[Unknown Target]

Give every object a bounding box and slice said box through the middle of each face to generate the axis?

[Nuke]

well the solid-body is more of a point with dimension (derived by cross sections), though a bounding box would be drawn to indicate the position, approximate size and orientation of the solid body. its not a physical model, just an approximation of some object's (or rather part of an object) aerodynamic properties. this would allow for flat objects such as wings at off angles if needed. this also provides a system for things like swing-wings or canards which rotate an entire static airfoil to generate control or some other function (trade off low drag for low speed lift in this case), rather than augmenting the cl to control lift. the point of the matrix is to be able to change frame of reference between the parent, and an angled airfoil. take the swing-wing concept into consideration, when you transform it into its delta configuration, the airflow is generated in its frame of reference and its cross section is computed with respect to airflow. presenting a larger cross section would increase drag, where presenting a smaller one will decrease it.

[Unknown Target]

I was thinking that the "solid body" would be a subobject; so multiple lifting bodies can  be taken into account, all referencing the lift direction as perpendicular to the x axis of the specific sub object. So the entire aircraft would be one object group, with, say, the wing as a subobject child. This would also allow you to also define what the wings are in the 3d program through simple naming conventions.

If you wanted to add an intermediary program, you could have it slice up the wing into sections and analyze the shape used to create a rough airfoil curve.

For simpler purposes you could define the airfoil used and it's relevant data (CoL, CoD, etc) for different subobjects via a table file.

Also, how do I go about installing your flight dynamics?

[Nuke]

well i intended to use existing models for my testing purposes. i had not intended to make a mod out of it, but a resource for modders who wanted to do atmospheric/newtonian stuff. so i had intended to keep the atmospheric model separate from actual geometry, for testing and potentially in practice. i was actually considering making it possible to bind a subobject to a solid model so they could share animation data (like for canards and other control surfaces), but i was not going to require a solid body to have geometry for the system to work. for example using a single submodel for the whole craft (such as the stock fs models it uses now), while having solid bodies denote wings, stabilizers and control surfaces, fuselages, and other parts of the ship. in the case of a bound solid body the subobject and solid body would share position and orientation data.

to get the flight model, you need to downlad it wit tortoise svn. pretty much the same thing you download the freespace source with. then checkout from here:
http://nukedspace.fsmods.net/svn/trunk/atmospheric_flight_model/

alternatively i decided to compress it and put it here:

clicky

[Unknown Target]

I could provide you with a test model if you wanted?