Due to my limited understanding, it seems like a solution to launching would come in the form of an "integrated SEXP," which would probably need to be added to the engine. FRED2 already allows for a ship arrival from a launch bay/hangar. Since I've not experimented with doing those operations yet in FRED2, I'm not entirely sure of what the absolute capabilities of that arrival cue are (thankfully you are... why am I talking?).
The problem with the Galactica or any other Battlestar is that it has multiple launch tubes. I assume for a MOVING Battlestar/other ship, the arrival cue (linked to an event in the "events editor" or merely a timed sequence) would have to slave a given ship/wing to the capship in question; these fighters, bombers, etc. would lauch from and appear from the ship the ship in question regardless of where they might be placed in the FRED environment. Next, the launch type would be selected; this would be dependant on the ship class (the carrier vessel and the ship being launched). If "catapult" would be selected in the SEXP, then a further condition to set would be the catapult number (this would be an inherent programming factor in the ship model itself and would determine which catapult the craft in question would launch from). Finally, one would select the "acceleration factor" I suppose this would take care of the "arrival cue" part of the "integrated SEXP." Note that this is primarily referring to launches from the 'cat as well.
Next, for the second part of this "integrated SEXP," there would be certain factors used in the operator which FRED already is capable of that the user would not be capable of changing. This would be in the form of XYZ-programming, except that the arithmetic would be automatic. From the capship already selected to launch the fighters in question, the speed of the capship in any particular direction would be translated to X, Y, and Z velocites. The fighter would be held as a rigid body along the line of the catapult, unmoving in the capship, perpendicular to the capship's XYZ-velocites. Similar to some training missions, the player as well as the AI would be unable to move the ship until free of the tube (this could be managed by an inherent factor determined by the launch ship). When launched, the "automatic arithmetic" of the operator would use the capship's perpendicular XYZ-velocites (you would have to use rotational velocities as well) along the line of the catapult to accelerate the fighter away from the ship's insides along a straight (though possibly turning) vector. Once the ship was free of the carrier, the player/AI would regain control of the ship. The ship would decelerate as neccesary (if neccessary) and compensate for any sliding movement from the vessel turning. If I could DO something, I think this would be my approach.
Of course, if it is possible to do the same thing with velocites along a pre-determined line (or if the possibility exists for such in the future), it may possibly be more simple to create a line (which is not drawn in-simulator) along a catapult of the capship to which a fighter would adhere from its axis by some offset. The craft would not be able to move until free from the tube and would accelerate along that path until free. This might be that translational movement as implied before...
-Thaeris