Reconsider solid ceiling in view of addition of shimmier? [DISCUSSION]

[Simon]

expand the level, repeating the terrain on the edges of the level infinitely.
Lemmings traveling to the left and right would continue
builder [...] will fall back down
makes sense for the sides to be deadly, but not the ceiling.

This is promising because it tries to remove the boundary instead of enforcing special boundary rules. The visible area will then define player influence instead of a living zone for lemmings. And it cries for an extended view once we get close to the ceiling, with a cutoff height for skill assignments.

But it has a catch. Lix has the infinite terrain extension. When you place the final builder brick under the ceiling, because of infinite extension, the brick creates a solid tower instead of a flat rectangle. The building lix will step onto the brick, thereby enter this tower and be stuck inside. Followers will turn at the tower and walk back down.

Since I don't like lixes stuck outside the visible area, I kill everything that's completely out of bounds. At least for the builder, the kill doesn't play a big role; the tower behavior comes entirely from the infinite terrain extension. 100 % logical from simple rules, but highly unintuitive. I had to test this in Lix while writing the post! Yet I'll keep it in Lix because rules should be simple, and level designs should leave the ceiling alone.

You can get around this with special rules for terrain that was always there vs. terrain that we created during play. But then we would heap special rules onto the ceiling despite our attempt to remove its specialness. Or, maybe special rules are best nonetheless.

Infinite terrain extension should kill climbers because they will never return.

I love that you brought this idea to the table. It's really the heart of the problem: The ceiling should matter less, not more.

ceiling's behavior doesn't usually result in backroutes

My hunch was the same 2 years ago. I haven't reconsidered since because the question hasn't come up since.

Icho's argument from 2 years ago was that deadly ceilings would force many level changes. Levels can break either by new backroutes or by failing intended routes; I assume here we would lose intended routes from randomly dying climbers.

-- Simon

[ccexplore]

(Sorry if below points had already been made on the thread.)

Climbers don't climb the level's left/right borders in NeoLemmix, do they?  They certainly don't in original Lemmings, despite other interactions present for those borders (eg. turning around walkers).

At least with climbers, even if they could climb the borders it would be quite difficult to exploit that behavior usefully.  Shimmying across the top border however would seem way more exploitable than any other ceiling interactions to date.  I don't think it is acceptable behavior to open up such an exploitable border behavior.  Heck, some people already feel even the existing border interactions are already too helpful.

[Nepster]

First of all: While I am against reiterating the same arguments over and over after having reached a conclusion, I think it is totally fair game to open up a discussion again, if new developments add substantial arguments to one side or the other.

Arguments for solid ceiling:
1) Existing content, that relies on it.
2) L1 had this behavior. As NeoLemmix already changed quite a lot of stuff, I suggest to disregard this point.
3) Climbers may accidentally vanish out of the screen.
4) Sky isn't deadly, cf. Dullstar's argument: Well, if you really want an in-game reason, one could always say lemmings will die due to asphyxiation due to thin air this high.   More problematic is the question: If the sky is considered air, then why do builders turn around when they hit the ceiling, instead of continuing forward?

Arguments against solid ceiling:
1) The ceiling is currently an exception: All other sides are air and lemmings die when they reach them. So there is no reason to expect that the ceiling behaves in any way differently.
2) Shimmiers should not be able to shimmy across the top border. With a non-solid ceiling, this is a non-issue. Otherwise it would require ugle special-casing to determine whether the shimmier has actually grabbed terrain or the level ceiling. And this special-casing would even be needed during each frame, because shimmiers may at first hold onto actual terrain and then reach the ceiling while already shimmying.
Moreover it is not clear to me why there should be a difference for shimmiers but not for builders for an empty ceiling compared to one that has a 1-pixel high terrain ceiling.
3) What should we do with jumpers, which will be implemented at some point, too? If we consider the sky to be empty non-shimmyable, then the jumper should be able to move through the ceiling with his head, instead of hitting and dropping down.

Upshot is: With the shimmier and the jumper ceiling interactions become much more frequent and muc more varied. If we keep a solid ceiling, then we will need a lot of special rules to define what happens to shimmiers and jumpers, while the rule-set would be a lot simpler for non-solid ceilings.
Therefore I am acutally much in favor of non-solid ceilings, which will be far easier to learn and understand given the new additions to the skill set.

I made a new-formats version on NeoLemmix with non-solid ceiling for you to try out, cf. attachment. Surprisingly few replays fail actually.

When I played through Lemmings 2, I found the deadly ceiling to be one of the most annoying aspects of that game (granted, the flinging mechanics made that much worse than it would be in NeoLemmix, but still). Considering that the ceiling's behavior doesn't usually result in backroutes, the ability to put builders close the ceiling without having them die randomly because they went slightly too high is rather nice - besides, it's logical that lemmings, if allowed to go offscreen towards the top, would eventually fall back into the level area due to gravity, which isn't the case with the bottom (because you can think of it as a bottomless pit) and the sides (where there's no particular reason why the lemmings should EVER turn around). It also has the convenience thing of not having to babysit climbers as much - and if the level designer really wants to fry rogue climbers it's easy to use a flame trap for that, especially with style mixing.

With the skill blueprints accidentally building out of the ceiling or jumping out of it will be more or less a non-issue. And jumping out of the screen with a shimmier... Let's just say I have faith in my fellow players that they will be able to recognize when there is no solid terrain above the lemming .
Regarding the climber: Yes this is an issue, but even now I would certainly encourage level designers to put terrain there, if the climber is intended to drop down again and not rely on the player knowing that the ceiling acts solid.

It is also logical that the top edge of the level represents sky.

Then please explain, why climbers should drop down if they reach the ceiling, if it is so logical for it to represent sky.

You cannot hold onto the sky, therefore shimmiers should not be able to attach to it. To be honest, I don't think it's even all that inconsistent to disallow the shimmier from using the ceilng, because while we sometimes consider the ceiling to be technically equivalent to a mass of steel just offscreen, there's no particular reason it needs to literally behave like a mass of steel at the top of the screen. The main reason I think it makes sense to have lemmings hit their heads on the ceiling is because you have to set an arbitrary "you can't build here" cutoff SOMEWHERE to prevent the lemmings from walking on top of the level boundary - and having them hit their heads certainly makes a lot more sense than them randomly dying because they wandered 1 pixel too high.

I am no really convinced here, that turning around due to illogical steel at the ceiling is in any way better than arbitrarily dying there. Certainly neither is prefect, but I cannot see why turning around is inherently more logical than dying. More convenient perhaps, but that wasn't your argument in this paragraph.
And finally I just have to ask: Why should a builder turn around, when hitting the ceiling with his head, if there is no steel at the ceiling? Why not continue until the builder's feet hit the ceiling? If the ceiling would be considered sky, then certainly this would be much more logical, wouldn't it?

[nin10doadict]

I wouldn't mind if the ceiling did become deadly... The only level I can think of that I've played recently where I used the ceiling was "Bridgeworks" from Sublems, where I built all the way up to the ceiling to form a wall up there, which I don't think was the intended solution anyway.

[Dullstar]

I focused primarily on the builder for my argument while mostly ignoring the climber. For the climber, I don't really like either option but I also don't really have a good reason for one over the other. Having them hit their head is certainly more convenient, but you'd hope level designers would avoid making level designs that rely on this behavior.

I kind of like some of the ideas Simon came up with regarding infinite terrain extension (which would kill climbers), though I'd make the following suggestions on how to handle builders/jumping shimmiers/etc.
- Don't kill
- Cut off building close to the edge of screen to avoid offscreen terrain, but don't turn the lemming around.

I don't think there is any fully logical way that also makes good gameplay sense to fully handle the level boundary, since such a boundary doesn't really have any real life equivalent. Even if you made the argument that the air is thin, they wouldn't asphyxiate immediately, after all - if it was just, say, a builder stair jutting of the screen, they could probably hold their breath long enough to survive.

Whatever the outcome, though, I'll make sure that any future levels I create are unaffected by the ceiling behavior.

[GigaLem]

I'll agree with Dullstar here, the ceiling shouldn't kill but it shouldn't allow lemmings to shimmy across it.
Essentially, the ceiling can be summarized as a One way down field.

[Nepster]

Given that the No-Ceiling version was downloaded 9 times, and only Dullstar and GigaLem complained about the non-solid ceiling, I think that this experiment went pretty well. So the next experimental version will have non-solid ceiling.

[Strato Incendus]

Are there any original levels (meaning from Lemmings / Oh no! More Lemmings) which are going to have their solutions affected by this?

(Just noticed that a level I created today happens to fall into this category .)

Spoiler

[Proxima]

There are a couple of levels in Genesis Lemmings that rely on a solid ceiling, notably Everyone's a hard nut and Lemmings' Ark.

For Redux, I've added terrain at the top edge in any case, since regardless of what's eventually decided, I don't think it's fair to rely on players knowing about hidden mechanics.

[Strato Incendus]

For Redux, I've added terrain at the top edge in any case

Ah, nevermind, I wasn't even thinking! The solution to such problems can be so easy...

Though I'd suggest to use specifically steel where such "alibi terrain" is needed. Otherwise, it would still be possible to build towards that top terrain and bash through it or something similar. Afaik this was possible in the Mac version of Lemmings and Oh no! More Lemmings, and you can easily break some levels this way (for example "Mind the Step", Mayhem 28) by just going above the entire landscape .

[IchoTolot]

I swor not to post in here again and I thought the discussion was finally over with the decision to respect the results of the old discussions, but the out of nowhere decision for change I cannot let just stand here.

I find a non-solid ceiling the wrong decision and as my researches of the past have shown a ton of content is affected by this, just to get rid of edge-cases of potential future skills. I would rather cut these new skills even.

When I played through Lemmings 2, I found the deadly ceiling to be one of the most annoying aspects of that game (granted, the flinging mechanics made that much worse than it would be in NeoLemmix, but still). Considering that the ceiling's behavior doesn't usually result in backroutes, the ability to put builders close the ceiling without having them die randomly because they went slightly too high is rather nice - besides, it's logical that lemmings, if allowed to go offscreen towards the top, would eventually fall back into the level area due to gravity, which isn't the case with the bottom (because you can think of it as a bottomless pit) and the sides (where there's no particular reason why the lemmings should EVER turn around). It also has the convenience thing of not having to babysit climbers as much - and if the level designer really wants to fry rogue climbers it's easy to use a flame trap for that, especially with style mixing.

It is also logical that the top edge of the level represents sky. You cannot hold onto the sky, therefore shimmiers should not be able to attach to it. To be honest, I don't think it's even all that inconsistent to disallow the shimmier from using the ceilng, because while we sometimes consider the ceiling to be technically equivalent to a mass of steel just offscreen, there's no particular reason it needs to literally behave like a mass of steel at the top of the screen. The main reason I think it makes sense to have lemmings hit their heads on the ceiling is because you have to set an arbitrary "you can't build here" cutoff SOMEWHERE to prevent the lemmings from walking on top of the level boundary - and having them hit their heads certainly makes a lot more sense than them randomly dying because they wandered 1 pixel too high.

Consider a level, one screen large, with the sides and ceiling accessible to lemmings. Now picture a builder that, if allowed to continue building as long as possible, reaches the point where it would die with deadly ceilings. Now, let's expand the level, repeating the terrain on the edges of the level infinitely. Lemmings traveling to the left and right would continue travelling left and right forever (assume the player lacks a skill to turn them around with), whereas the builder, building in the same position as before, will fall back down after passing the former ceiling. This is a compelling reason why it makes sense for the sides to be deadly, but not the ceiling.

To summarize, I am strongly against deadly ceilings, and against shimmiers attaching to the sky.

I 100% stand behind Dullstars post here and highly protest against a non-solid celing!

[Strato Incendus]

If I understood Dullstar's post correctly, this is mainly an argument against a DEADLY ceiling, not necessarily against a non-solid one (i.e. one that would allow you to "build over the level" and then fall back into it).

I agree that deadly ceilings were annoying in Lemmings 2, mainly however for jet packs and super lemmings and all that other nonsense which was difficult enough to control with the mouse in the first place. A deadly ceiling would probably be less problematic if it triggered at the lemmings feet rather than its head (which would pretty much require almost the entire lemming to be outside the level before the trigger occurs, so there would be less "random" deaths). But still, I'd also prefer a non-deadly ceiling .

That doesn't necessarily mean it has to be solid, though. Since I'm currently creating levels for NeoLemmix 1.43 again, "solid like steel" in that case actually meant climbers could go up the sides of the levels. So if we consider the ceiling this type of "solid", it would make sense for shimmiers to attach to it. However, since level sides have been deadly for a long time now, for the current state of affairs it would make more sense if neither climbers nor shimmiers could interact with the ceiling in any way.

Therefore, GigaLem's analogy of the "one-way down field" seems to make most sense to me. Of course we technically don't have one-way down fields yet, but considering what the left- and right versions do: You can build against them or use them to turn around, but you can't climb them.

Transfering this principle to the ceiling would mean: You can turn around by bumping your head, but you can't shimmy along it.

[Simon]

"one-way field" is a bad name and clashes with "one-way wall", which is implemented as a field, too. Rename the thing to leftforce, rightforce, downforce. Or leftpusher, ...

Now, whoever wants the ceiling to be air with downforce: Please spec your proposal such that a computer can understand it.  "can be summarized as" or "principle means that special case X is resolved as" won't cut the mustard. I'm sure there is a nice idea to investigate here, but you're too vague.

Example spec: Whenever you would enter downforce as a walker, instead turn. Whenever you enter downforce as a faller, ignore (behave as if there were no downforce). Whenever you would enter downforce as a non-walker-non-faller, instead become walker or faller depending on whether you're standing on ground or not, and also turn around (I despise this turn, it's for back-compat).

Beware: Whenever your behavior becomes cute or introduces new mechanics, your ceiling becomes an interesting gameplay element. Your ceiling might be interesting in level designs, exactly what we'd like to avoid. In the above spec, the ceiling would catch jumpers mid-flight and have them drop straight down from the apex of the jump.

Also beware: Nasty corner cases of walkers getting stuck in downforce after walking into it from slope below.

-- Simon

[Nepster]

A deadly ceiling would probably be less problematic if it triggered at the lemmings feet rather than its head (which would pretty much require almost the entire lemming to be outside the level before the trigger occurs, so there would be less "random" deaths). But still, I'd also prefer a non-deadly ceiling .

The ceiling would only be deadly if a lemming is completely outside the level area, i.e. if not even a single foot pixel is still visible.

I agree that deadly ceilings were annoying in Lemmings 2, mainly however for jet packs and super lemmings and all that other nonsense which was difficult enough to control with the mouse in the first place.

Yeah, I totally agree with that (and I hopefully mentioned that before in this thread - at least I should have done so). But this is mainly a consequence of all these flying skills and the horrible fan, neither of which is or will be present in NeoLemmix in the future. So I don't think the L2 experience is a good guideline for the choice we have to make here.
On the contrary: Lix has flinging and lix die if flung through the ceiling. Similarly for lix jumping out through the ceiling. As far as I know, this works pretty well and there is almost no complains. The main reason is probably that it's far easier to estimate what would happen, so players who lose lemmings that way think "Oh, my bad, I should have seen this" instead of "Stupid ceiling! I tried to fly my lemming within the level boundaries - why did you vanish nevertheless". At least these are my own reactions...

Now, whoever wants the ceiling to be air with downforce: Please spec your proposal such that a computer can understand it.  "can be summarized as" or "principle means that special case X is resolved as" won't cut the mustard. I'm sure there is a nice idea to investigate here, but you're too vague.

Well, although I certainly wouldn't mind if you do, noone has to write computer-readable specs here.

The main problem with all these suggestions that let climbers fall down but disallow shimmying across the ceiling is the following:
There are functions that answer the question "Is pixel at position (x, y) terrain?" and similarly for steel, any trigger area types, etc. Given the code structure these functions are the only way to check a certain pixel coming from a lemming action (at least without duplicating lots of safety code, too). This means that it doesn't matter whether a climber or a shimmier asks whether the pixel (0, -1) (the first ceiling pixel) is solid, this function will always return the same value.
So the lemming himself would have to check whether he is at the ceiling of the level or not. But what does a lemming care about the global properties of the level?
At the moment, such information could be read by a lemming, but this is more of a design fault in the code than anything else. All this means that to implement such a rule, I would have to introduce a rather tight coupling between a lemming and the global properties of the level, which creates problems in ensuring the code is bug- and glitch-free and in supporting the code over the next years.

Therefore, GigaLem's analogy of the "one-way down field" seems to make most sense to me. Of course we technically don't have one-way down fields yet, but considering what the left- and right versions do: You can build against them or use them to turn around, but you can't climb them.

Transfering this principle to the ceiling would mean: You can turn around by bumping your head, but you can't shimmy along it.

There is one big difference between the existing force fields and the downwards version: Lemming have a direction that is left or right, which simply can be reversed, and the resulting position totally makes sense. But lemmings don't have a downward or upward direction, so it's far less clear what should happen there. I am almost certain that you would not want a builder to continue building a downwards bridge . 
Should lemmings turn around? If not, then walkers on sloped terrain will be stuck at the ceiling position and climbers will simply fall down and climb the same wall again. If yes, then shimmiers and jumpers hitting the ceiling would turn around as well, which doesn't make sense at all (at least to me).
Should skills stop when hitting down force-field? If not, then builders would continue placing another brick, move a little bit forward and then fall through the bridge because they hit the end of the previous brick, while all other lemmings would be stuck at the top. If yes, then this would be a huge change compared to the other force fields, and gliders in updrafts would move along incredibly funny paths, because they would regularly revert to fallers and have to start gliding again.
How should bashers behave? Force fields are not solid, so bashers would stop much easier at the top and revert to walkers. Currently they continue, at least if there is usual terrain around the top. In my opinion, this makes a lot more sense, especially as it's much more common to have a huge terrain wall when bashing on the top (think "Going Up...", Mayhem 23) than having a pathway at the very top where lemming could walk along and the only bashable terrain might possibly be outside the whole level.
If jumpers are allowed to go through the ceiling with their head, but climbers will fall down when hitting their head at the ceiling,, what should happen if a jumper hits a vertical wall? I think it would be nice and helpful to transition into a climber. But what if the jumper's head is above the ceiling. Then we would transition into a climber where a climber should never, ever be!

[mobius]

maybe this was said before but I don't remember:

Why do so many levels depend on ceiling behavior? I personally think this is not the best design and in my own levels I (now) try to avoid anything finicky like messing around near the ceiling. I personally don't (usually) care for levels that are all about the greyer areas of the game like edge behavior, stacking 7 builders to make a wall etc.

In my own level pack, I can't remember off hand how many (if any) of my own levels depend on ceiling behavior. In the current pack I'm working on so far only 2 levels would be affected by it and they are very easy to change (one way or the other)

Anyways; that's way this isn't really a huge issue for me. But for the record; I would prefer deadly ceilings for reasons I guess already covered;
-it's consistent with all other sides.
-it allows for simpler; neater design. If you want lemmings to die above; no need for traps; just open the ceiling. If you want a safe ceiling just add terrain or steel. Seems simple to me.
-all of this other discussion about 'downward force floors' or the ceiling acting like sky is just silly imo. This is too complicated; I think the game should be simpler. It is already complicated enough.