Sharing geographical data sets is highly desirable for economical and technical reasons. Economically, the cost of acquiring the digital data can be distributed across different organisations resulting in either lower costs for each organisation or greater funding for capturing higher quality data. Technically, sharing data sets is made easier if the underlying framework on which these sets are captured is the same. That is, the underlying framework is itself a shared data set. Consider, for example, if utility organisations providing telecommunication, power, and water services all used the same cadastral data set as the basis upon which they added their information. Individual organisations could more easily share their own information with others because these sets have a common basis.
Communication networks facilitate access to data sets provided the location of each data set is known or can be found. The notion of a geographical data directory has been discussed (Baker, 1996; Newsome, 1995; Anderson, 1995; Pascoe & Penny, 1993), although for different purposes. The New South Wales National Resources Data Directory (NRDD) (Baker, 1996) provides an electronic database in which there are entries describing a variety of data sets related to natural resources such as land, soil, water, catchments, estuaries, atmosphere, biodiversity and so on. Included in each description is information such as the custodian's name, an abstract of the dataset's content, a description of the spatial extent, the representation in which the set is stored, and the name and contact point from whom the data can be supplied. The NRDD is similar to a directory of geographic databases provided by Manaaki Whenua - Landcare Research NZ Ltd (Newsome, 1995).
Anderson (1995) discusses the development of a GIS directory for the purpose of organising the storage of resources on a computer system. This directory structure is required to: be easy to learn and use; support the needs of a multi-departmental user community; support incremental implementation and expansion; support database management and system administration functions; and reinforce the concept of a commonly shared and integrated system. In this instance the discussion is on the hierarchical (tree) structure of the directory, deciding what resources are stored in particular branches of the directory and what these branches and sub-branches are called. Individual resources are not described in this directory. In the author's earlier development of interfaces (Pascoe & Penny, 1990), the function of these software packages was as 'a mechanism by which one data structure can be directly converted into another for the purpose of communication between systems or sub-systems' (van Roessel et al., 1986). Consideration of software packages operating within a network environment lead to the definition of a communicating interface as being 'an interface that sends data to, or receives data from, some other interface through a communications network' (Pascoe & Penny, 1995).
With the introduction of a network environment, the functionality of a communicating interface could be easily expanded to include other functions. In an earlier paper, Pascoe & Penny (1993) descried the use of a primitive geographical data directory to search for data sets that conformed to user defined criteria. Essentially the user defined the spatial extent of interest and the interface then searched the directory for data sets that overlapped this extent and retrieved these sets from the locations specified within the directory entries. Retrieval of these data sets included transforming the data representation if the required representation differed from that in which the data was available; and moving the data from one computer system to the user's computer system if they were different.
Searching a geographical data directory for data sets that match some user criteria clearly extends the functionality of a communicating interface and introduces the possibility for adding other functions such as: query processing to enable users to define sophisticated criteria that data sets must match; searching for other interfaces to perform data translations where these translations are beyond those provided by the interface searching for the data; the integration of several data sets, each matching the user's criteria; and so on. Given the increasing number of complex functions to be performed by an interface, in this paper the author suggests a paradigm switch such that an interface becomes one or more specialised software agents, which collaborate to achieve the collection of geographical data sets.
In this paper a project is described that aims to significantly extend the author's earlier use of a geographical data directory by developing a standard X.500 Directory entry to describe a geographical data set and the location of any associated software to be used for supplying this data set through a communications network. The X.500 Directory service is a facility for storing and retrieving widely distributed information. The Directory is hierarchically structured with branches of the tree stored on different computer systems and independently administered. A wide variety of information can be stored within the Directory. For example, names, addresses, photographs, and pointers to software that provide publicly accessible services. The Directory can be browsed by users or searched by the specialised data agents mentioned earlier. ISODE (ISO Development Environment) contains an implementation of the X.500 Directory which the author has been using for practical demonstrations of the research discussed in this paper.
Anderson, J.S. 1995. "Building a Useful GIS Directory: Snohomish County, Washington", URISA Journal, 45-52.
Baker, H. 1996. "Data Directory Blazes Trail for Metadata Standards", The Australasian Geographic Information Systems Applications Journal, GIS User, 15.
Newsome, P.F.J. 1995. Directory of Geographic Databases within Manaaki Whenua - Landcare Research, Technical Report, Whenua - Landcare Research New Zealand Ltd., Private Bag 11052, Palerston North, New Zealand.
Pascoe, R.T. and Penny, J.P. 1990. "Construction of interfaces for the exchange of geographic data", 4(2), 147-156.
Pascoe, R.T. and Penny, J.P. 1993. "Transforming geographic data between different concrete representations". In Gupta, G., Mohay, G. and Toper, R. eds. Proceedings of the Sixteenth Australian Computer Science Conference, 653-663.
Pascoe, R.T and Penny, J.P. 1995. "Constructing interfaces between (and within) geographical information systems", 275-291.
van Roessel, J., Bankers, D., Connochioli, V., Doescher, S., Fosnight, G., Wehde, M. and Tyler, D. 1986. Vector data structure convcersion at the EROS data center, final report, phase I, Technical report, EROS Data Center, Sioux Falls, South Dakota, 57198.