Automatic Generation of High-Coverage Tests for Complex Systems Programs(stanford.edu)
stanford.edu
Automatic Generation of High-Coverage Tests for Complex Systems Programs
http://www.stanford.edu/~engler/klee-osdi-2008.pdf
http://www.stanford.edu/~engler/klee-osdi-2008.pdf
It is a very useful approach to prevent regressions while working on legacy code.
Alberto Savoya (who is behind agitar) has a nice definition of legacy code : code that has no unit-test.