Instructor: Pavel Panchekha, [email protected]
Assistant: Taylor Allred, [email protected]
Lecture: Mon/Wed 3:00–4:30 in WEB L122
Appointments: Pavel: https://shorturl.at/dzDJN; Taylor: https://calendar.app.google/WJaC7S5Nk29Gx8fy6
Office Hours: Thu 12:00–1:00 in MEB 3115
Discord: https://discord.gg/fHvWXgdFv2
Github: https://github.com/utah-cs4470-sp23/class

The goal of this class is to teach you how a compiler works, including all of the major components from front to back. By the end of this course, you will be able to:

• understand concepts in lexical analysis, and implement a lexer
• understand how a programming language is represented by a grammar
• understand LL and LR parsing, and write an LL parser
• understand concepts in type checking, and implement a type checker
• understand target-independent program optimizations, and implement several relatively simple optimizations
• understand code generation, and implement generation of x86-64

This is a programming-intensive course: to understand compilers, you will you'll write a compiler for a simple language called JPL. Prepare to spend a significant amount of time programming.

This document is the syllabus.

The JPL specification details JPL: its syntax, semantics, and some simple implementation tips.

The x86_64 Assembly Handbook details the assembly code you'll need to know to write your compiler. You don't need to read this until we get to the relevant part of the course.

You'll find the assignments for this class linked above as they are released.

These and any other materials developed in this class are always works in progress. If you see typos, inaccuracies, or anything confusing, let the instructors know (on Discord).

No textbook is required, but if you find textbooks helpful while studying, I hear good things about Engineering a Compiler by Cooper and Torczon.

If you found lecture confusing, don't understand a homework assignment, or are stuck debugging some issue: get help. Think about it: if you've been stuck for three hours on a problem, how likely is the fourth hour to help? Getting help from others is an important part of real-world programming and developing that skill now is important.

There are three ways to get help in this class.

You can always post on the class Discord channel, linked at the top of this syllabus. There, you can get help from other students or the instructors. This is a great place to get clarifications, coding help, or quick sanity-checks of your solutions. Naturally, do not post complete solutions to homework problems or copy huge chunks of code there (see the academic misconduct policy below), but short snippets are typically find.

You can also join office hours with TA Allred; the time and place is given at the top of this syllabus. Office hours are a good way to check that you're approaching the homework the right way, get a quick code review, or have someone help you out debugging a problem. We recommend you show up to office hours whether or not you're having problems, and just work on the homework while there. That way, TA Allred can answer any questions that show up, or you can help out other students if they run into a problem you've seen. This will be particularly helpful once we get to generating assembly code.

Finally, you can meet with Prof. Panchekha by appointment, at the time and place given at the top of this syllabus. This meeting is a great way to do an in-depth code review, clarify anything that's confusing or that you didn't understand in class, or discuss any larger issues you're having in the class. Prof. Panchekha may also reach out to you and request you make an appointment. You're not in trouble---Prof. Panchekha likely wants to discuss what the instructors can be doing to help you in this class.

Class periods will be used for lecture and discussion. Often, we'll do some live coding as well. You will be required to attend. Class time will not be recorded; if you must miss class, please let the instructor know and arrange with a classmate to share notes.

It’s really hard to learn by only listening: people mostly learn by participating. We’ll have discussions during class time and online and students should participate in these. During class, instructors will call on students to answer questions and to present code from their compiler implementations to the class.

There will also be a short Canvas quiz after every lecture. The intention of these quizzes is to help you recall the major topics and ideas that you are learning in class.

No one participates in a discussion if they will be shamed or embarrassed for it. All students must uphold a positive, welcoming, and supportive environment during discussions both in class and online. Remember that every piece of code has to debugged before it works---in class, you'll be participating in that debugging.

You will have a weekly homework assignment. These will be cumulative, implementing a compiler for a new language that the instructors have created for this course. You’ll spend a lot of time testing your implementation to make sure that is actually works as intended. The instructors believe that writing compiler code is the most effective way to learn about how compilers work.

Homework will be submitted and graded via Github. Assignments will be due at midnight at the end of Fridays. However, we will automatically grant a 48hr extension (without any grading penalty). Because the assignments are cumulative, if you fall behind, it's very difficult to catch up, so we will not accept submissions that are later than that.

CS 4470 will follow the School of Computing’s policy, please read it carefully. Basically, it's OK to discuss the assignment with other students at a high level, but in general it is not OK to share code or look at another student's code, or similar code found online. It is the policy of the School of Computing that academic dishonesty results in a failing grade for the course. The School of Computing has a two-strikes policy, where two instances of academic dishonestly lead to expulsion from the major.

Grades will be assigned on a standard 90/80/70/60 scale, but the instructors reserve the option to curve raw scores before assigning grades if they judge that (for example, because assignments were more difficult than intended) students’ numeric performance is not representative of the grades they deserve.

Components of course grades are:

There will be no exams, final or otherwise.

Additionally, a couple of students who prove particularly helpful to others in class or online will get up to 5% extra credit at instructors' discretion.

Note: This schedule is aspirational---a guide for our course, not a promise that will be kept no matter what. It will be updated and changed as the course continues.

Other parts of the syllabus can also be changed by the instructor with reasonable notice.