The Digital South: summary for ACS GIS Symposium, 28 February 2003

Hugh Blackmer, Washington & Lee University

(n.b. 7 PM version, 26 Feb...)

It occurred to me, looking back to the Richmond GIS Symposium of December 2001, that we at W&L have continued to work on what the concluding session identified as a priority: the means to access spatial information.

The fundamental idea of the Digital South is that we can share resources: we can build structures --essentially, a distributed digital library-- that help us to solve common problems, which come down to challenges that we all face in the support of teaching and learning. The Digital South project is attempting to build a basis for distributed management of spatial data resources,

in the expectation that the resulting digital library will prove to be useful to participants. It's really important to think of this digital library as something that we build, by our cooperative and collaborative efforts.

What brings us to Southwestern is a common interest in the problems of support for and development of GIS on our separate campuses. Each campus has a different constellation of resources and interests --not just at different stages of development, but with different personal and institutional visions of what GIS is for, where it belongs in disciplinary senses and pedagogical niches. What we have in common is an INFORMATION problem, and underlying that is a stratum of DATA problems.

Consider the INFORMATION problem first: the problem is an embarassment of Information, which we must contrive the means to integrate into what we do --into teaching and learning.

Accuracy of the estimate that 80% of information is spatial is dubious, but a LOT is, certainly. We know that the spatial dimension is vitally important in just about every discipline, and specifically that it's somewhere near the center of our own interests --that's what brings us into this room. We all have, to greater or lesser degrees, knowledge of the enabling technology --GIS-- for study and learning in that spatial dimension. And we all know that most of our colleagues and institutions are significantly benighted in this regard.

In order to attract the institutional resources we need, and we can see that others need, we have to raise the levels of awareness, interest, and knowledge of our colleagues. We need to figure out how to make the spatial dimension (1) salient and (2) accessible to them. GIS Days and other showcasing activities do this, but it's also necessary to find other ways to involve potential users by working with their problems and augmenting their resources.

A generation ago the spatial dimension was largely the concern of the discipline of Geography, though a few others (Geology most prominently) made significant use of maps and spatial technologies. A generation ago there were no desktop (let alone laptop) computers, and we weren't bound together by the just-emerging Internet, and we weren't awash in digital data, and if you wanted to find a book in the library you used the card catalog... I typed my dissertation on a rented Selectric, and drew maps by hand because there was no other way to do those things.

But now everywhere we turn we see exhortations about the necessity for interdisciplinarity, and opportunitires for new ways to explore and represent and convey data, and new demands to expand what we know and [think] we understand... e.g., O'Reilly offers a Basic Linux for Bioinformatics book... a conjunction that wasn't on the radar even 5 years ago, and symptomatic of the ways in which links between disciplines pervade.

If the Information problem has to do with volume and how to integrate into teaching and learning, the data problems include many technical issues having to do with collecting, finding, massaging, converting, processing, storing, managing, curating, documenting, distributing... and probably others we could name. It's all too easy to get entangled in the TECHNOLOGY of GIS and forget that it's fundamentally an enabling technology, and what really matters is what one can do with it: (a) visualize data and (b) explore patterns observed. The descriptive task, which generates questions (basically, why do the data look like that?) then provokes analyses of the data.

We can summarize the data agenda thus:

The point is that our activities with GIS should be building a growing knowledgebase, useable by ourselves and by others. We should have the means to integrate our GIS activities with emerging and developing digital libraries, not just so that we can find the data we need, buit also so that others can find what we have done.

How do we get there? In the short run we all have some obvious needs: to attract local support for the expenses of effective implementation of GIS, and to get access to the requisite expertise to carry out GIS projects. I'm hoping to hear diamond-cutting ideas from others on how to accomplish this...

In the longer run, we have to prepare for what we'll do with the results of our GIS activities: how we'll store datasets so they can be found again, and shared, and used to display results of analyses. This is not just a "library problem" --indeed, few librarians recognize it as their responsibility. Looking further down the line, we recognize the need to develop the distributed management resources beyond ACS --and NITLE is the next sensible unit to think in terms of, but there's no reason to stop there, and eventual interlinkage with DLESE and with NSDL are surely possibilities toward which we should be working.

The Digital South began as a pipedream, a speculation about ways we could use technologies to build a common resource for the member schools of ACS. The focus and the technical details of the Digital South have morphed and drifted in the years since the term was coined, and there's no reason to assume that everybody associated with the project --or with the label-- has the same set of ideas about what it is or where it's going, or how it might get there.

You're mostly going to hear my late-February 2003 version of the Digital South, but I expect to be interrupted with questions and alternative visions. What I want to get across is how the Digital South fits with ACS's efforts to develop and support GIS --how what we are trying to build can support multiple agendas, and multiple versions of what GIS is about.

The Digital South is the infrastructure for a digital library that will be built by contributions from the members of ACS. Its collections will not "be" in any one place, and indeed its hardware can be dispersed. The library consists of pointers to digital resources --map layers, images, texts, sound files, etc. A user who needs, say, census data for South Carolina or a satellite image of Memphis might find the resource by querying the Digital South, to discover that the desired resource was on a server at Furman, or at Rhodes. Likewise, a user who has created a dataset on stream health or forest succession or Civil War battlefields might add the resource to the Digital South by making a pointer to the location of the dataset, and supplying the necessary metadata.

The goal is not to have "everything" in the Digital South, but rather to have a place to build a shared resource base --and eventually to facilitate the linkage of Digital South contents to other emerging digital libraries. NSDL is an interesting model, and we could do worse than to try to build the Digital South on a path that eventually converges with NSDL --so that a researcher in Nebraska could find and retrieve a resource tha the Digital South has declared public. The functionality exists for just this sort of service, exemplified by the Alsos Digital Library on Nuclear Issues at W&L.

The core technologies here are

So the Digital South is a database project, and the content we're interested in here is GIS data --though clearly other media are includable and, as we'll see, linkable with GIS data. The Digital South is also a foot in the door to raise the profile of spatial information and GIS on our campuses, via the ArcIMS Web interface. The idea here is that we can only attract the $$ and staff resources to support GIS if we can raise the demand on our campuses, by demonstrating the utility of the technology as a means to augment pedagogy (well, maybe not the only way, but an effective one).

My objective is to see GIS as a cross-disciplinary tool --to get geo-graphy back into teaching and learning. My strategy for that is to contrive support for ANY sector or department that can make use of GIS, but to concentrate especially on interdisciplinary applications --and Environmental Studies seems the most promising way forward on most campuses.

It still requires an active imagination to project to what the Digital South will look like and how it will work, but a sketch of the modules will clarify the basic functionality we're working toward. This sketch is also meant to exemplify the importance we see in situating GIS within the liberal arts --in connecting the technology with teaching and learning. The issue is not just how do we implement and support GIS (nor focused how do we deliver the technical skills), but also how do we encourage its application across the curriculum, and how do we collect and manage the results of GIS activity, so that they can be redistributed.

Our eventual deliverable is an integrated application that's designed to converge with other digital library developments. Along the way there will be changes of direction as we are able to link to other developments, and as new technical possibilities move to practicality. I want to show you some of the pieces, so you'll have a better idea what we're trying to build and what its modules are:

In Summer 2002 we built Pirarucu, a personal digital library environment with facilities for creating and managing collections and collaborations, and searching the complete library. This project was aimed at the information needs of a FIPSE-funded consortial relationship focused on sustainable development in the Amazon, involving two Brazilian universities in partnership with Washington & Lee and Fairfield University.

In the last few weeks Skip developed SnipIt, a toolbar utility that allows a user to highlight text on a Web page and send the text and URL to Pirarucu.

We have been working with ArcIMS for more than a year, and have a variety of demos of functionality, including

What I did here was to go through the steps a resource creator might follow, using W&L's Frank Reader Diary (written by a Union soldier as he moved down the Shenandoah Valley, March-July 1864):
  1. looked up coordinates for towns named in entries in the online version of the Frank Reader Diary in the Getty Thesaurus of Geographic Names
  2. entered coordinates and URLs and brief text into a database table
  3. used ArcMap 'add xy data' to create a map layer
  4. grabbed metadata for the layer
  5. exported data as .shp
  6. saved .shp and related files in Pirarucu
  7. added the layer to the Civil War protoprototype
The summer Civil War project will give us considerable experience with the creation and collection of spatial data and metadata, both from the Comp Sci point of view of engineering the interfaces, and the historians' digitizing and georeferencing candidates for inclusion. Once the basic structures are built and tested, we will seek contributions from other ACS institutions, and gradually broaden the scope and size of the digital Civil War collection. Along the way we will have to deal with problems like communicating standards, vetting content, controlling access to some materials, and managing servers in a distributed environment.

The next stage in the evolution is still on the drawing board, but I'm working on a proposal to NSF's NSDL initiative, involving the extension of the emerging Digital South structure to support Environmental Studies programs, and thinking that ACS schools would be the appropriate testbed. You may want to look at some very preliminary notes for that proposal.

Here's where we're heading:

Links to various previous documents, tracing the evolution of the Digital South:

June 1999 e-mail to Bob Whyte re: ACS GIS

November 1999 GIS: an insurmountable opportunity? (remarks from ACS Information Fluency Symposium)

December 1999 A GIS Server for ACS Partners

January 2000 Developing Geographic Information Systems for ACS Partners: Toward a GIS Server

March 2000 Geographic Information Systems (GIS) Across the Curriculum: Implementing Distribution to 15 Campuses (LAAP proposal)

December 2000 GIS in Undergraduate Teaching

July 2001 GIS in midsummer 2001

December 2001 Collaboration to Manage Spatial Information on ACS Campuses: prototyping an information infrastructure (Information Fluency Grant Proposal)

Summer 2002 Pirarucu Project

November 2002 GIS on ACS Campuses

January 2003 Updating and Extending the Digital South Project

February 2003 R.E. Lee Summer Project Proposal: The Civil War and the Digital South

February 2003 Thoughts on NITLE and NSDL and GIS and Environmental Studies

February 2003 Outlining our presentation(s)

===

fragments:

Metadata harvesters are the key element that binds collections into the operation and effectiveness of a digital library: the harvester tends the constitutent units, queries them for new content, and augments a central repository of pointers to resources.

So we have exploratory, analytical, communicative as functions of GIS technology. Each has a place in teaching and learning.

I come back to why GIS matters: fundamentally, GIS gives us the tools to visualize spatial data: to see and then to analyze patterns in space and time, and to communicate the results to audiences. A learning tool, a teaching tool, a communication tool. The technology also enables efficient communication of complexities: a picture of data is efficiently grasped, while a page of numbers is much more difficult to assimilate.