An approach for categorizing end user programmers to guide software engineering research

Over 64 million Americans used computers at work in 1997, and we estimate this number will grow to 90 million in 2012, including over 55 million spreadsheet and database users and 13 million self-reported programmers. Existing characterizations of this end user population based on software usage provide minimal guidance on how to help end user programmers practice better software engineering. We describe an enhanced method of characterizing the end user population, based on categorizing end users according to the ways they represent abstractions. Since the use of abstraction can facilitate or impede achieving key software engineering goals (such as improving reusability and maintainability), this categorization promises an improved ability to highlight niches of end users with special software engineering capabilities or struggles. We have incorporated this approach into an in-progress survey of end user programming practices.