DAVE-ML is intended to convey an entire flight vehicle dynamic simulation package, as is traditionally done with engineering development and flight training simulations. It is intended to allow a programming language independent representation of the aerodynamic, mass/inertia, landing gear, propulsion, and guidance, navigation and control laws for a particular vehicle.

Traditionally, flight simulation data packages are often a combination of paper documents and data files on magnetic or optical media. This collection of information is very much site-specific, and is often incomplete. Many times, the preparing facility makes assumptions about the knowledge the receiving facility has about the way the preparer's simulation environment is structured; these assumptions are not always true. As a result, the "rehosting" of the dynamic flight model can take weeks if not months as the receiving facility staff gets their hands around the contents and arrangement of the data package, the model structure, the various data formats, and then spends additional time running check cases (if they are lucky enough to have received any) and tracking down small differences in implementations.

There are obvious benefits if this tedious, manual process could be somewhat automated. Often, when a paired set of facilities has exchanged one model, the receipt of another model is much faster, since the receiving facility will probably have built some computer scripts and processes to convert the data (both model and checkcase data).

The purpose of DAVE-ML is to define a common exchange format for this data. The advantage gained is that any simulation facility or laboratory, after having written a DAVE-ML import and/or export script, could automatically receive and/or generate such packages (and updates to those packages) extremely quickly from other DAVE-ML-compliant facilities.

To accomplish this (lofty) goal, the DAVE-ML project is starting with the bulkiest part of the most aircraft simulation packages: the aerodynamic model. This initial version of DAVE-ML can transport a complete aerodynamics model, including descriptions of the aerodynamic build-up equations and the data tables, as well as include references to the documentation about the aerodynamic model. This format also lends itself to any static subsystem model (i.e. one that contains no state vector) such as the mass & inertia model, or a weapons loadout model, or perhaps a navigational database. The only requirement is that model outputs can be unambiguously defined in terms of inputs, with no past history information required.