[REJECTED][Suggestion] Ghost as a skill

[Strato Incendus]

In the same vein as Crane proposing the reversal of a change that has been made a while ago , I'm going to bring up another thing from the past:

If you remember, around the time of the major format switch last year, the idea of neutral lemmings was being discussed. Skills were not supposed to be assigned to them, like with zombies, but in contrast to zombies, they could still be saved, or even had to be saved.

I don't know if that idea is still being ventilated, but I think a lot more puzzle potential actually lies in a thing we already had back in the day, which was the exact opposite: Ghost lemmings. They could not be saved, however they could be assigned skills, and they didn't interact with any kind of objects. This allowed them to walk through fire, traps, and water alike, doing things like mining or digging at the bottom of a water pit, which isn't possible otherwise.

Ghosts were removed together with gimmicks, and with regard to the latter, I understand how they would have slowed down every new implementation of a feature a lot. Having to balance e.g. a new skill with all these extra rules would have been pretty chaotic, obviously.

Ghosts, however, actually require less interaction in total than even regular lemmings! They simply treat all objects, including exits, as if they were background objects. So I can't imagine they would require a lot of code, plus the sprites are already there. Ghost sprites for new skills introduced in the meantime (=fencers and shimmiers) can easily be recoloured, just like we have created different lemmings sprites for the L2 tilesets.

There was one issue with the original ghosts, however: They always came as pre-assigned lemmings, i.e. either they were pre-placed lemmings or came out of pre-assigned hatches, like zombies.

That made me realise what a ghost actually is: Just another skill! So why not re-introduce ghosts as skills that can be assigned like any other?

This would make the ghost a unique hybrid of lethal skills (like bombers and stoners) and athletic skills (like climbers and floaters):

You can get past certain obstacles with him which no other lemming can get by - for example, fire traps - but in exchange for this enormous power, the lemming itself is sacrificed (since an exit is also an object, so he can't enter it anymore either). Athletes, often employed as pioneers, usually take paths the rest of the crowd can't take, and their task is to then create an alternative path for said crowd.

In some sense, ghosts are actually a more balanced version of the disarmer - because they can go through traps like disarmers can, but leave them active. The disarmer seems to be the least favourite skill at the moment, mainly because levels involving it boil down to
1) collect the disarmer as a pickup skill, the level is auto-solve from there;
2) pure timing-challenges, stalling the crowd until the disarmer gets his job done;
or 3) the crowd takes exactly the same path as the disarmer pioneer, because there's no difference between a disarmer and a regular lemming once a trap has been deactivated.

Some people even advised for disarmers to be culled during the format shift. While I of course will always be against such a move, I do have to say, if we still have even disarmers, there's no reason why we shouldn't have ghosts as well .

Also, it's in spirit (*badum-tss*!) of the Halloween season !

[IchoTolot]

I didn't like ghosts back then and that pretty much has not changed.

They just introduce a whole new set of behaviors to the game and as a result making it a lot more complicated. The lemmings fear ghosts thing also can lead to a ton of fiddly stuff.

The shimmier as an example doesn't require to learn a lot of new behaviors, so it's a rather simple skill compared to the ghost mess.

I still support the gimmick and ghost culling back then with all my heart, it was nessesary to remove this clutter and I am a very careful person when it comes to culls.

[Strato Incendus]

The main case I'm making here is that the ghost and gimmick removal are not necessarily the same thing. Unless you want to call ghosts a gimmick, too, because they were part of the list, but the same is true for zombies (these two gimmicks were activated by default). So in that case, we could be nitpicky and say gimmicks were never fully removed, becaues zombies are still around .

But since the colloquial definition has now become that zombies aren't gimmicks, I consider ghosts and zombies just different types of lemmings, like water lemmings in Lemmings Revolution, whereas gimmicks are legitimate physics changes.

A gimmick changes the rules of everything, at least several skills, if you think of gimmicks like hardworkers, lazy lemmings, wraparound, no gravity etc. Naturally, this is going to come with a lot more coding effort.

Ghosts, in contrast, actually do less than regular lemmings, but precisely in that lies their strength. So if Lemmings were programmed in Python and I wanted to add ghosts, I'd simply copy and paste everything I need for normal lemmings, and then delete everything that defines an interaction with an object. Done! I don't know whether this is more difficult in C++, but on a general level, I can't see how fewer interactions between ingame objects could be more complicated than what we already have now.

The lemmings fear ghosts thing also can lead to a ton of fiddly stuff.

Well, at least now I finally know why the ghosts push other lemmings around! I never thought of flavour reasons for that!

To be honest, that is kind of an odd additional behaviour. One could consider removing that / not re-adding it, in case ghosts should be re-introduced one day. To avoid confusion with original ghosts, call them spirit lemmings from now on, if you prefer.

Zombies have undergone a change as well, from Classic Zombies, which could collect pickup skills and press exit buttons, to the current zombies, which cannot. (Btw, in my first couple of days on the forum, I proposed Zombies should be able to do this, and while people offered me a lot of valid explanations why they shouldn't, I think no one actually mentioned that once, they had indeed been able to do this! )

Also, since other skills usually end on "-er", I'd also be fine with "spiriter", or something along those lines. Not "ghoster", though, since by that I'd understand someone who engages in one or another form of activity termed "ghosting" .

I think it's just pretty plain to see that ghosts are in many ways strictly more versatile than disarmers, while at the same time not being a lot more additional effort, since the sprites are already there and the coding part should largely just be "regular lemming minus X". Especially if you refrain from re-adding the "intimidate other lemmings" behaviour.

Although people who like using compression method levels, like nin10doadict, might object to that .

[IchoTolot]

Unless you want to call ghosts a gimmick, too, because they were part of the list, but the same is true for zombies (these two gimmicks were activated by default). So in that case, we could be nitpicky and say gimmicks were never fully removed, becaues zombies are still around

Then let's call zombies a gimmick. They were used by quite a lot of people and were generally accepted (that was the main reson we wanted to keep them), ghosts saw only use in very very few levels (namida and Giga were the only ones with ghosts levels I recall).

Ghosts, in contrast, actually do less than regular lemmings, but precisely in that lies their strength.

That still doesn't reduce the actual clutter they are going to add.

So if Lemmings were programmed in Python and I wanted to add ghosts, I'd simply copy and paste everything I need for normal lemmings, and then delete everything that defines an interaction with an object. Done!

That is exactly what you should not do! Things like this make your code a total mess!

Copying a huge code chunks over and over is not good, you really want to avoid that in large programs.

I think it's just pretty plain to see that ghosts are in many ways strictly more versatile than disarmers

That is not a clear reason to add a skill. Let's propose a teleporter skill that teleports a lemming, as it is way more versatile than a climber!

Also less versatile skills are often easier to learn for players. This can be a double edged sword. Disarmers are easy to understand, ghosts not (also I wouldn't call them intuitive).

Furthermore I wouldn't call the non-interaction with objects a huge pool for extra level ideas and really mostly just an addition of extra game rules. There are enough ways to go around objects.

Because of things like this we had 20+ gimmicks back in the days. Also there is already a huge list of things we want to implement. We had resons to burry our past sins and we should not open the old graves again.

You won't convince me for the ghost case here really. I disliked them back then and my mind hasn't changed.

[Simon]

simply copy and paste everything

-- Simon

[Strato Incendus]

@Simon: Trust me, I know that issue ! But isn't that only a problem when programming something entirely new? Regular lemmings, as they currently are, don't seem to have any errors, they work just as they should?

Obviously, the old code for ghosts can't be revived again, because it was a different format. But it should be possible to develop them out of standard lemmings, if there is enough use for them.

@IchoTolot: You wouldn't have to copy it over and over - just once .

The teleporter skill you proposed as an example, in contrast, would be something completely new. It would require a lemming to be clicked twice (once for assigning the skill, the second time for defining a target position), it wouldn't build on anything that is already present, and it would most likely be completely broken, since that lemming could go literally anywhere in a level.

Ghosts, however, have very clearly defined uses and limitations - which makes it easily visible that the disarmer is a lot more limited in design space.

I'd also make the case that disarmers are less intuitive than ghosts, because they only work on triggered traps, whereas ghosts bypass all traps (triggered, fire, water) and benefits (pickup skills, teleporters, buttons, updrafts) alike. Also, disarmers remove the trigger of a trap the second they step on it, even though the animation is still playing. This would confuse me even more as a new player; I'd expect the trigger not to be removed until the disarmer animation is finished.

The most confusing part about the ghost is the intimidation part, and as I said, that could be left out, as far as I'm concerned.

We had resons to burry our past sins and we should not open the old graves again.

Since this seems to be your standard argument against reintroducing anything that has been removed once, may I remind you of features that got removed from Old to New Formats that are now being considered for reintroduction, like the replace piece function in the editor, or have just been reintroduced, like the level replace function in the pack toolkit.

These changes show that not everything gets removed for a reason; some things just got removed because no one found a use for them back then. This can change however with new members joining the forum.

A lot of people also still don't find a lot of use for many of the NeoLemmix skills, and prefer to stick to classic skills instead. I generally don't get a lot of ideas involving buttons, pre-placed lemmings, and pickup skills. But that may just be a lack of creativity on my part, as far as a given feature is concerned, rather than that feature being useless in general.

[IchoTolot]

@IchoTolot: You wouldn't have to copy it over and over - just once .

That's how it all starts.

Ghosts, however, have very clearly defined uses and limitations

I'd also make the case that disarmers are less intuitive than ghosts

I know that ghosts can fly so I expect these ghosts to also fly around the map. Ghosts are far far far away from being intuitive.

Also I see the object ignoring not really valuable, there are enough substitutes for it as well. The fear part would be the only thing remotely interesting, but it can lead to very fiddly stuff, which I don't like.

may I remind you of features that got removed from Old to New Formats that are now being considered for reintroduction

All the cases you mentioned here, came together under vastly different circumstances than the big gimmick cull.

Back then we all actively discussed and decided to remove gimmicks.

All features you mentioned here are casualties to the great tool switch.

Backstory:
Nepster created a personal editor for himself back then. When the new formats came along, this freshly built editor, which was mostly built around Nepster's level creation needs, became the base for the new editor. In the next months we gave a ton of feedback for this editor to make it more generally accepted.
The way of inserting terrain was one of the most criticised things about the old editor, so this one took another route.

The toolkits had to be coded from skretch, so things like the "replace" button never were actively culled, they were just forgotten/overlooked.

[namida]

Ghosts, however, actually require less interaction in total than even regular lemmings!

This does not mean they will not cause a mess in the code. Consider the regular lemming to be the "base scenario". Any deviation from this - doesn't matter if they do more, do less, or just do something differently - adds complexity to the code. Just because our test is "if the lemming is a ghost, do nothing" rather than "if the lemming is a ghost, do this extra thing" doesn't change that much - it's still an extra test, and an extra special case to handle differently from normal lemmings.

In the case of zombies, there had been a lot of really good puzzles making use of them, that warranted keeping them. Ghosts, on the other hand, turned out to be another one of those ideas that sound great in theory, but are actually really hard to make anything useful of in practice.

[Strato Incendus]

Also I see the object ignoring not really valuable, there are enough substitutes for it as well.

Fair enough, I just find these ways of circumventing objects disproportionately often result in just building over it . And every builder you have to provide to get past an obstacle, even if it's just the pioneer lemming who has to get past it, rather than the entire crowd, is another builder which might potentially break the level (given that it's widely accepted as the most powerful skill).

Ghosts don't just simply walk past fire and water, but more importantly, also past exits. How many levels do you remember where you had to build over the exit to prevent the pioneer lemming from leaving the level too early? I am already being creative with doing this in different ways than building over them (stack + climb, bash underneath and cancel the basher with a walker, so that the ceiling isn't actually penetrated and lemmings can still walk over it). But I don't want to overdo these tricks either.

Plus, let's not underestimate the ability to perform skills inside water areas! Fire-related stuff isn't useful for regular lemmings, but swimmers are a thing, however they can only bomb and stone while in water. Ghosts can bash / mine / dig / fence from inside a waterpont to allow swimmers to get into that pont. Or to kill zombies, of course , but ghost-zombie interactions are next-level stuff, of course!

I know that ghosts can fly so I expect these ghosts to also fly around the map. Ghosts are far far far away from being intuitive.

I appreciate you using flavour arguments rather than game-mechanical ones, for a change! Usually, it's the other way round...

Well, if we go there, I could also say builders are unintuitive because it shouldn't be possible to build in mid-air without the staircase still being firmly connected to terrain at its foot. Or, as said before, when I was new to NeoLemmix, I expected the disarmer to also be able to disarm fire traps.

You can't control subjective player expectations anyway; players get used to what they are being taught by the game. The behaviour of ghosts is very consistent, they ignore all objects equally, even those you may not want them to ignore (like pickup skills, teleporters, and buttons).

The sacrificial nature of the ghost-skill would add in an extra layer of complexity that old ghosts did not have, because you'd have to make a conscious decision as a player how many of them you want to use - like with bombers and stoners. Just like the latter, level creators could even occasionally provide them as red herrings to increase the entropy of a level!

This does not mean they will not cause a mess in the code. Consider the regular lemming to be the "base scenario". Any deviation from this - doesn't matter if they do more, do less, or just do something differently - adds complexity to the code. Just because our test is "if the lemming is a ghost, do nothing" rather than "if the lemming is a ghost, do this extra thing" doesn't change that much - it's still an extra test, and an extra special case to handle differently from normal lemmings.

Okay, I get that, but isn't that the case for any new skill we introduce? For example, the standard rule that skills get canceled when a lemming loses ground under his feet has to be suspended for the shimmier.

That's why propose to indeed view "ghost" as just another skill, rather than a gimmick or different type of lemming. Currently, it's a skill that has been culled, and in "very Old Formats", it was a skill that could only be pre-assigned to a lemming. Pretty much as if disarmers could only be pre-placed lemmings. That wouldn't make sense either, would it?

Zombies, in contrast, are different types of lemmings, because it wouldn't make sense to have them as a skill and voluntarily make one of your own lemmings a zombie (except in a hypothetical 2-player game, perhaps, to kill off your opponent's lemmings ).

[ccexplore]

They just introduce a whole new set of behaviors to the game and as a result making it a lot more complicated. The lemmings fear ghosts thing also can lead to a ton of fiddly stuff.

I believe Strato's proposal will not include the "fear ghost" behavior.  As I understand it, essentially it just grants immunity to all object interactions to a lemming, but otherwise nothing else changes.  It would be a permanent skill in the vain of climbers and floaters.

As far as programming goes, I believe you would first check whether a lemming is a ghost before going into the code that checks for specific object interactions.  I think that basically might be all there is to it.  So on first glance, it shouldn't require too much changes code-wise (at least to the physics; you'll still have to work out details of the appearance changes that distinguish them from normal lemmings, and to distinguish whether the ghost is also climber or whatever), though ultimately that will depend a lot on how the existing code is structured.

It still sounds somewhat overpowered to me, even with the fact that assigning it automatically excludes the lemming from being able to exit.  But it's hard to judge puzzle potential from just a description of the behavior.  namida's opinion that the old ghost gimmick didn't ultimately contribute much is perhaps a good data point to take into account, even if that gimmick had other extra behaviors (eg. fear ghost) that we don't want nor care about here.

[namida]

(at least to the physics; you'll still have to work out details of the appearance changes that distinguish them from normal lemmings, and to distinguish whether the ghost is also climber or whatever)

This part would be very little trouble. NeoLemmix already has code to handle recoloring of sprites for special cases (see: athletes, zombies); no separate sprites exist for these, just a small text-based datafile that says what color to swap with what other color. (In cases like the Xmas lemmings where only part of the lemming is recolored? That's simply done by having those parts have a very slightly different color, usually only off by 1 on a single color channel.)

namida's opinion that the old ghost gimmick didn't ultimately contribute much is perhaps a good data point to take into account, even if that gimmick had other extra behaviors (eg. fear ghost) that we don't want nor care about here.

To be honest, the best (and I still use that term pretty loosely) ghost-based levels, relied more on that than the object-ignoring. A lot of the better ones also relied on the "become ghost upon death" gimmick too. While I can't rule out that there is good potential I just never found, I'd definitely from experience lean towards thinking they're not that useful in practice.

[GigaLem]

If I had to say anything on the subject, it'd be better to reinvent the ghosts in a new intuitive way. Like having certain objects be affected by ghosts if the check box for em is turned on. And if so an obvious clue would tell you that ghosts can be affected by it and lemmings cannot. Or have a skill that can be assigned to not just ghosts but zombies too, making them come back to life as normal lemmings. A "revitalizer" if you will.

Instead of just declining the idea of the problems it had, why not find a way to turn things on their head and make something better or work. I rather give this a shot if there's a new way to make em work. (Like stated above) because I'm willing to converse and sell ideas to add depth to level creation rather than take away. We can't rely on standard stuff forever and it wouldn't hurt for once to give it shot. Because personally I rather stop with the instant and constant naysaying. Even unconventional things can work if remade and used properly.

[Strato Incendus]

"Ghost upon death" was basically the way to make ghost a skill back in the day: Assigning a bomber or stoner with that gimmick switched on essentially translated to "assign a lemming a ghost skill".

I wouldn't mind either if the fear-part remained, especially if other people could profit more from ghosts in that case . In fact, I'd actively try to make good use of that, too. I just wanted to state that this intimidation part isn't top priority for me personally, and thereby meant to offer an option to spare coding effort.

In the end, this mainly comes down to Nepster, how much additional effort the intimidate-part would mean for him . Current top priority seems to be the shimmier, which I fully support.


Of course, just like the shimmier, reintroduced ghosts would probably go through a test phase: You could try ghosts without intimidation first, since that means less coding effort, and provide the disclaimer that levels using them at that point may have to slightly change once the final behaviour is fixed. And then you see what people come up with, with and without intimidation. (Unless that disclaimer would intimidate creators from creating levels while the final ghost behaviour hasn't been determined yet ...)

I'll certainly create levels for the experimental version of the shimmier, at my own risk, so to say, but I can't speak for others, of course :/ .

If I had to say anything on the subject, it'd be better to reinvent the ghosts in a new intuitive way. Like having certain objects be affected by ghosts if the check box for em is turned on. And if so an obvious clue would tell you that ghosts can be affected by it and lemmings cannot.

Well, me personally, I'd always welcome such flexibility! And with the obvious cue telling you, I think there might be less people who object. Others however would probably regard this as inconsistent behaviour. Meaning, if in one level a ghost can push buttons, and in the next one it can't, based on a box in the editor being checked or unchecked, even if there were a cue above the ghost's head, player expectations would be subverted on a regular basis. This would make ghosts a lot more complicated and confusing, i.e. push them into the direction IchoTolot warns of.

Or have a skill that can be assigned to not just ghosts but zombies too, making them come back to life as normal lemmings. A "revitalizer" if you will.

Sounds like a nice idea in theory, I'm just wondering how that would work in practice. Do you have to kill the ghost or zombie and then revive it (how can you click a lemming whose sprite is no longer there)? Because if the skill exists to invert losses created by a ghost skill, why shouldn't it also reverse bomber and stoner-based deaths?

Basically, if it's just about replacing casualties, that's usually what cloners are for.

Instead of just declining the idea of the problems it had, why not find a way to turn things on their head and make something better or work. I rather give this a shot if there's a new way to make em work. (Like stated above) because I'm willing to converse and sell ideas to add depth to level creation rather than take away. We can't rely on standard stuff forever and it wouldn't hurt for once to give it shot. Because personally I rather stop with the instant and constant naysaying. Even unconventional things can work if remade and used properly.

That mindset in general I can fully support!

[IchoTolot]

We aren't simply recejcting everything! But we simply shouldn't and can't implement everything so that the game becomes hopelessly overloaded with a lot of stuff near to no one cares about (like it was in the past).

We already had a huge list with things we voted on to implement (https://www.lemmingsforums.net/index.php?topic=3399.0), but after the format change basic maintenance needed to come first before adding the new stuff we voted on.
New features are in development, but still need a bit of time and the top things we voted on will come step by step.

Now extra rescources shall go to a long discarded idea that did not work out before (general dislike, only 2 people used it, one of these people even admits it didn't work out now). Additionally with the suggested removing of an effect that failed idea had will make it even less appealing (even if that extra feature was fiddly).
This sounds insane to me and our focus for new stuff should be on the things the majority agreed upon would be good to have.

[Strato Incendus]

Sure, I'm not saying this should be done right now! It just needs to be proposed at some time in the first place so that the idea is out there; other skills may be suggested in the meantime, too (people have been toying with the idea of a downward builder or an upward digger of some sort). These aren't taking precedent over the shimmier either, but it's good to have them in the back of our heads.

As you can see, GigaLem would welcome the idea, and ccexplore doesn't seem to outright reject it either, since I understand his stance as "we can't predict how many puzzles people are going to create with this, especially if we slightly change the behaviour". namida has outlined where the challenges may lie, but also that some aspects of this are not as complicated as people may believe they are (like the colour change for ghost athletes).