Experiments with Information Ecology in Landscape Studies

In the summer of 2012 the Landscape Studies department at Dumbarton Oaks funded internships in Landscape Studies. Three of the interns, Alexis Delvecchio, Siobhan Aitchison and Robin Abad participated with Paul Cote, GIS Specialist at the Harvard Graduate School of Design in some experiments in collaborative place-based research. This is the fourth year of collaboration between Paul and Dumbarton Oaks. To date, the collaboration has focused on assembling several epochs of detailed survey data regarding the ground plan and plantings in the garden into a systematic Geographic Information System. For background on these earlier phases, see Data Model for a Managed Landscape. This year the focus of the project became a study how to build an infrastructure for collaborative landscape research, and to investigate the application of an ecological perspective to landscape research.

This page is a working draft of the ongoing documentation of this summer's findings.

Infrastructure: Open Source Services on the Amazon Elastic Cloud

The setting for our experiments is the server on which you are viewing this page. www.planetable.org is a Linux virtual computer running on the Amazon Elastic Compute Cloud. Our virtual computer (also known as an instance) is running the open-source Ubuntu Linux operating system, version 11.1. We started our project with our virtual computer running in an amazon micro instance which is hosted for free. Halfway through the summer upgraded to a small instance which costs about $40 per month. Our server is running several pieces of open source software:

Information Ecology in the Cloud

The notion of information ecology in landscape research will be fleshed out in more detail later. For now, it will suffice to say that an ecological approach to research takes a look at the elemental activities, like collecting and documenting images, and attempts to understand how new information emerges from arrangements of the elements and activities at multiple scales. While ordinarily we may think of research on a project by project basis, the following experiments are an attempt to see how the interaction of several actors and activities may create effects that span multiple projects.

Web-Based Management of the Dumbarton Oaks Trees Database: Building on Our continuing project to build an information system for managing the garden at Dumbarton Oaks we have harnessed geoserver, along with OpenLayers a powerful set of tools for creating browser-based mapping applications, to create a mobile application for managing the trees database. This application interacts with the GPS in your mobile device to indicate your current location in the Garden and permits the user to recored their observations of trees.

The Collaborative Image Repository: Alexis, Siobhan and Robin were both concerned with documenting aspects of the evolution of the Garden at Dumbarton Oaks and the greater Washington DC setting. Their work involved collecting hundreds of images form various sources. Our Omeka Repository provided a place for them to store their images with documentation that makes their work accessible to each other and for others who may be interested in carrying their work forward.

Interpretive Exhibits: Once images have been deposited into Omeka with their critical reference information, they are accessible to be organized into any number of narratives. Alexis learned how to apply styling to Omeka to create this Home Page that serves as a setting for Alexis's History of Dumbarton Oaks with particular focus on the Arbor Terrace or Siobhan's history of the Kitchen Gardens with particular focus on the excavation and simulated reconstruction of the mysterious frame house. And Robin's In-depth exploration of the history and evolution of Washington DC's distinctive triangle parks.

Interactive GeoTemporal Exhibits with the Omeka Neatline Extension: As if the interpretive exhibits presented above were not enough, Alexis took her work into a couple of new dimensions with a brilliant demonstration of the Neatline plugin for Omeka in her exhibit on the History of Dumbarton Oaks. The Neatline extension is a terrific demonstration of how the effort of adding geographic and temporal references to the images in Omeka and the scanned maps hosted in geoserver generates a terrific amount of new information as the historic and geographic relationships among elements emerges in the composition of a Neatline Exhibit.

Four Dimensional Modeling and Off-Line Compositions

The use of cloud-based infrastructure for research exposed a couple of important issues for our team. One of these was the problem of independence. As a temporary intern beginning a long-lived research project, Robin, for example had real concerns about building too much dependence on an infrastructure that was not under his control. An alternative to the using omeka as the repository and vehicle for organizing images and text with spatial and temporal references is to use a strictly local repository that may be pulled together into an interactive 3D model using Google Earth. I currently don't have an example of Robin's Google Earth work. Alexis also used Google Earth as an alternative way of exploring the history of Dumbarton Oaks. You are encouraged to download her KMZ archives from Alexis Delvecchio's Omeka Collection. We can go into all of the wonderful ways that information can be referenced using this terrific medium as work on this project continues.

It should not go unnoticed that from a standpoint of information ecology, there are drawbacks to having researchers develop their projects off-line, just as there are advantages. One of the conclusions that we will get to as we flesh out this analysis will be how we may establish ways of working that preserve the advantage of off-line compilations with shared on-line repositories.