New course

		     6.893: TOPICS IN PROGRAM ANALYSIS
		       http://pag.lcs.mit.edu/6.893/

			    Prof. Michael Ernst

		Tentative meeting times:  MW 1-2:30, 38-166
		First meeting:  Wed., Sept. 5, 1:00, 38-166

This 12-unit graduate seminar examines a variety of program analysis
techniques.  Topics include static analyses (as typically found in
compilers and other programs that are provided source code), dynamic
analysis (as found in profilers and other tools that work over program
executions), and hybrid analyses that combine the two approaches.  Sample
analyses to be covered include model checking, theorem-proving, type
inference and checking, dataflow analysis, concept analysis, profiling,
test suite coverage, dynamic invariant detection, race detection, and
visualization.

The focus of the course is the development and use of tools based on
program analysis to enhance human capabilities and assist in program
development tasks, including design, implementation, testing,
documentation, and maintenance.  (An alternative would be to concentrate
on, say, optimization for the sake of faster execution, but people form the
real bottleneck in software development.)

In addition to weekly readings of research papers and small exercises that
provide experience with a variety of tools, the course includes a
substantial team project performed in groups of 2-3.  Sample projects will
be provided, but students are also free to propose their own.  Examples
include proposing and evaluating a fundamental new technique; developing
and assessing new algorithms to replace currently-used ones; translating a
methodology to a new problem domain; evaluation of proposed techniques (for
instance, via a user study); and applying known techniques to new problem
domains, such as operating systems, networks, embedded systems, security,
biology, aerospace, etc.

Prerequisite:  6.035 or instructor's permission.
The course awards Grad H and EC credit.