Why do we need architecture for information systems?
My own first exposure to Architecture dates way back to about 1988, when I first learned about Information Engineering, as in the Information Engineering Methodology (IEM) then offered by James Martin & Associates. Its core was 3 layers of Architecture. First, Data and Function models made up the Information Architecture, describing the key concepts/entities of the business for which data was collected, and the functions that managed that data to be valid and useable. The catch-phrase was “Data + Function = Information”.
These architectures were not an end in themselves; they each played a role in leading to delivery of an integrated set of information systems meeting the needs of the business. The scope of these architectures was an enterprise, not one product or project. They showed that the systems of an enterprise supported different business areas, such that highly cohesive and de-coupled systems could be built, almost in parallel, and they would all work together as a portfolio of systems to support the information needs of the enterprise.
The architecture(s) of Information Engineering were a great discovery for me, but it became clear over time that they assumed, especially as automated by CASE tools, that other aspects of Systems would be handled outside their scope, such as user roles/security, or locations/networks.