intentional glides

I've been wandering about this, there's an old demo thread about actually getting through it but has anyone figured out the magic formula (Aside from it being 32px wide and exactly NS/EW) for *creating* one in a map? Maybe it's just me and the ports I've used because I can never get through them.

I don't believe the exact circumstances are known, other than that the gap has to be at least 32 units and that it has to have parallell walls going straight east->west or north->south for a glide to be possible at all. I'm also not sure if glides work in boom compatibility, however I vaguely recall doing e.g. the Doom 2 map 16 glide in ZDoom.

An idea could be to send the map around to some people from the demo crowd and see if they can do it (I could take a look at it, for instance).

By dint of experience, I would say that the relatively easy glides are ones where there is a wall and a small square pillar 32 units away from it. For example, ksutra map32, darkcvnt map10. The hardest seem to be ones where there are two pillars that are symmetrical as you approach them (e.g. at the exits of TVR! map22 and Plutonia map19 - both undone AFAIK). Of course, until someone has done them, one can't even say for sure that they are possible at all.

More marginal cases seem to arise where there is some irregularity in the geometry. TVR! map13 (second glide - the first one is a classic case of an easy glide) is very difficult, as is the third glide on Extremal e1m7.

And of course, the difficulty of a glide decreases as you get more experience attempting it, and will depend on the method being used to attempt it.

But no, I have never seen a proper analysis of the mechanics of these glides. And maybe some ports have changed them. Use Doom2.exe (compat) if you're having problems with one you know to be possible. But also be aware that some of the glides you see in demos only worked after hours of attempts. :/

Links to some of those referred to in this post:
http://www.d.kth.se/~d99-kro/demos/ks32-020.zip
http://www.doomworld.com/sda/dwdemo/dc10p004.zip
http://www.emil.vbox.co.uk/doom2/tv13p008.zip
http://www.emil.vbox.co.uk/doom2/tv25p003.zip
http://www.emil.vbox.co.uk/doom1/x1t7m229.zip (actually what I refer to as the "third glide" is the second one here, as he wisely skipped the first one in this proof demo)

Scientist 2 map11 has a very easy glide which almost never fails.

Here we go

http://www.cyclomedia.co.uk/doom/wads/glides.zip

total glide possibilities: 24

12 short and 12 long
8 aligned with origin (0,0)
8 aligned with 128 grid
8 aligned with 16 grid (marked with dead imps)
6 cross two sectors
6 use doors instead of pillars
12 vertical and 12 horizontal

I couldnt get any to work in WinMBF

Playing around for a couple of minutes (i.e. not testing rigorously), I could do *all* north/south glides in doom2.exe compatibility, boom compatibility and MBF compatibility, and *none* of the east/west glides. The westernmost of the three northernmost glides I managed to get the cleanest glide I've ever seen in a map (it didn't even look like I was using a glide!) three times. Would you like me to send you demos of the glides?

UPDATE: Managed all of the glides with MBF compatibility (I'm sure it works with doom2.exe and boom compatibility as well). Interesting notes: The east/west glides only worked when running west, and none of those glides exhibited any wobbling whatsoever (which all the north/south glides did).

ooh, share!

IIRC the western N-S glides are at coords ( -128 , ... )

Here are the demos, I split them up into two parts: the 12 glides done north-south and the 12 glides done east-west.

glides

Recorded with doom2.exe compatibility, but it works exactly the same in Boom compatibility and MBF compatibility (for me, anyway).

I got it mixed up in my previous post, and have edited it accordingly. Here's what I found:
The north-south glides exhibit "wobbling" and work for me in both directions (which I show in the demo).
The east-west glides exhibit no "wobbling" and, in fact, seem to be dependant on something other than the alignment of the player. Once I've found the perfect angle (lining up at the pillar) I just go straight back and then straight forward, trying from different positions, and all of a sudden without notice I'm let through. I also could only make the east-west glides work when going from east to west, and never the other way around.

well a wholly unscientific look at these by using the automap and zooming in as far as you can go suggested that it only happens when you are exactly lined up with the gap and not a fraction of a degree out.

Next I tried running a quick custom version of winmbf to printf the player's (desired) speed at particular moments in time. which seemed to discount this. the first north-south one worked when (roughly):

momx = -300
momy = 400000

if my theory was correct that -300 whould have been zero, when you're bobbing up and down and not getting through the gap the y speed is approx 100000

needs more code analysis but i'm busy right now so will have to wait. a quick look shows that the wall (line) detection code runs lines from the corners of the player box to see what they hit and then attempts to slide the payer along whatever walls it encounters. at a guess in this example the glide works when the corner of the player that wants to slide along the wall (your left in the first NS glide) wins over the one that wants to stop dead on the front side of the pillar (to the right)