Interoperability of Geographic Information: From the spreadsheet to virtual environments

Pedro Pereira Gonçalves, Nelson Neves, João P. Silva, Joaquim Muchaxo and António Câmara
Grupo de Análise de Sistemas Ambientais
Faculdade de Ciências e Tecnologia
Universidade Nova de Lisboa
email: pmg@uninova.pt

Introduction

Environmental modelling, statistical data analysis and GIS are three examples of computational activities that are usually relegated to separate, large and sophisticated computer systems. Usually each one of them has more functionality than is required by the user, whose needs exceed the range of a single one (e.g. both GIS and statistical data analysis). In the first two, the data organisation benefits from a GIS perspective and some very simple issues like the quality of spatial coverage of the data can be resolved by a simple geographical display. Likewise, the quality of the time-dimension over a geographical region can be combined with a geographical display by enhanced visualisation. For these two examples (spatial and temporal coverage of data), for interfacing models over a large range of possible input parameters, and for many other circumstances, it is beneficial to have a level of integration which is much greater than a simple file transfer mode.

One of the most often critic to Geographic Information Systems (GIS) is that has failed to give adequate attention to principles of cartographic design, or for regarding the map as a simple store of information rather than a tool for communication [Goodchild, 1990]. If the database could be considered as truth, then the produced map would be no more than a store device, since would exist a simple correspondence between objects on the database map and objects on the map. But, usually, the database is an approximation of the geographical truth and, since it can affect the user's view of the world, the design of the output display is critical. The actual of electronic display go far beyond conventional cartography.

The advent of the Open is Geodata Interoperability Specification (OGIS) opens for the first time a real opportunity to develop data structure-independent GIS applications [Buehler, 1994], [OGIS 96]. The Open GIS Consortium (OGC) does not attempt to define high level operations. Its specifications are restricted to low level database (SQL-like) and topological operations based on the work by Egenhofer [Egenhofer and Franzosa, 1991] and [Egenhofer et al., 1993]. This is achieved through the use of a common language for sharing geodata and standardised definitions of interfaces to functions that operate on geographic information. While their initial work is focusing on traditional geoprocessing, such as spatial selection, thematic overlay, measurement, and distance analyses, other services which access geographic information, such as hydrodynamic models, seismic prediction, and allocation functions, will also be able to directly access geodata stores (as well as other geoprocessing functions).

But the effects of the development of object-oriented databases and object-oriented programming systems in computer science are likely to be much more profound. The argument that has to be done is that GIS software does not have to be visible to the end-user. By the contrary, the GIS as to be understood as another variable available to the end-user. The GIS community needs to transform all the GIS concepts, and functions, into a variable as simple to define an integer, byte or float. In this context the NovaGIS project [NovaGIS, 1997] pretend to establish the concept of an "invisible GIS". Adding to the computer system the variables necessary to work with Geodata.

System Description

The NovaGIS software is an OLE server that creates an interface to Geodata. One of the main goals is to transform Geodata into a normal variable in common macro language. Using this object traditional geoprocessing services can be added to our software. At the same time, the GIS application remains invisible to the user. When the variable is define the user works directly in the data structure and functions.

For example, a simple Basic programme can do the query of a map:

The simplicity of defining our map, study area or set of data is given at the same time by the potential of many macro languages present in common office applications. At the same time there is no need to invent or rebuild new script or macro to work with this structure. The full potential of software dedicated to environmental modelling, statistical data analysis simulation, or even data visualisation can enhanced by the add-on of GIS concepts and functions.

The NovaGIS server can be used in any OLE compliant program like Excel, Visual Basic, Delphi or any other OLE-compliant software. All these software products have the capability or to run macros or even to compile the source code. The system functions are well separated, but to the end-user the system works as one. This interconnection is achieved by means of the Object Linking and Embedding (OLE) from Microsoft. The foundation of OLE is the Component Object Model (COM), that dictates how OLE applications behave and interact, and provides mechanism that lets one application connect to the OLE interfaces that another supports.

Application

The case study chosen was forest fire modelling, using the model FireGIS [Gonçalves and Diogo, 1994], and to prove the feasibility of the systems, two intertwined applications were built. In the first, the forest fire model interface was build in a spreadsheet of Microsoft Excel, with all the data input, model parameters, map and charts visualisation (Fig.1b). At the same time the outputs of the model are seen in a Virtual Environment (VE) using the VirtualGIS system [Nelson et al., 1997].

The VirtualGIS application updates the Digital Terrain Model (DTM), where a correspondence established between cell sizes in the virtual environment and the simulation environment that is established when the model starts running. As new simulation steps are computed the information about the cells that changed is sent to the VE, reporting the cell location (x,y) and the new cell value. At the same time the user can operate the data inputs and model results in a common spreadsheet program. In Fig. 1a a snapshot of the visualisation of the outputs of a forest fires simulation model in the VE are shown.

Fig. 3 - Visualising a forest fire in the Virtual GIS Room (a) and in a spreadsheet (b).
So far all the tests were done in a single computer by means of OLE. In the case of several computers will be used the distributed version of OLE. The Distributed COM (DCOM) is a technology that extends the local capabilities of COM to cross network communications between objects on different computers. This will enable the GIS and VE system in different computers and even in completely different locations. The use of a virtual environment system to visualise and interact, with spatial information and associated simulation models, allows the user to explore information recreating their usual interaction in a real GIS room.

The objects created can be used in different applications, allowing their use in a diverse range of applications, from word processors, spreadsheets, programming tools (like Visual capabilities wazzu Basic, Delphi or C++). This structure fits as well to the development of different versions with access through HTTP protocol. The interoperability of the system allows the design of different user interfaces according to the different needs as well a more integrated analysis of the results.

Acknowledgements

This work was partially supported by JNICT under research contract PEAM/C/FGI/633/95. The PRAXIS grant 3/3.2/AMB/04/94 supports the work of Pedro Pereira Gonçalves.

References

[Buehler, 1994] Buehler, K., 1994. OGIS Project Document 94-019. OGIS Geodata Model Overview. The Open GIS Foundation, Cambridge, MA.

[Buttenfield and Ganter, 1990] Buttenfield, B.P. and Ganter, J. H., 1990. Visualization and GIS: What Shoul we see? What might we miss? Proceedings of the 4th International Symposium on Spatial Data Handling, Zurich, Switzerland, Vol.1, pp. 307-316.

[Clapham, 1992] Clapham, S., 1992. A formal approach to the visualization of spatial data quality, in Proceedings of GIS/LIS'92, San Jose, CA, Vol.1, pp.138-148.

[Egenhofer and Franzosa, 1991] Egenhofer, M. and R. Franzosa, 1991. Point-Set Topological Spatial Relations. International Journal of Geographical Information Systems 5(2), pp. 161-174.

[Egenhofer et al., 1993] Egenhofer, M., Sharma, J. and D. Mark, 1993. A Critical Comparison of the 4-Intersection and the 9-Intersection Models for Spatial Relations: Formal Analysis. Proceedings, Eleventh International Symposium on Computer-Assisted Cartography (Auto Carto 11), pp. 1-11.

[Gonçalves and Diogo, 1994] Gonçalves, P. and Diogo, P., 1994. Geographic Information Systems and Cellular Automata: A New Approach to Forest Fire Simulation in Proc. of The European Conference on Geographical Information Systems (EGIS 94), Paris - France, 1994.

[Goodchild, 1990] Goodchild, Michael F., 1990. Keynote Address: Spatial Information Science. Proceedings of the 4th International Symposium on Spatial Data Handling, Zurich, Switzerland, Vol.1, pp. 3-12.

[Goodchild, 1992] Goodchild, M., 1992, Geographical Information Science, International Journal of Geographic Information Systems, 6(1), pp.31-45.

[NovaGIS, 1997] NovaGIS homepage. http://virtual.fct.unl.pt/~pedro/NovaGIS/

[Nelson et al., 1997] Neves, N, Silva, J.P., Gonçalves, P., Muchaxo, J., Silva, J. and Câmara, A., 1997. "Cognitive Spaces and Metaphors: A solution for Interacting with Spatial Data", Computers & Geosciences, Vol. 23, No.4, pp.483-488.

[OGIS 96] OGIS. Open GIS Consortium. http://www.ogis.org/