Lots of testing tonight (over 3hrs), and the results are… I now have more questions than answers! That said, I think I’m beginning to better understand how AI traffic works. For consistency, all tests were done repeating the same KPIE-KEYW flight I’ve been doing for a few days.
First test was with Just Flight’s FS Traffic. Aircraft parked all over the place and although few were actually disappearing after landing (only one A319, but it did manage to vacate the runway before vanishing on a taxiway), they were taking some… interesting routes from runway to stand! Overall, though, relatively successful with no floating aircraft. Since FST lets the sim control AI aircraft, the haphazard nature of the taxi routes came as no surprise, with the core traffic logic still being broken under SU5. The dodgy parking is also simply explained, by how the sim categorises wingspan — i.e. ‘small’ is anything between something like 17 and 26 metres, so an E-Jet can appear on a stand meant for a Citation and a dev will have zero control over it. Mods like BATC can circumvent this because they use AI averages — so if, in real life, an E-Jet has never parked on a stand reserved exclusively for bizjets, it won’t use that stand in the sim, regardless of whether it has a wingspan common to the ‘small’ category.
Next test was using BATC, but with FSLTL models prioritised (rather than FST). This one produced the most interesting (and pleasing) results, with no anomalies whatsoever. No floating planes, no weird taxi routes, and all traffic successfully taxiing to stand after arrival.
In conclusion, the whole issue at KEYW doesn’t appear to be down to BATC, per se, but is instead related to how different aircraft models are interpreted. Since most people use FSLTL for their type models, it seems BATC took those models as a base. I did notice that FSLTL models were using much less runway to slow down, so KEYW’s runway length could still be a stumbling block. A point worth noting is that both FST and FSLTL actually reduce the wingspans of their aircraft models to better fit airports that have less precise parking radii (though in FST, this is optional).
So going forward, it’s unlikely this issue will be addressed, since it would necessitate BATC, FSHud, SI, et al checking every model available in FST, FSLTL and AIG. It’s just not a realistic expectation. If they adapt for FST, they break FSLTL compatibility, and so on.
I will still raise a bug report with BATC, just in case there’s something they can do to better accommodate FST aircraft models. But at this point, I’m fairly certain the blame doesn’t squarely lie with BATC. I suspect it’s one of those little niggles we’re going to have to learn to live with (and believe me, having re-experienced the chaos of the sim’s core AI logic via FST, it’s an easy annoyance to live with!).
. My Community folder is full of these.