JFIT: a framework to obtain combined experimental results through joint fits

A framework is presented for obtaining combined experimental results through joint fits of datasets from several experiments. The JFIT framework allows such fits to be performed keeping the data separated, in its original format, and using independent fitting environments, thus simplifying the process with respect to data access policies. It is based on a master-server architecture, using the network communication classes from ROOT. The framework provides an optimal way to exploit data from several experiments: it ensures that correlations are correctly taken into account and results in a better determination of nuisance parameters. Its advantages are discussed and illustrated by two examples from the domain of high energy physics.