[SOME CHANGES] [DISCUSSION] [PLAYER] Sides of levels?

[namida]

Edit by Nepster: Currently planned behavior for V1.44: Bottom, right and left edge deadly; top solid.

So - it's been discussed before but, how should the sides of the levels work? Current behaviour is that left / right / top are steel, and bottom is a bottomless pit.

Some packs have used the "Deadly Sides" gimmick on every level because the author prefers their levels to work that way.

Most likely, the Deadly Sides gimmick will be culled if the gimmick cull does go ahead, so I think this is also a good time to consider what should be the one consistent behaviour of the level sides.

[IchoTolot]

I think that the current way is the best way here:
Sides + Top solid       Bottom endless-pit

And i think this is WAY more logical than making every side deadly. 

Lemmings die when they fall in a pit and not when they reach the ceiling or the walls. And yes I count the three level edges as those!

Also so many things will break with a change for literraly no real improvement and I find deadly sides highly irritating.

[Nepster]

If we wouldn't already have lots of levels for NeoLemmix, I would argue as follows: There is no inherent reason any side should be solid or non-solid.
With the same reason as for the sides, one might argue that the bottom should be steel as well, because just why should it be any different than any other side? If one wants it deadly one could just place water there.
On the other hand, one might argue, that there is no reason why any of the sides should be walls: The lemmings world might just continue there are the lemmings are just wandering outside the reach of the player.

Considering practical aspects: It is much easier to add steel sides (which would even be directly visible to the player!) than adding traps on all sides, which in the case of the gimmick is not even obvious to the player.

But given that currently most levels use solid sides as default, I see how it might be a huge amount of work to adapt all of them (though I wonder how many of them make actually use of solid sides?). This might be the strongest argument for keeping the sides solid.

And for the record: I find solid sides highly irritating . They are only another way backroutes may sneak in.

[IchoTolot]

Considering practical aspects: It is much easier to add steel sides (which would even be directly visible to the player!) than adding traps on all sides, which in the case of the gimmick is not even obvious to the player.

Well to simulate deadly sides: Simply increase the level size so you have bigger gaps at the edges.
And I must say sometimes big steel edges look just plain bad especially if you have "regular size" levels and you have to add vertical scrolling just for a steel ceiling.

[mobius]

I argue to go with the practical aspect and make all deadly sides.

I admit when designing most of my levels I’ve put steel or walls around the edges [if it was appropriate] partly in case you ever changed this feature and partly because I think it looks better and makes more sense: It’s something the player might not know either way so better to be sure.
Maybe it was partly because I used to make levels for Lix but I just think it looks better too. I've always thought lemmings climbing up the side or just hitting and turning looked weird.

[Simon]

tl;dr: Would advocate for all 4 edges to be a deadly void. Levels that rely on either side edge behavior should be redesigned. I haven't checked existing levels, which are the strongest counter-argument.

Treating 3 edges with one rule, and the bottom edge with a different rule, that's inconsistent. An inconsistent design needs very good reasons for its defense. Nepster is right how there is no clear standard rule for either edge, but simplicity and consistency are arguments to treat all 4 edges the same.

Lemmings turning at the edges feels weird. Yes, when lemmings reach a wall, they should turn, but we don't see any walls at the edge.

Lemmings walking out of player's reach is weird, too, but less weird to me than turning at thin air. The bottomless pit doesn't have this problem: No matter what is deep down there, the lemmings would splat anyway.

It's harder to (turn mandatory walls into void) than it is to (place steel near the void). Ignoring existing levels, I'd propose deadly edges. This isn't the strongest argument on the table. Nepster is right how there is surprisingly little rightness in either choice.

Maybe allowing to reach the edges is bad level design in the first place. Last time Icho was here, he has shown me a one-screener in the cavelem styleset. The one-screener had extra steel near the right side, because the solid right side was important. Icho has not relied on the steel edge. The level was clear and honest.

Likewise, even if we have deadly edges, instead of relying on them, we should enlarge the level or put a trap.

And yes I count the three level edges as those [walls]! And i think this [bottom deadly, other sides steel] is WAY more logical than making every side deadly.

Why do you count the edges as walls, even though you don't see anything there? I will accept "because it feels right to me".

But then -- why do you find it logical/natural/... that the bottom edge kills, yet the side edges don't kill? What makes the bottom edge so special?

And I must say sometimes big steel edges look just plain bad especially if you have "regular size" levels and you have to add vertical scrolling just for a steel ceiling.

Nobody will add steel to the ceiling, no matter how we decide. The ceiling is the least concern.

At most, we might have to put climber-blocking terrain in a few places.

-- Simon

[Gronkling]

If neither option is preferred by a majority of people, it may be best to just allow it be an option. Possibly put an indicator on sides that are deadly.
I personally prefer steel edges, but this seems to be completely divisive. To me a steel edge feels like it can be used in a lot more different ways than a trap edge, and feels a lot more natural to me (I can't really explain why that is though). Being punished for building too close to the top of the screen also seems rather annoying.

[Simon]

Being punished for building too close to the top of the screen also seems rather annoying.

Yes! If we go for a deadly void, it should only kill what's completely outside. Being killed for being only close to the void, that sucks.

If we go for steel edges, I'd advocate to make them more visible. Maybe allow terrain to override this visible edge.

-- Simon

[ccexplore]

Treating 3 edges with one rule, and the bottom edge with a different rule

But then -- why do you find it logical/natural/... that the bottom edge kills, yet the side edges don't kill? What makes the bottom edge so special?

One word: gravity.

I think the setup with the bottom edge being deadly and the other edges being walls/ceilings is kind of modeling after the real world, where a "bottomless" pit (which is really just a very deep bottom) is "automatically" deadly by virtue of gravity, while most obstacles in the other dimensions are inert solid walls/ceilings unless hazards are explicitly placed.  Obviously you don't want to take this model strictly (eg. a strict reading implies floaters should survive a bottomless pit), but I think it is the underlying pinning for the intuition leading to the deadly-bottom-other-edges-walls mental model.

Boundaries are arguably weird by definition anyway.  Real world physics don't come with built-in boundaries; any boundary behavior, whether an obstacle-like versus hazard-like, implies the existence of something that is actually causing the boundary behavior.  In my mind it is just as mysterious for the boundary to be solid as it is to be deadly.  Arguably the only boundary behavior that's truly natural is toroidal scrolling.  Anything else is just something the player needs to learn.

The official lemmings franchise has been remarkably inconsistent with regards to boundary behavior.  In Lemmings 1 the underlying code definitely specifies the "wall-like" behavior, but the right boundary used by the game mechanics appears to be unreachable in many versions (PC unreachable, maybe Amiga, Genesis definitely reachable) relative to how far right you can scroll.  Lemmings 2 switched to the "all-deadly" behavior.  Revolution went with toroidal scrolling in the horizontal direction.  I don't quite remember how the top and bottom behave especially wrt to lemmings that have switched gravity, but I want to say that the top boundary is not automatically deadly to a normal-gravity lemming?

Having all edges deadly do have some other advantages besides more consistency (though that's not always the overriding factor--even in Lix, some behaviors like flinging are granted temporary immunity to the top edge for example).  For example if the physics is ever expanded to support multiple gravities (eg. the lemming's gravity is not fixed to one direction), the inconsistency of singling out a "bottom" edge (or have a boundary's "bottomness" vary by lemming based on their gravity) would become a lot more glaring.

I can understand that having to depict a solid boundary visually can ruin the look, and switching to all deadly may require changes to many existing levels.  On the other hand, how many levels truly would be difficult to change to accommodate deadly boundaries without adversely compromising solutions and visuals?  I'm not really convinced there are that many that would truly posed difficulties.

Ultimately I have little stake in this since I don't have any NeoLemmix levels, and having dealt with many lemmings games and clones that have different take on this (to say nothing of other video games), I've come to expect that boundary behaviors have be learned on a per-game basis, and that fortunately like Simon said, good designs usually try to avoid forcing you into boundary behaviors, at least for intended solutions.  The important thing is to have the boundary behavior be relatively consistent and discoverable.  I think I can probably even be okay with having behaviors vary from level to level (though again, it's hard to dismiss the suspicion that wall-like boundary levels can probably be converted to hazard-like boundary levels), as long as there's some easy to learn indicator somewhere in-game showing the type of boundary in effect for the level (and maybe this is taught early on via a tutorial level or something).

Likewise, even if we have deadly edges, instead of relying on them, we should enlarge the level or put a trap.

I do like having all obstacles and hazards be visible in principle.  I think it's just a little tough to adhere strictly to this especially when it comes to the top boundary, especially if you want to achieve an outdoor look visually.

[IchoTolot]

Treating 3 edges with one rule, and the bottom edge with a different rule

But then -- why do you find it logical/natural/... that the bottom edge kills, yet the side edges don't kill? What makes the bottom edge so special?

One word: gravity.

I think this is the best explanation for my thought process.

Lemmings die when falling into a pit (= bottom level edge). Lemmings are stopped by walls/ceiling (= other edges). I see the level edges as a "solid steel boundary" and not as a deadly "electric fence".

And I will use the words: "It just feels right to me." 
In most other games "I" played the level edges are not deadly so encountering deadly edges is unnatural. The bottom is a logical exeption, because of the gravity point (----> "Mario" is an example i can think of that uses a pit at the bottom)

[Nepster]

Well to simulate deadly sides: Simply increase the level size so you have bigger gaps at the edges.
And I must say sometimes big steel edges look just plain bad especially if you have "regular size" levels and you have to add vertical scrolling just for a steel ceiling.

If you have a big level, adding some empty space to both sides is totally fine and fits. But if you have a level fitting within one screen (i.e. 320x160px), adding 90 pixels of empty space to both sides is just wrong: Now over one third of the level is just there to keep the player away from the steel edges. Take as an example "Everclimbing Lemming" (Black Hole 17) from NepsterLems.
Some other like "Small Steps for Lemkind" (Comet 7) would need walls of continuous traps on both sides to prevent building to one side, blocking and then building back. Granted, this might fall under "bad level design", but it shows that increasing the size is not always the solution.

Regarding your second point: Steel borders need not look bad (depending on the level), usual terrain suffices most of the time and quite often you don't need to cover the whole border with terrain (just the relevant parts).
And I have yet to see a level where you have to add vertical scrolling and steel at the ceiling to make the level work.

Likewise, even if we have deadly edges, instead of relying on them, we should enlarge the level or put a trap.

If you really thinks so, quite a few levels in the Lix lemforum level pack need to be modified. Quickly skimming through Hopeless I see like 6-7 levels where solid edges can be used to backroute levels or at least simplify the solution.


I support the suggestions from Gronkling and ccexplore to make this an option. But I would make it an option in the Flexi Toolkit allowing to fix one behavior for a whole level pack, instead of a per-level basis. This way players can rely at least on some consistency.
As an in-game way to show the boundary behavior, why not color the edges of the minimap green if the edge is solid and red if the edge is void (or deadly)? Alternatively one might change the backgound color of the minimap for those parts outside of the actual level.

[IchoTolot]

I know a level from PimoLems ("Lemesis") which would need a steel ceiling. ("Going Up" from the original could also be a canidate).
I've not actively used the solid edges a whole lot as I mostly put quite a lot of decoration there. Nightmare 30 would be a canidate which uses the solid edges and a bunch of PimoLem levels. Another full check-through would be the case though and more stuff that need to be fixed would follow.

I can see the advantage of full deadly sides in Nepsters case and the more I follow the discussion I see the big mess in general we've gotten into by allowing 2 options in the first place.
The few deadly pixels overhang is sth which also bothers me. The deadly edges function as I understand them that  the outter ~8 pixels will be deadly instead of just the last one. It is still weird as hell in general seeing Lemmings get burned when they got near the edge.

If the deadly sides remain as an option there must be a way to notify the player though.
Or the "one-screeners" with them enabled can rely on a "decoration zone" instead of just empty space.

[Nepster]

I know a level from PimoLems ("Lemesis") which would need a steel ceiling. ("Going Up" from the original could also be a canidate).

Do they really need steel at the top? Would not adding some usual terrain wall of 7 pixels work as well and look much better?

The few deadly pixels overhang is sth which also bothers me. The deadly edges function as I understand them that  the outter ~8 pixels will be deadly instead of just the last one. It is still weird as hell in general seeing Lemmings get burned when they got near the edge.

I agree with everyone that deadly sides should only kill lemmings once they really hit the edge of the level and not before that.


And here is just another suggestion how to deal with this problem: Add to every style some kind of continuous trap like the vertical and horizontal red laser traps in Lix (found in geoo/construction). Then level designers who want deadly sides may frame their levels with this trap. It wouldn't be the the nicest-looking option, but at least this makes the deadly sides visible to players, does not require new options and uses only existing object types.

[IchoTolot]

And here is just another suggestion how to deal with this problem: Add to every style some kind of continuous trap like the vertical and horizontal red laser traps in Lix (found in geoo/construction). Then level designers who want deadly sides may frame their levels with this trap. It wouldn't be the the nicest-looking option, but at least this makes the deadly sides visible to players, does not require new options and uses only existing object types.

This "could" be the solution!

I like this idea, but maybe in some tilesets it can look kinda unusual ("Outdoor" or "Dirt" with laser beams ). Maybe match the color of the tileset (or builders) with them.

EDIT: Maybe the deadly edges option adds 8 pixels of map to each side and puts a laser beam in there automatically -----> lesser hassel for designers

[Simon]

I see the level edges as a "solid steel boundary" and not as a deadly "electric fence".

I don't see the edges as an electric fence either. Lemmings die when outside of player's range, not because they're grilled by the edge. The current NL burner edge doesn't feel right. The egde should do as little as possible.

My Feeling: open edges > steel > electric fence.

It is still weird as hell in general seeing Lemmings get burned when they got near the edge.

100 % agree!

One word: gravity.

other games "I" played the level edges are not deadly so encountering deadly edges is unnatural. The bottom is a logical exeption, because of the gravity point (----> "Mario"

In Jump'n'runs, gravity pulls you towards the bottom, where you're killed. To go towards the edge, you must want to go there, by steering towards the side. At the edge, the character ignores input towards that side.

Gravity in Lemmings, too, pulls you towards the bottom. But lemmings walk uncontrollably towards sides. You have to be active to prevent them from walking somewhere! The sides attract lemmings like the bottom does.

In Lemmings, it is natural to walk outside of the level. There is no magical barrier that prevents what lemmigns would do on their own. And while we might return in the real world upon hitting an obstacle, the lemmings get stuck.

But I would make it an option in the Flexi Toolkit allowing to fix one behavior for a whole level pack

Pack-dependent options open a new can of worms. What is with playtest? What is with standalone levels? Is the complexity warranted, even though authors have been consistent within a pack?

The per-level option is not as bad as it is now, should the boundaries get visual hints.

Always the same > option per-level > option per-pack.

Edges are weird, no matter what they do. I care about consistency. Let's not be coy. If we evade a decision, the players will suffer.

We're not in a hurry to decide, and can sleep over this for a week.

Likewise, even if we have deadly edges, instead of relying on them, we should enlarge the level or put a trap.

If you really thinks so, quite a few levels in the Lix lemforum level pack need to be modified. Quickly skimming through Hopeless I see like 6-7 levels where solid edges can be used to backroute levels or at least simplify the solution.

I'm okay with a few explicit hazards. I was speculating that edge reliance is bad design, but I'm not sure.

I agree that explicit hazards can be ugly. Like ccx, I don't want to require traps. When you have levels with large open areas, I agree that the void looks nicer than a load of traps.

Regarding your second point: Steel borders need not look bad (depending on the level), usual terrain suffices most of the time and quite often you don't need to cover the whole border with terrain (just the relevant parts).

Yes! Normal terrain suffices often and looks good.

Then level designers who want deadly sides may frame their levels with this trap

EDIT: Maybe the deadly edges option adds 8 pixels of map to each side and puts a laser beam in there automatically -----> lesser hassel for designers

I like this push for explicit edge behavior. I don't like mass traps.

When you want a void, the traps look horrible, and don't produce the right feeling either.

When you want steel, terrain or steel looks okay, and match what you perceive.

Once again. The edges aren't an electric fence. When edges are deadly, the lemmings get lost in the void outside of the level. The edges do as little as possible!

-- Simon