1.Methodology or paradigm: The selection of the programming paradigm, which determines what should be taught, must precede the selection of the first language, which influences how to teach it (Esendal, 1994).

The paradigm defines the framework within which the students are taught particularly concerning their [url=http://iprojectmaster.com/final-year-projects-materials/COMPUTER%20SCIENCE]computer science undergraduate projects[/url]. Programming paradigms are differentiated by the concepts that they emphasize (Watt, 2000). Imperative programming emphasizes procedures operating on (unencapsulated) variables; object-oriented programming emphasizes methods operating on (encapsulated) objects; functional programming emphasizes functions operating on immutable values; logic programming emphasizes predicates operating on immutable values; concurrent programming emphasizes processes exchanging messages (Watt, 2000).

The related issue of a single language versus multiple languages is another critical decision. The number of languages to cover in a course or curriculum involves a trade-off between breadth and depth (King, 1992).

2.Real or Customized: Regardless of the paradigm, a choice must be made between professional-grade and customised languages. A professional-grade language is one used in industry and is taught in its entirety (with the exception of very advanced features), via a commercially available environment (Esendal, 1994).

A professional-grade language like Java or C++ provides computer science undergraduates with experience and a real-life environment. Any problems encountered are those that will be encountered in real life, preparing students better for their professional lives (Esendal, 1994).

A customized language is one developed especially for teaching purposes or one that is a subset of a real language (Esendal, 1994). Customized languages such as Haskell and Eiffel separate learning programming from learning the details of a particular language (de Raadt et al., 2003), minimizing the technical problems that may distract from the learning process. Once students learn the fundamentals, they can apply their programming knowledge to any language (Esendal, 1994).

3.Cost of changing: Some educators naively think that there is little cost in changing programming languages because there are no serious consequences if a decision turns out badly. Lee and Phillips (2002) point out, however, that considerable overhead occurs in adopting a particular language, including preparation of lecture materials, developing projects and student exercises, evaluating and learning development environments and installing the chosen language, requesting and evaluating textbooks, and training personnel. Such overheads indicate that care should be taken when choosing a programming language because that choice is likely to impact the educator for several years.

Other criteria are:
• Availability of textbooks
• Language’s stage in life cycle
• System requirements of student/academic/full version
• Operating system dependence
• Open source (versus proprietary)
• Academic acceptance
• Industry acceptance
• Marketability (regional and national) of graduates
• Easy to use development environment
• Ease of learning basic concepts
• Support for target application domain (such as scientific or business)
• Full-featured language (versus scripting)
• Support for teaching approach (function first, object first or object early)
• Object-oriented support
• Good debugging facilities
• Support of web development
• Support for safe programming