ABSTRACT
The inclusion of error handling capabilities within GIS is often viewed as crucial to the future commercial and legal viability of the technology. Error handling has remained high on the geographical information science research agenda for close to two decades. By contrast, commercial GIS have arguably not benefited significantly from this research. Systems capable of storing, managing and manipulating spatial data quality are a rarity outside the research laboratory. There are a variety of contributory factors to this undesirable situation, including:
• the lack of agreement on exactly what constitutes “error” and how it should be collected and reported;
• the cursory nature of most metadata, providing only limited raw materials upon which error-handling routines might operate;
• the highly technical, statistical or specialised nature of many error handling routines,
thereby precluding their practical use for many general GIS applications and users; • the negative connotations of the word “error”, outside the research literature, which
may deter software and data producers from emphasising error-handling capabilities.
It is certainly not possible to provide a technical solution to all these difficulties. Indeed, Goodchild (chapter 2, this volume) questions whether the problems posed by error handling may in fact require a complete rethink of GIS design. However, in this chapter we argue that it is possible to develop practical error-handling software using existing technology that at least eases some of the difficulties sketched above. An architecture capable of practical error handing is presented in three stages. Section 5.2 reviews the development of an error-sensitive GIS, which provides core error handling functionality in a form flexible enough to be widely applicable to error in spatial data. Section 5.3 introduces the concept of an error-aware GIS, which aims to provide intelligent, domain specific extensions to the core error-sensitive functionality, enabling users to better use and understand spatial data quality. Section 5.4 explores the deployment of distributed systems technology as a stable bridge between the flexibility of the error-sensitive GIS functionality and the specificity of the error-aware extensions. The resultant system, applied to a telecommunications legacy data capture example in Section 5.5, is used to indicate the suitability and robustness of the approach.
