook Study

We interviewed 13 biologists in detail about their use of field notebooks for their work. As part of this effort we scanned a number of notebook samples and analyzed the data types in their contents, occurrences of data collaging, frequency of error correction, layout habits, and so on. This activity has proved to be highly valuable for informing our designs of supporting tools (see below).

Beyond the field notebook study a mixed group of computer scientists and field biologists went out into the field together to experience authentic field outing sessions. Relevant experiences included power and transport capacity limitations, as well as issues of glare, and limitations to the use of paper based dichotomous species identification keys.

Project-Internal Contact Building

We organized two all-hands meetings for which all three project partners convened: the California Academy of Sciences, the Jasper Ridge Biological Preserve, and the Stanford Computer Science group. During the kick-off meeting we established the most immediate common interests that offered starting points for action.

One such area was Prof. Dirzo's plan to install large-mammal camera traps in the Jasper Ridge Preserve. This project will converge with some of the geography-based information organization and retrieval work that is ongoing in the computer science arena.

During the second all-hands meeting, which we held half-way through this first 6-month period, we learned about results of examining the suitability of several camera products.

Broader Community Impact

In November, we organized a meeting around the large-mammal monitoring program. This get-together was attended by a dozen people, mostly docents at Jasper Ridge, who live in the surrounding communities. The purpose was to introduce them to the goals of the mammal monitoring program, and to introduce them to Prof. Dirzo. He explained the camera-trap project, and invited the attendees' participation.

Following this meeting, we began monthly camera meetings with a smaller group that was dedicated to establishing the data gathering protocols. We held three of these regular meetings this spring, conducted numerous camera tests, and generated a large volume of email discussions that eventually led to the final choice of equipment.

One of the computer science students attended a two-quarter class to qualify him as a docent in the Jasper Ridge Preserve. This activity proved very important as it enables the computer scientists to enter the otherwise closed preserve and interact effectively with the environment there.

Early Protoypes and Designs

We engaged in the design and construction of three early tool prototypes: a collaborative species image labeling experiment, an exploration of improvements for manipulating online dichotomous keys, and a field notebook transcription experiment in the Los Tuxlas Mexican rain forest.

The labeling experiment was conducted on a DiamondTouch display table. This instrument looks like a regular rectangular table, but it is in truth a touch sensitive computer display. We implemented for this installation an application that allows field biologists to add labels to photographs that were previously taken in the field. The collaborative nature of the application allows several experts to pool their expertise during the labeling procedure (see technical reports list).

We invested time in studying the state of the art in adding computer support to the practice of using dichotomous keys for identifying species. We developed preliminary mathematical models for identifying optimal inquiry sequences. In contrast to existing methods, our approach is to include knowledge of historical sightings at a particular site, as well as context information, such as the current temperature and the season during which the sighting and associated identification attempt is taking place.

Our third prototype/design effort addressed the problem of transcribing field notebook materials to the computer. We are utilizing a special, commercially available pen, which uses a miniature camera in conjunction with specially marked paper to capture all pen strokes during regular writing activities. The captured strokes are then transferred to the computer where they constitute a true image of the written material. That is, the pen's ink produces a normal written record on paper, while the tiny camera simultaneously produces an online version that can be scaled in size, saved, duplicated, or analyzed in various ways.

One student wrote a number of tools for this system. He then accompanied a class of Stanford biology students and their Professor to the Mexican Las Tuxlas rain forest. Thanks to this grant's funds we were able to equip each of the participating biology students with a laptop loaner, one of the optical pens, and notebooks to accompany the pens. The grant student then conducted usability experiments for our tools right there in the rain forest.

The tools were released as open source.