Prof. Arup Dasgupta
Honorary Managing Editor GIS Development
[email protected]
There are three key words that describe the field of Geospatial Information Systems today: enterprise, convergence and democratisation. I have deliberately used the word ‘geospatial’ instead of the more common ‘geographical’ to stress the fact that “the GIS train has truly left the geography station”.
I have picked up this allegorical reference from an article by Mike Goodchild in which he has opined that GIS, in terms of technology and applications, is likely to move beyond the scope of traditional geography. This is not to downplay the importance of the science of geography in a geospatial system. Physical geography does provide the theoretical underpinnings of GIS. What is more important is that geospatial technology also provides geography with an opportunity to go beyond its confines and rediscover itself. Incidentally, there are many who think the term geospatial is an oddity and geographical – or geographic, if you are from the US – encompasses all that is happening in the world of digital spatial technology. I will not venture to add to this debate – in any case my preference is for geomatics which also happens to be the accepted ISO term for what we are going to discuss.
ENTERPRISE
The lifecycle of geospatial information systems closely follows the life cycle of general information systems. The advent of computers gave birth to the technology of data processing and the establishment of the electronic data processing (EDP) unit in most enterprises. The EDP cell remained in the periphery till it was given due recognition as the management information system (MIS) department and became an integral part of the enterprise. The advent of desktop computing moved many of the MIS functions to the individual desktops but the need for data synchronisation, security and redundancy avoidance forced the move to clientserver systems with the EDP/MIS managing the data. As IT became more integrated with the enterprise, we saw the advent of office automation, enterprise resources planning, supply chain management, back office operations and customer relationship management to name a few of the applications. This resulted in a multiple client server systems within an enterprise and the rise of the enterprise- wide distributed data processing paradigm with multiple servers across divisions, many geographically dispersed, linked together in a network. The advent of the Internet and the World Wide Web provided a network which could be used to configure a virtual private secured network for the enterprise and an open network for public access.
The GIS lifecycle began as a distinct unit variously called the digital mapping centre or GIS division. GIS was mainframe based and access was through command line based graphic terminals. Networking was not on the radar. In mid 1980s, a prominent GIS vendor expressed more confidence on a minicomputer based system rather than a workstation based product. Even as late as 1999, ESRI was distributing GIO lapel badges at its user conference. GIO stood for Geographic Information Officer, a variation of Chief Information Officer – a term popular in the EDP/MIS era. The advent of the desktop GIS led to a proliferation of GIS installations and the headaches of data
duplication and synchronisation across the enterprise. Enterprise GIS was the solution. An additional factor was the discovery of the utility of spatial information by new users. Spatial information is a given for defence services, engineers and natural resource managers but now areas like marketing, CRM, ERP, governance, health, education and law enforcement have begun to appreciate the power of spatial information. As a result a need has arisen to create a core spatial database which could be used by many users. This is the emerging picture of GIS. We are seeing such cores becoming an integral part of traditional IT systems like ERP, CRM and e-governance. The enterprise GIS is thus a distributed system where the core geographical data is managed by one agency or department and each branch of the enterprise adding its own unique dataset to the core and running its own analysis models on the ensemble. The importance given to standards and interoperability issues is a mark of the emergence of such enterprise solutions. As GIS becomes an integral part of an enterprise solution, users will be from diverse disciplines. Here, ease of use holds the key to success.
GIS designers have found the Web to be the ideal medium networking and the open Web technology as the vehicle for interoperability and for ease of use. Web GIS thus has made the best use of the developments and standards of the WWW Consortium, W3C to create innovative open solutions. Web 1.0 provided a simple solution for map publishing; Web 2.0 has made possible a degree of interactivity which enables not only data but service dissemination over the Web. ISO and OGC standards like GML, WMS, WFS, CSW and many more, based on Web 2.0, provide the framework for spatial data interoperability across heterogeneous distributed systems. 
The W3C, ISO and OGC standards are in a process of evolution and keep step with evolving technologies. Heterogeneity extends from hardware to operating systems to applications software and data. In order to maintain a degree of independence of individual systems, the technology trend is towards a coarse grained service oriented architecture along with SOAP or REST protocols to provide inter-system communications. Such a system will insulate individual systems from getting affected by external changes. One of the most important evolving applications systems is mashup. Mashups provide applications development on the fly. It consists of the client, usually a browser; a server which provides a Web page containing an application that uses the server’s own data and data from other content providers; and servers of other content providers that provide data through APIs using SOAP or REST protocols or other Web services. Much of the interaction in mashups happens between machines and this brings up the problem of heterogeneity in naming and cognitive semantics. Web 3.0 is expected to provide solutions to such problems through research on the semantic Web.
CONVERGENCE
Web GIS represents one face of the convergence of technologies. Many other technologies are becoming important elements in the geospatial world. Convergence can happen at device level, like GPS enabled cellular phones and GPS enabled PDAs for navigation; at data level, like maps and high resolution imagery at street level and at application interface level for mashups. Navigation systems like GPS and GLONASS are not only important for surveying but have become important elements of location based services like vehicle and personal navigation, asset tracking and point of interest (POI) location. GPS augmented cameras can tag scenes with position and time, enabling their location in space-time on a map. GPS tagged, metric quality street imagery as provided by for example, Cyclomedia can be used to update urban databases rapidly. Mashups enable creation of new applications on the fly. Many new services are coming up like Geocommons which enable novice users to create data mashups. Geospatial data and services will be accessible seamlessly over different platforms ranging from desktops to cellular phones enabling a person to ‘carry’ his work from the desk to the field and vice versa seamlessly. Mobile devices also enable rapid acquisition of data from field stations. For example, electric utilities use WiFi enabled meters in residences which can be queried and their data downloaded by mobile vans equipped with readers connected to the central SCADA station. This overcomes the problems of locked residences and the labour of a meter reader going from house to house.
DEMOCRATISATION
Spatial context has become an important factor in everyday life of the common man. Car or personal navigation, location of points of interest, issues related to common property resources are some of the applications which are of relevance. Enterprise GIS and convergence have led to services that have put spatial content and the ability to manipulate this content in the hands of the common person. This democratisation of spatial technology, data and applications has given rise to the neogeographer. A neo-geographer is a person who uses the simple tools made available by services like Google and Microsoft Virtual Earth to add geographical information about his environs to a database. Another term coming into prominence is ‘crowd-sourcing’ where individuals are encouraged to contribute geographic data to a common platform. The data is then put through a process of verification before being incorporated into the database. An excellent example is Google Map Maker which is now available across several countries. Neo-geography spatially enables communities and helps them to participate constructively in governance and development. Spatially enabled communities can ensure better services like for example municipal services and also contribute to community friendly development of their environment, like location of recreation areas, health facilities etc.
GI SCIENCE
All these developments require research beyond the confines of traditional geography. For example, population studies treat figures as static numbers but population density can change drastically in the course of a day, for example in an area around a school. Missing these variations can result in improper resources planning, for example inadequate parking area for vehicles or inadequate public transport. Studying such problems require innovative computational methods and algorithms. Map algebra takes on a new meaning with raster processing algorithms. Geostatistics can be run on digital maps and the results visualised. An entirely new field, geospatial information science has developed to look at these issues. If we look at geography in this context, then many new research areas emerge. Some of these are :
• Ontology of the geographical domain
• Representation of geographical phenomena
• Qualitative spatial reasoning
• Computational geometry
• Indexing, retrieval and search in geographical databases
• Spatial statistics
• Map algebra
• Cognitive models of geographic phenomena
• Human interaction with geographical information and technology
• Acquisition and quality of data
• Spatio-temporal modelling, analysis
• Societal impact of geographical information
TOOLKITS
The geospatial professional has a range of tools at his disposal to configure these systems. GIS software has gone the server way. For example, ESRI’s ArcGIS Server is fully standards compliant and interoperable with other systems. Integration with SAP, SQL databases like Oracle, PostGres and DB2 are built in. The server can provide coarse grained SOA Web services using SOAP API as well as fine grained application development like Microsoft Dot Net, Visual Studio and Java. OGC compliance is built into all GIS server products whether it is from ESRI, Bentley or Autodesk. The latter has a product Autodesk Map, which has an Open Source versionand a priced product. The Open Source (OS) version enables developers to contribute modules, that find a place in commercial version. OS seems to be a mantrathat is catching on. Open Source Geospatial, OSGeo, has a strong following and many mashups are coming out of the OS world. However, Open Source is not for the faint-hearted or for those looking for quick solutions.
Having said that, I must add that some of the best OGC compliant map servers for WMS and WFS are available in Open Source. The interesting fact is that OSGeo supported products work on both Linux and Windows. This should appeal to those who are not comfortable with Linux. An interesting development is the release of ITC’s ILWIS as an open source product supported by a commercial entity, 52North. ILWIS is a product that combines both GIS and image processing functionalities.
Similar integration can be seen in commercial products like ERDAS, ERMapper and many others. This integration is a pointer to the importance of high resolution imagery in GIS. Such imagery provides most current geospatial information and during events like disasters they are the only authentic information available to the rescue and rehabilitation teams. Google Maps and Microsoft VE depend heavily on imagery. Google Map Maker provides high resolution imagery in the background to enable neo-geographers to identify and input information. High resolution stereo imagery is required for high level 3D mapping using digital photogrammetry. Digital aerial stereo imagery has replaced conventional photography. Though satellite stereo imagery is also available, studies show that aerial imagery still has an edge. Software which automates the photogrammetry workflow is increasingly getting integrated with the data acquisition itself. This provides rapid turnaround of the data from acquisition to application.
What next?
Forecasting is a dangerous proposition! However, I think it is safe to predict that GIS will be embedded to the point of becoming ubiquitous. Both high tech professional applications and popular social applications will thrive and perhaps feed constructively into each other like Microsoft’s Photosynth. People and professionals will discover new applications – many which we may not have even dreamt about! The concept of place and time will enable the location of all objects from a house to the house pet in a geospatial framework. As Franz Leberl puts it, people will need to add computational thinking to the three R’s.
