I agree with GTAuron that the railway deck should have a fence so it won't be as frustrating to drive along. Railway bridges have fences in reality to improve safety for maintenance crews, so it would be realistic. The fence should be very simple, probably just thin metal posts with alpha mapped railings. The collision mesh wouldn't include gaps for individual railings as that would add loads of polygons; could just have one face (two triangles) between each post.
The fence should be fairly low, at most one metre high. The collision file for it could be very simple, just a vertical face running the length of each railway section. Making thin collisions for the wire and the posts would add a lot of polygons and could make vehicles get stuck in them during collisions.
|QUOTE (JasonB @ Feb 19 2006, 03:33)|
This isn't quite what I meant. With this design, you've made the middle of the concrete deck thinner than at the edges. There isn't any supporting framework in the middle and the supporting framework is offset to the outer edges of the pillars. This would make the middle of the bridge very weak, creating shearing stresses through the deck. In reality, the spans would be likely to split down the middle along their length when a train went over them!
Also, the support beam doesn't go up to the edges, so they aren't supported and the outer portion of the frames don't fix into the beam directly. The beam slopes into where the frames start, which would make it difficult to fix the other parts of the frames securely. The slope doesn't include the middle of the beam, so the beam is narrower at the point where it as the least support from the pillars beneath.
Obviously GTASA isn't reality but the unbalanced design of these supports looks weird, doesn't it? I didn't explain my initial ideas very well.
I could make some sketches of what I mean, scan them in and upload them. If JB is on MSN at the same time as me, I could work with him to create another beam design combining my initial ideas and his new ones. Then we could compare the new beam with the current one and see which is better. There will be an awful lot of these supports and they'll be visible from most of MI, so we should make sure they look totally feasible.
(EDIT) An alternative to the dense, thin framework might be a few massive girders
instead. This would save polygons but might look weird around the curves? Obviously the supports will be closer together during curves (especially the sharper curves) so it might be alright?
(EDIT2) I suck at drawing, so I've made a Lego model of what I mean instead. I'll take pictures tomorrow when there is some daylight, that way the faces will have better shading so the shape will be more obvious.
Edited by Cerbera, 22 February 2006 - 09:26 PM.