The source files in this directory create osl, a solver based on IBM's OSL (Optimization Subroutine Library) that handles linear, mixed-integer, network, and quadratic optimization problems. You need to obtain OSL from IBM or OST: see http://ism.boulder.ibm.com/es/oslv2/features/welcome.htm If your float.h doesn't #define DBL_MAX to be a constant, compile osl.c with Bad_DBL_MAX #defined, e.g. by adding -DBad_DBL_MAX to the CFLAGS assignment in the makefile. Directory osl1.2 contains source for the OSL 1.2 AMPL/OSL driver. OSL 1.2 is no longer sold, but some people still use it.