Examples are important tools for programming education. In this paper, we investigate desirable properties of programming examples from a cognitive and a measurement point of view. We argue that some cognitive aspects of example programs are "caught" by common software measures, but they are not sufficient to capture all important aspects of understandability. We propose a framework for measuring the understandability of example programs that also considers factors related to the usage context of examples. Research shows that examples play an important role for cognitive skill acquisition. Students as well as teachers rank examples as important resources for learning to program.

Students use examples as templates for their work. Therefore examples must be consistent with the principles and rules of the topics we are teaching; free of any undesirable properties or behaviour. By repeatedly exposing students to "exemplary" examples, desirable properties are reinforced many times. Students will eventually recognize patterns of "good" design, gaining experience in telling desirable from undesirable properties. We therefore need to be able to tell apart "good" and "bad" examples. To do so objectively and consistently requires some kind of measurement instrument.

In this paper, we describe the development and initial validation of an evaluation instrument for example programs, based on three aspects of quality; technical quality, object-oriented quality, and didactical quality. Validation was performed by six experienced educators using examples from popular introductory programming textbooks.

Results show that the instrument helps indicating particular strengths and weaknesses of example programs. In all but one example program, we identified aspects that need to be improved. However, inter-rater agreement is low. The instrument can therefore not be considered reliable enough for evaluations on a larger scale. Further work is necessary to fine-tune the instrument and provide guidance for its usage.