Customizing an Open Source ERP System

Abstract

At a point in time when the use of computers and software within companies and corporations is ubiquitous there is still reason to question their purpose and validity. The mere use of computers is not enough to enable efficient management of our time and work processes. Off-the-shelf software has a tendency to streamline organizations into facsimiles, leaving no room for developing and defining superior business processes. Refining business processes while figuring out what the true value of our endeavors is, be it in the manufacturing of goods or providing a service is, I believe, a path that is remarkably rewarding for both market and staff, as the cost of a commodity may begin to approach its perceived value. The business process becomes a true competitive advantage. This thesis explores the century-old, but often forgotten, models of business process refinement, combining them and the theory behind them with modern era business software into a system that is easily customized to fit any organization. A business process can be characterized as the description of how to complete a certain task. A process has a starting point and an end point with subtasks in between. Depicting the process in the form of a graph is a natural progression from that fact. In this thesis I create a software application for this purpose, storing the graph in a graph database. Since this type of graph has a natural path from starting point to end point it is possible to follow the graph programmatically from start to finish, in other words it is possible to use the graph as a form of source code, given that the correct parameters are available on the graph. Some tasks may involve decisions that are hard to program due to their circumstantial nature. In such cases the process must be halted and a notification be sent to whomever is responsible for allowing the process to continue. As part of the system a workflow engine is developed that reads the graph and performs tasks as they are encountered. In the simple process given as an example in this thesis, the performed tasks are all functions in an open source ERP software, but the implied possibilities for automation are endless, as long as the task can be performed by a software programmable device. With this system I show that it is possible to automate and orchestrate an off-the-shelf software system in a way that is less limiting with regard to business processes than a standard system, without the enormous effort to create a custom system from scratch.