This paper presents a study of computer based text development processes performed by ten native English language writers. The aim of this study is to analyse the text development process at a micro-level using a computer to explore whether unique keystroke patterns exist that can be used to identify individuals and different writing tasks. Keystroke logging techniques have been used together with fuzzy-logic comparison engines to identify unique keystroke patters across different tasks. Results indicate that patterns can be defined that are stable within writers but different between writers. Finally the paper suggests future improvements to the comparison engines and discusses ethical issues that need to be considered if such an identification mechanism were to be introduced into online assessment systems.