Yeah I was thinking on the same, I think the solution for that would be using 2 empties, one at each side of the character and the character tracking them depending on the last pressed button (left or right).
Decreasing to detection rate looks promising, back to the code.
Wow! Double posting…
Well, it isn’t useless because I found the reason of the corner glitch:
http://cloud.gl.googlepages.com/3DSideScrollerOriented.blend
It was because of the IPOs, compare this file with the previous one, in the IPO windows, the bright blue curve, solving that soved the problem.
Now what does the curve change means?
->The previous curve made a 360 rotation in very little time so it seemed that the box moved well untill that point, made a bug, then continued, so moving a piece of the curve in a way the movement where softer made disapear the bug. ^^
http://cloud.gl.googlepages.com/3DSideScrollerSoftlyOriented.blend
Hey 3DGuru, there you have it, softly oriented 3D sidescroller! The original box I leaved it as a “shadow” because didn’t wanted to bother erasing it and changing it with an empty, but the overall idea can be understood, right?
Triple posting… XD
WOW @_@ , nice update Cloud_GL … that’s somthing we can use man …
the members may use it a day …
thanks for blendenzo and Cloud_GL 
edit :- i mean the second one the softly oriented …
Wow, I didn’t expect to see this thread today…
Cloud_GL, that’s a nice set of changes you’ve made there! Glad to see you were able to work around my sloppy setup.
The only thing that I see that could still use improvement is that the character does not match its orientation to the ramp. That could probably be done using an invisible plane set to “shadow” and having the character copy its orientation, though.
Be my guest ^^
Yeah, not being automatically oriented may suck a little, but so far this demo plays and fell really well, several things can be made with it that could look impresive.
The only thing that really bugs me is how would we make the ramp split in a “Y” like intersection, doing it only verticaly isn’t big deal but in the plane may get nutcracking, the choice or the way normally is done be pressing up or down in the intersection, where one of the routes is the default (no up/down pressed).
It seems to me that the easiest way to do that would be to use multiple empties, each with their own path animations defined. Then at an intersection, use Python to switch which empty is being followed if the user presses up or down. The default could be to continue following the current path (the empty that the player is following at the time) unless a key is pressed.
Mmm, that sounds to be it. But all those extra empties will have to have the same animation in the frames where their path is the same as the default one, and all empties’s IPOs and locations of intersection must exactly fit, so the switching doesn’t become aparent. Probably because of this those 3D sidescroller have so little differents paths 
thanks for sharing such valuable views.
Wow, good stuff. :]
i wonder if this might be catched by anyone then even more updated … .
why didnt I see that awesome thread yet?!?!
Wow, this stuff is great! Great work Blendenzo and Cloud_GL.
To be honest, this is the best 2.5D setup I’ve seen yet.
oh, i think its time to post my 2.5d experiments (a multi layer system, without ipos or curves, just some special walls) but unfortunatelly i have little time, but as soon as i develop it a bit more i’ll post something
Wow this is pretty sweet . I’ve always liked platformers. Now let’s see what I can make.
Can I see some screenies?
Heeeeeeey! SPIFFEROO!!! No seriously… this is interesting… i will have to look into this…
thanks
terry
Funny how a 6 months old thread suddenly comes back to live this way XD
Well there aren’t any screenshots since it’s rather ugly looking (grey extruded cubes on blue background), the idea is to see it working :P.
This is really interesting, I’ve toyed with the demo a little and was originally quite excited about the potential of this approach. Then unfortunately I hit a brick wall (metaphorically) that I can’t find a way round. That is, how do you get the player to hit a brick wall!?
If you put any obstacles in the way, the method of movement just forces the player through them, and causes some headaches to the physics sim on the way.
Does anyone have any ideas how this could be overcome? I’ve spent some time thinking about it, and haven’t come up with a workable solution yet. Also, I was thinking about a game mechanic that involves parts of the scene affecting the movement of the player, i.e. swinging things that might project the player forward or back, again, can’t think of an easy way around this either. I suspect the solution to both of these is to change the movement mechanism to not explicitly assign the position of the follower, but instead make an attempt to get to the follower somehow.
Any thoughts appreciated.
Paul
cyborg_ar , that seems quit good , a blend ? , or not yet ?