Making compiler construction projects relevant to core curriculums

Abstract

Having 300 students a year implement a compiler is a debatable enterprise, since the industry will certainly not recruit them for this competence. Yet we made that decision five years ago, for reasons not related to compiler construction. We detail these motivations, the resulting compiler design, and how we manage the assignment. The project meets its goals, since the majority of former students invariably refer to it as the project that taught them the most.

Design and Implementation of an Interpreter Using Software Engineering Concepts

“…Demaille [9] states that compiler construction is a challenging process that requires material from virtually all computer science courses on the core curriculum. While the idea of compilers is usually furthered and explored in detail later on in an upper level course such as Compiler Construction, Xing [2] argues that the idea of interpreters rarely gets the same "treatment": There is no such a course targeting on interpreter constructions in most undergraduate computer science curricula at universities and colleges [9].…”

A set of tools to teach compiler construction

“…To ensure that students correctly manage the memory, we promote the use of checkers [9]. They have revealed a well-known fact: it is impossible to properly reclaim memory during error recovery or when special actions such as YYACCEPT or YYABORT are used.…”

A set of tools to teach compiler construction

“…Parts declared "teacher only" are not given to the students. This tool helped us change the architecture to address the shortcomings spotted during the yearly debriefings with the students [9]. It also helps us change the assignment from year to year, to diminish the interest of cheating by stealing code from previous classes.…”

“…In [9] we introduced "yet another compiler construction project": the Tiger project. Contrary to its peers (e.g., [2,4,5,6]) its primary objective is not to implement a compiler, since "students will (most likely) never design a compiler" [8].…”

“…So over the years, to improve their educative value and/or to ease the management of such a large scale project, we improved several of these tools and created new ones taking student advice and requests into account. In [9] we skimmed over these tools; reviewers and readers asked for more details. This paper addresses their demand.…”