Just to get the correct impression: Are your comments meant ironically due to the complexity of jetways.nas?
I spent some time during the last weeks for checking opportunities for building ground services. There are already existing ground service components in FG providing Pushback, FollowMe, Ramp Marshall and Jetways (maybe even more). Integrating these with the existing AI system appears straightforward. Routes for driving vehicles on the apron could be derived from groundnet.xml.
I didn't succeed in understanding the existing AI system with its moving vehicles from waypoint to waypoint, neither by checking the code nor by building a sample scenario. The Nasal approach of tanker.nas however looks more comprehensive. Reading groundnet.xml from Nasal apparently isn't possible yet without doing explicit XML parsing. Though the airportinfo function in nasal provides parking and taxiway information, these might be derived from different sources than groundnet.xml.
Having the groundnet.xml information available in Nasal in combination with the logic in tanker.nas might provide a simple way for having vehicles moving on the apron.
But thinking further reveals more challenges:
1) The existing ground service components are based on different concepts. (the FollowMe vehicle eg. is a pseudo aircraft if I understand that correct). Integrating these different concepts might lead to the loss of some features/ideas/concepts or to a competition of concepts (often resulting in endless discussions).
2) Vehicles driving exactly on the groundnet will collide with others. The routes could be modified a few meters beside the route. But a pushback car for example needs some special route for reaching the front of an aircraft without driving through a building. All this might lead to complex geometry calculations.
Sounds daunting to me. However, I think I'll continue my current efforts reading groundnet.xml manually and moving vehicles on that routes.
Thomas