Answers to \Which programming language is the best?” can ignite fierce arguments among zealots who see no reason for any language other than the one they use. Fortunately, most programmers do not hold such extreme opinions, and often would like to see a rational evaluation of different languages from which they can draw their own opinions. Many researchers have proposed methods for comparing and evaluating languages 2, 9, 10], but they tend to focus more on the languages than on the needs of language users. Although languages are intrinsically worthy of study, their real purpose is as tools in problem solving.

Users do need to know the strengths and deficiencies inherent in a language, and how well a language applies to an application domain. But, even within an application domain, requirements for two distinct projects may vary widely. One product may have to be highly reliable and portable, while another may have to be extremely efficient. Thus, knowing how well a language supports an application domain may not su ce we also need to know how well a language supports the needs of particular projects within a domain. We need to be able to evaluate languages for their applicability to a specific project. This paper proposes a language evaluation method with just that focus.

In the following sections, we review major categories of programming language evaluation criteria, and propose an evaluation scheme that could help software developers determine the best language for their particular task. Finally, we reflect on the significance of our proposal. Language Evaluation Criteria. To help us better sort and understand proposed measures of language goodness, we settled on four major categories. Three contain criteria that could be applied even if humans never read or wrote a line of code. The fourth category contains criteria for those characteristics that enhance or impede use of a language by human beings.

Download pdf A Project-Based Approach to Programming Language Evaluation