Brandon Plewe
Assistant Professor
Department of Geography
Brigham Young University

The Potential Academic Contribution to GIS Interoperabilty

     To date, the interoperability of Geographic Information Systems (GIS) has been seen largely as a technical issue: "How can we enable software from different vendors, which use very different data structures, to communicate and share data with each other?"  The OpenGIS Consortium has been quite successful at facilitating the communication between vendor communications which is necessary to develop the technical solutions to interoperability.  The release of products which are based on potential OGC standards, such as Intergraph's GeoMedia and LAS' GRASSLAND, while not perfect, show the great deal of potential which lies in the sharing of heterogeneous data.

     However, in addition to the technical issues being dealt with, there are also many scientific and societal issues in interoperability (and the superset field of Distributed Geographic Information, or DGI) which have not received nearly the same level of attention.  This is because these issues are not part of the stated mission of OGC (for good reason, they intend to keep their focus on the technical side), but are left to the academic (geographic information science) community.  How can (or should) the GIS community, and the public at large, use this new technology?  In turn, how might interoperability radically or subtly change the community itself?  In many ways, these questions are much more difficult to resolve than the technical obstacles.

     There are many issues which should be considered by the geographic information science community.  Some are new problems which are created by interoperable GIS, while others are old issues which are made more important (or more problematic) by this new technology.  Some of these are listed below:

Overlay of Disparate Information.  One of the core capabilities of GIS is the comparison of multiple themes to display (in a map) or analyze the relationships between them.  The basic assumption behind the overlay functionality is that the themes being used are comparable.  For example, if one is creating a multiple-criteria region, one must assume that the input themes have similar levels of accuracy, dates of source information, projections/coordinate systems, etc.  In a time-series analysis, the date assumption is removed, but the others are still in force.  OpenGIS gives us the opportunity to build GIS projects which include data from many sources, which may or may not be comparable.  This potential conflict is augmented by the fact that the GIS user will
likely not be as familiar with these remotely obtained sources as with traditional locally generated data sets, and will thus be less likely to recognize conflicts.  The research community should be involved in developing technical and educational means (probably involving metadata) to assist users in avoiding and resolving conflicts.

Conflation.  This special case of overlay has its own issues.  Since the data sets which are being combined represent the same geographic entities (i.e. roads from multiple sources), the standard of comparability is much higher than for standard overlay.  Conflation will be incorract if the data sources have very different levels of positional accuracy, different ages, levels of detail, or classification schemes.  One may wish to prevent inappropriate combinations, or weight the data sources so that arbitration decisions will favor more accurate or more recent data.  Again, OpenGIS increases the ability to include data sources with which the user is not intimately familiar, and thus makes this issue more pressing.  Developing automated yet intelligent means for conflating disparate data sources is a prime area for academic research, especially in partnership with GIS vendors.

Data Sharing and the GIS Community.  Although it has been espoused for several years (the U.S. Federal Government probably being the most vocal proponent), widespread data sharing has not caught on among the general community.  This is largely because several important obstacles have not been fully overcome.  These include legal issues such as copyright and protection of privacy, technical issues such as security, automated data purchasing and effective marketing of available information, and societal issues such as the tendency to hoard data sets in which you have invested a considerable amount of time and money (which occurs not only between organizations, but within them as well).  Why should public and private organizations share data with each other and between individuals within themselves?  How is this best accomplished, in terms of both the technical approach, and organizational policies?  Some of these concerns may be resolved by the industry, some by governments, but the academic community can also contribute to this issue.

GIS and the Public.  The technical advances of interoperability will likely increase the access which the general (i.e. non-GIS-savvy) public has to geographic information, whether via the Internet or other means.  This raises many academic issues: how can the software assist naive users (naive in terms of GIS techniques and/or the subject matter) in effectively obtaining the information in which they are interested?  Should (and how should) the public be better educated in the principles of geography and geographic information to enable them to better use and understand GIS?  The recent research into naive and cognitive geography may have pertinence here.

Information Retrieval.  As interoperability increases the number of sources (internal and external to the organization) from which data may be obtained, it becomes increasingly important for those sources to be easy to find and obtain.  While initiatives such as NSDI and the Digital Libraries Initiative (especially the Alexandria Project) are making strides in developing effective search and retrieval mechanisms for very large stores of spatial data, there is much more to be researched.  One question which has not been studied much is the scalability of search mechanisms to handle perhaps millions of data sets from thousands of servers.

Translation Loss.  As with any translation process, the conversion of native data archives to the standard OpenGIS transfer formats may result in a loss of information.  The vendors participating in OGC have worked hard to minimize this loss, and thus make the standards appropriate for most applications.  However, there may be some applications in which the lost information is vital, and thus are not implementable using the standards.  Scholarly research would assist in locating these applications, and subsequently augmenting the standards or developing new methodologies for implementing the application which will work within the OpenGIS framework.

     These issues, among others, constitute a full research agenda for the academic community to contribute to the fields of DGI and GIS interoperability.  Due to the nature of the GIS community, this research has and will continue to happen in cooperation with the GIS industry, as well as governmental institutions.  The primary contributions of the geographic information science research community to this area are the same as they have always been: to develop fundamentally new solutions to difficult problems (leaving incremental improvements to the software developers), and in discovering the most effective use of GIS technology, to use the tools to accomplish the aim of Geography: a better understanding of the world.