SIGCSE 2001 DC Application -- Peter DePasquale

SIGCSE '01 Doctoral Consortium Application - Research Summary
Peter J. DePasquale
Department of Computer Science
Virginia Polytechnic Institute and State University (Virginia Tech)
pjdepasq@vt.edu

"The application of a spiral approach to language utilization
within a programming environment in university/college curriculums."

Introduction - Throughout CS1 (1st semester programming) courses, syntactic language elements and programming concepts are generally introduced in an interactive manner, where concepts are built upon in successive iterations. Yet, while we introduce our students to these elements, we generally require them to use a development environment or production-level compiler that inadvertently assumes the student has a wider breadth of knowledge about the language then they do.
This work focuses on the application of a new type of programming environment for use in CS1 courses which layers these syntactic elements and programming concepts such that students are only permitted to utilize that which is specified by their instructor.
Previous research in the area - The only identified previous research in this area (directly related to the application of layering (spiral approach) the language) was an effort by the Rice (University) Programming Languages Team [1]. Their Dr. Scheme project developed a programming environment that layered the Scheme programming language. Student users of the software are bound to the layers dictated by the DrScheme developers. My research will differ because it will expand the environment to allow course instructors to create and set their own levels thereby mirroring the pedogical experience of the classroom. Additionally, I will be examining the concept of providing programming problems or assignments to the student, as defined by the instructor, via the development environment. Such problems would be delivered with an associated set of level information and would force the student to adhere to the language constructs that the instructor mandates.
Additional research has been conducted in the area of simplified programming environments for first time users (http://www.holtsoft.com/ready/home.html) as well as the examination of various coding/analysis environments which assist in object browsing [2] or the graphical representation of student's programming [3]. While these areas are related in terms of programming environments, I don't view them to be directly related, since they don't attempt to match the pedogical model of how students are learning a programming language in the classroom in CS1 courses within their environment.
Goals of the research - We seek to compare the results of current students who use the production level environments to students using a layered environment which enforces access to those concepts the student has been exposed to, as defined by their instructor.
It is conjectured that this approach will show measurable pedogical gains over standard development environments with the use of this new style programming environment, in part because the syntax and compilation error messages can now be tailored to the student based upon the knowledge of what concepts and syntax the student is permitted to utilize. It is also conjectured that the nature of the programming environment which a student uses impacts their comprehension of the language and programming in general.
In a larger sense, these efforts may very well impact future CS1 syllabus design. Such an environment may be used as a tool to force the student to "cope" with solving problems in a limited domain in an effort to understand the problem prior to using high level constructs to solve it. This idea is not unlike the restriction of not using tools such as lex and yacc in a compiler class during the development of a compiler/interpreter, so that the students can get an understanding of the fundamentals before using high level tools.
Current status - As of the date of this submission, November 19th, 2000, I am currently overseeing the development of a prototype environment by two undergraduate students for independent study credit. It is hoped that by the end of the fall semester 2000, we will have a small scale development system that allows for a student user to obtain a list of language constructs that are permitted for use, and which permits C language code to be entered, recognized and interpreted. The student's programming environment will enforce the accessibility to specified language contstructs and error messages will be tailored to the constructs that the student is permitted to utilize (and therefore should provide a higher quality of error messages than which production-level compilers currently produce).
Interim conclusions / Open issues - As of this date, none. Anecdotal evidence is in favor of our goals, and it is believed that this type of environment may indeed hold promise to help novice students move into the use of programming environments and learning to program.

Current stage in your program of study - As of the Fall semester of 2000, I have passed our department's qualifying (comprehensive) examination, and I am in the process of writing my Ph.D. research proposal with an eye toward presenting my preliminary defense in the spring (possibly prior to SIGCSE '01).
What you hope to gain from participating in the Doctoral Consortium - I'd like to continue to be motivated by the SIGCSE Doctoral Consortium. I was a participant in 1999 and I found the activities and feedback to be invaluable. In the last year, my research direction has changed somewhat and I'm looking forward to presenting my new ideas to the participants.
Bibliographic references -
[1] Matthias Felleisen, Robert Bruce Findler, Matthew Flatt, and Shriram Krishnamurthi. Functional programming: The DrScheme Project: An overview. ACM SIGPLAN Notices, 33(6):17-??, June 1998.

[2] Michael Kolling. Why BlueJ? An introduction to the problems BlueJ addresses. http://bluej.monash.edu/html/why.html.

[3] James Cross, Introduction to the Control Structure Diagrams, http://www.eng.auburn.edu/department/cse/research/grasp/