There has been healthy debate in the past decade or so about GI Systems versus GI Science (or as Wright et al (1997) called it: tool or science?). Today the debate is over and there is consensus about the value of both GI Systems and GI Science and a requirement to use both in tandem. GI Science allows us to consider the philosophical, epistemological and ontological contexts of geographic information and GI Systems provide the infrastructure, tools and methods for tackling real world problems within acceptable timeframes. It is possible to undertake GI science without GI Systems thinking and technology (for example using a pencil and paper, or a calculator), but it is very inefficient and unproductive. Similarly, using GI Systems without a clear understanding of the scientific context and key scientific methods will produce results which are suspect at best and may even be wholly inappropriate or even incorrect. 
GI Systems can be used to turn data into information using the many tools and geoprocessing methods common to today’s systems. GI science provides the underlying principles and contextual understanding of when to use the tools, how they work and the way to interpret the resulting information. More importantly, GI science gives us a framework and a workflow for allowing us to turn information into evidence and then knowledge. By applying knowledge successfully over multiple years, we can build our wisdom about the world. GI scientists are interested in examining the fundamental principles that underlie GIS, that is to say the basic models, methods and generally held tenants of geography and geographic information. They are also interested in using GI Systems in scientific investigations to create new geographic knowledge.
Source: GIS development magazine Jan-2010