Course News

• This course uses Piazza for class discussion and questions about class material. The site is located at the following URL: https://piazza.com/rice/spring2020/comp411/home. The Comp 411/511 Piazza discussion board will be monitored by the course staff and will also appear on each student's Piazza home page. The Piazza website is designed to quickly provide you with course help from classmates, the TAs, and myself. Rather than emailing questions to the teaching staff, I encourage you to post your questions on Piazza. If you have any problems or feedback for the Piazza developers, email team@piazza.com. I may experiment with moving our class web site to Piazza but the primary course web site is:
  
• Browser-based reference Jam interpreter developed by Nick Vrvilo
• Finding project partners:
  All projects in Comp 411/511 will be completed in teams of up to two students. Students may work alone but we recommend working with a partner so that you can audit each other's code. This process serves as a simple form of code walkthrough. In this course, your programming assignments will be graded on style and the quality of your unit tests, as well as correctness against our unit testing suites. We will provide class solutions for the first three projects that demonstrate what we consider to be good Java style, minimizing mutation and leveraging inheritance where appropriate.

  If you would like to find a partner for completing the projects this semester, we encourage you to use the Search for Teammates! feature on Piazza. The following are some issues to keep in mind when searching for a partner.

  • Ensure that you can work together on the project. Do not choose a partner with an incompatible schedule.
  • Ensure that at least one partner has Java programming experience, preferably with a good understanding of the core Java design patterns: composite, interpreter, visitor, strategy, and singleton.
  • Try to find a permanent partner for the entire class; it's a lot easier to work with the same student than to switch partners on each assignment.
  • You may choose to work alone. We do not allow groups of more than two students, even if there is an odd number of students in the class.
  • You are responsible for understanding all of the code that you submit for each project. You may by chance be asked to explain how it works..

Summary

COMP 411/511 is an introduction to the principles of programming languages. It focuses on:

• identifying the conceptual building blocks from which lanugages are assembled and
• specifying the semantics, including common type systems, of programming languages.

A secondary theme of this course is software engineering. All of the programming assignments in this course are conducted in Java using test-driven development and pair-programming, two of the major tenets of Extreme Programming. A compelling reason for using Java is that it is widely used for software development in industry and it supports low-level programming where appropriate. Java Virtual Machines are expensive to start (the Achilles heel of Java IMO) and Java applications are commpiled "Just In Time" (JIT) which has visible costs early in program execution. On the other hand, well-written Java code that has been "warmed up" (run on sufficient inputs to force the JIT compilation of all of the compute intensive parts of the application) is surprisingly fast.

COMP 411/511 consists of three parts:

• The first part focuses on specifying the syntax and the semantics of programming languages. The former is introduced via simple parsing (translations) of program text into abstract syntax. More detailed aspects of parsing (such as lexical analysis and advanced parsing methods) are left to COMP 412, a course on compilers. In practice, most domain specific languages do not need a sophisticated parser; in fact, so-called "internal DSLs" rely on the parser and compiler of the host language. Scala, a putative successor to Java that compiles to Java byte code, is specfically designed to support the easy implementation of internal DSLs. In addition, "external DSLs" (which are implemented by interpreters akin to the interpreter projects assigned in this course) often rely on simple manually written parsers based on "recursive descent" which is the approach to writing parsers taught in this course. The other attractive option for parsing external DSLs is to use a parser generator like ANTLR which is based on a simple extension of recursive descent parsing.

  In the course, you will implements the semantics of a comprehensive collection of programming language constructs using mostly functional interpreters derived from simple reduction (textual rewriting) semantics for the constructs. The constructs include arithmetical and conditional expressions, lexical binding of variables, blocks, first-class functions, assignment and mutuation, basic control constructs (loop exits, first-class continuations, simple threads), and dynamic dispatch.

• The second part illustrates how interpreters and reduction semantics can be used to analyze important behaviorial properties of programming languages. Two examples are covered: type safety and memory safety. Type safety guarantees that programs respect syntactically defined type abstraction boundaries and never raise certain classes of error signals. Memory safety guarantees that programs release memory if it is provably useless for the remainder of the evaluation. The standard technique for ensuring memory safety is automatic storage management, which is typically implemented using reference counting or garbage collection. Automatic storage management (which is present in all high-level programming languages including Java, Python, Javascript, and Swift) never deletes data objects that are reachable in subsequent program execution. Type safety typically depends on memory safety. We will study the formal type-checking process incorporated in the ML family of languages, which has been adopted in part by sophisticated OO languages like Scala and Swift.


• The third part shows how interpreters can systematically be transformed so that they use fewer and fewer language facilities. The key transformations are the explicit representation of closures as records and the conversion of program control flow to continuation-passing style. Using these transformations, a recursive interpreter can readily be be re-written in a low-level language like C/C++ or even assembly language. These transformations can be applied both to interpreters and to arbitrary programs. If we apply these transformations to an interpreter, the result is a low-level interpreter that is easily implemented in machine code. The process of applying these transformations to arbitrary input programs yields a high-level description of program compilation, a process which is explored in more detail in COMP 412.

This course material enables students to analyze the semantics and pragmatics of the old, new, and future programming languages that they are likely to encounter in the workplace (e.g., C, C++, Java, JavaScript, Swift, C#, Python). It also enables students to design and implement efficient interpreters for new languages or DSLs embeded in software applications. Most importantly, it equips students to become master software developers because they will be able to define and implement whatever linguistic extensions are appropriate for simplifying the construction of a particular software system.

My notes on object-oriented program design briefly describe the design patterns that I recommend using to express functional programming abstractions in Java. If you have little prior experience in writing functional programming code in Java, I highly recommend skimming them.

For students interested in operational (syntax-based) semantics, I recommend reading notes by Walid Taha (now at Halmstad University in Sweden) on big-step versus small-step semantics. I strongly prefer small-step semantics, so it is the only form of operational semantics that we will use in this class, but big-step semantics is often used in the programming languages literature.

Differences between COMP 411 and COMP 511

COMP 511 includes all of the material from COMP 411 plus a modest amount of supplemental material. Portions of projects 3 and 6 are mandatory for Comp 511 students but optional for 411 students, COMP 511 students are also required to complete 3 short written assignments over the course of the semester, discussing concepts from the lectures.

More Course Information

Please take a look at the course information/policies page, which is referenced in the first few class lectures.

Course Schedule

Language Resources

1. Java
   1. SDK Download
   2. API Reference
   3. DrJava Programming Environment
   4. Elements of Object-Oriented Program Design by Prof. Cartwright
2. Scheme (Racket)
   1. How to Design Programs
   2. DrRacket Programming Environment
3. Browser-based reference Jam interpreter

Additional References

1. Krishamurthi, Shriram. Programming Languages: Application and Interpretation. This book is a descendant of lecture notes created by Shriram for a version of this course when Shriram was a teaching assistant over a decade ago.
2. Friedman, Wand, and Haynes, Essentials of Programming Languages, 2nd ed. (MIT Press, 2001)
   You can take a look at the following two chapters, which the authors prepared for the second edition, without buying the book:
   1. Parameter Passing ( local file, PDF)
   2. Types and Type Inference ( local file, PDF)
3. Evaluation rules for functional Scheme ( PDF)
4. References on evaluating Jam programs
5. Lecture Notes on Types I
6. Lecture Notes on Types II
7. Introduction to System F (Polymorphic Lambda-Calculus)
8. Scheme code from Class Lectures
9. The Essence of Compiling with Continuations by Flanagan et al.
10. Uniprocessor Garbage Collection Techniques by Paul Wilson
11. Garbage Collection [canonical textbook] by Jones and Lins
12. Space Efficient Conservative Garbage Collection by Hans Boehm ( local file, PDF)
13. Hans Boehm's Conservative GC Webpage
14. JVM Performance Optimization. The first article in a five part series (with the remaining four parts linked from the first article) on JVM internals and how to write Java source code to use them efficiently.
15. Java Memory Model
16. Revised Thread Synchronization Policies in DrJava (doc) ( pdf). Since DrJava is built using the Java Swing library, it must conform to the synchronization policies for Swing. Unfortunately, the official Swing documentation is sparse and misleading in places, so this document includes a discussion of the Swing synchronization policies.
17. Lesson: Concurrency in Swing. This lesson discusses concurrency as it applies to Swing programming. It assumes that you are already familiar with the content of the Concurrency lesson in the Essential Classes trail.
18. Java concurrency (multi-threading): Tutorial. A tutorial on expressing concurrent computation in Java using threads.
19. The Last Word in Swing Threads. An article on the perils of accessing Swing components from outside the event dispatch thread.
20. Why Functional Programming Matters (as analyzed in 1990).
21. Why Functional Programming Mattered (gazing at programming technology in 2017).
22. Old Course Website

Accomodations for Students with Special Needs

Students with disabilities are encouraged to contact me during the first two weeks of class regarding any special needs. Students with disabilities should also contact Disabled Student Services in the Ley Student Center and the Rice Disability Support Services.