Management Information Systems
Sunday, April 17, 2005
To avoid any misunderstanding it is probably not superfluous to define what we mean by "Management Information Systems" (or IS, for short). This term is generally used about most of what comes to mind regarding how computers can be used in a firm: to compute payroll, process banking transactions, airline tickets, orders, invoices, etc. But it normally leaves out the likes of rocket control computers, industrial robots or electronic car ignition controllers, usually grouped into the categories of "Embedded" or "Automation" systems. To make this a little more clear cut, we will define Management Information Systems as data processing applications interacting solely with humans through screens, keyboards and printouts (or sometimes voice-mail) to the exclusion of those which interact directly with a physical process through sensors and actuators.
This distinction is important because the impact of these two categories of systems on business practices is completely different. Industrial robots and embedded systems have had a tremendous impact on productivity and on the very nature of many industrial products. But this impact has been largely restricted to the design bureau and the factory floor, outside of which it has not been felt so much. On the contrary, the influence of Management Information Systems has been pervasive, though less easy to pinpoint and measure.
One consequence of this definition is to highlight the link between the philosophical question of meaning and IS practice. Computers and humans interact by exchanging signs and only signs. So what goes on in such a context depends primarily on how meaning is given to signs, or fails to be given, or is lost, etc. Post "linguistic turn" 20th century philosophy also happens to focus on these issues. Progress, in the analytic tradition, has often been understood as the ability to propose a new, and better, account of how meaning attaches to signs, or to objects. Similarly, in the discourse about MIS practice, it has always been taken for granted that success is linked with the ability to capture the meaning of the signs being processed so as to reach the best possible understanding of the rules governing their use. Both analytic philosophy and IS practice share a sort of "semantic optimism": a belief that semantic analysis of natural language is both possible and able, by itself, to avoid vagueness and resolve ambiguity. IS professionals are great believers in clarity and formalism. They hold logic in great esteem, as a way leading both to accuracy and to justice.
Development Methodologies
Monday, April 18, 2005
When computers were first introduced into a limited number of corporations during the 1950s, there were no preexisting guidelines determining how people should perform the task of programming them. Therefore, this was done on a case by case basis, in the way that seemed most natural to the people involved. Later on, these early experiments were codified into a set of concepts whereby a project was divided into a number of successive phases : requirements analysis, specification writing and approval, coding, developer testing, acceptance testing. Through the years, many variations have been tried around this theme, some minutely detailed and formal, others nothing more than vague guidelines of the "this-is-the-way-we-do-things-here" type. All these variants are grouped under the generic name of "waterfall" methods, and still represent the dominant approach followed in most MIS organizations.
Alternative approaches were attempted as early as the mid 1980s and have kept up appearing since, under various names. The most widely known of them are Rapid Application Development (or RAD) and Extreme Programming.