This course covers both foundations and practical aspects of the automated analysis of programs, which is becoming increasingly critical to find software errors and assure program correctness. The theory of abstract interpretation captures the essence of a broad range of program analyses and supports reasoning about their correctness. Building on this foundation, the course will describe program representations, data flow analysis, alias analysis, interprocedural analysis, dynamic analysis, and symbolic execution. Through assignments and projects, students will design and implement practical analysis tools that find bugs and verify properties of software.
This course fulfills the Logic and Languages constrained elective category for the Computer Science major as well as the Theoretical Foundations requirement of the Computer Science master's degree.
Why take this course?
- Explore the meaning of programs. One of the most basic questions that programmers ask is "What does this program do?" Program analysis is all about understanding programs--automatically!
- Learn deep theory. The theory of abstract interpretation stands with type theory as the most important and beautiful foundations of programming languages. Abstract interpretation is the fundamental theory of abstraction: how to precisely relate the concrete execution of a program to an abstraction of that execution. Naturally, this has many applications, which brings us to the third reason to take this course:
- Build awesome tools. Using program analysis, you can build tools that find bugs, prove important security and correctness properties, automatically generate useful tests, and much more--and you'll have a chance to do all of this in course assignments and a project that you can design yourself (if you want).
Logistics and People
(all times below in US Eastern Time; see Canvas for Zoom meeting links)
Office hours: Mon 8-9pm via Zoom
Course Syllabus and Policies
The syllabus covers course learning objectives, supplemental textbooks, assessments, late work policy, and policies. Consult Canvas and especially Piazza for up-to-date announcements and discussion (links in header).
As of January 2021, we have endeavored to make this schedule, including assignment due dates, as accurate as possible. However, it is subject to change as the semester advances. "Hybrid" modality indicates in-person + remote.
|Date||Topic||Reading/Material||Assignments Due||Modality||Optional Reading|
|Feb 2||Introduction, Program Representation, and Syntactic Analysis||Text ch. 1-2, slides||Remote||PPA ch. 1|
|Feb 4||Program Semantics||Text ch. 3, slides||Remote|
|Feb 5||recitation CodeQL||notes||Remote|
|Feb 9||Dataflow Analysis and Abstract Interpretation||Text ch. 4, slides||Remote||PPA ch. 2|
|Feb 11||Dataflow Analysis and Abstract Interpretation (contd.)||Text ch. 4, slides||hw1 hw1||Remote||PPA ch. 6|
|Feb 12||recitation Semantics||exercises, notes||Remote|
|Feb 16||Dataflow Analysis examples||Text ch. 5, slides||Hybrid|
|Feb 18||Dataflow Analysis examples (contd.)||(see previous)||hw2 pdf mathpartir||Hybrid|
|Feb 19||recitation Specifying/Implementing Dataflow Analysis||Remote|
|Feb 23||No class|
|Feb 25||Dataflow Analysis termination and correctness||Text ch. 6, slides||Hybrid||PPA ch. 4|
|Feb 26||recitation Correctness of Program Analysis||exercises, notes||Remote|
|Mar 2||Widening Operators and Collecting Semantics||Text ch. 7, slides||hw3 pdf||Hybrid|
|Mar 4||Interprocedural analysis||Text ch. 8, slides, simplified algorithm||Hybrid|
|Mar 5||recitation Introduction to Soot||notes||Remote|
|Mar 9||Context-Sensitive Analysis||Text ch. 8 & ch. 10.2 slides, simplified algorithm||hw4 pdf||Hybrid|
|Mar 11||Pointer Analysis||Text ch. 10 slides||Hybrid|
|Mar 12||recitation Midterm review||Midterm Study Guide||Remote|
8am take-home exam released
|Mar 16||Hoare-style Verification||Text ch. 11, slides||hw5 (checkpoint) pdf||Hybrid||PPA ch. 3|
|Mar 18||Control-Flow Analysis for Functional Languages||Text ch. 8 slides||midterm
take-home exam due
|Mar 19||no recitation|
|Mar 23||Hoare-style Verification (contd.)||Text ch. 11, slides||Hybrid|
|Mar 25||Symbolic Execution||Text ch. 13, slides||hw5 pdf||Hybrid|
|Mar 26||recitation||exercises, notes||Remote|
|Mar 30||Concolic Testing||Text ch. 14, slides||Hybrid|
|Apr 1||Fuzz Testing||Text ch. 16, slides||hw6 pdf||Hybrid|
|Apr 6||Satisfiability Modulo Theories (SMT)||Text ch. 12, slides||Hybrid|
|Apr 8||Program Synthesis||Text ch. 15, slides||Hybrid|
|Apr 13||Program Synthesis||hw7 pdf||Hybrid|
|Apr 15||No class|
|Apr 16||no recitation|
|Apr 20||Program Repair||Hybrid|
|Apr 22||Program Repair||hw8 (incl. project proposal) pdf||Hybrid|
|Apr 27||Model Checking||Hybrid|
|Apr 29||Model Checking||Hybrid|
|May 4||Gradual Verification||checkpoint for project||Hybrid|
|May 6||Dynamic Analysis of Concurrent Programs||Hybrid|
|May 14, 1—4pm||Project presentations||Project presentations||Remote|
|May 17||project final report due|