The Direct Drop Topic

[WillLem]

A better test would be to transfer a usermade pack made in a non-DD environment to a DD environment and see how many backroutes open up

Had a look at Mike's Lemmings (SuperLemmix Edition) today. This pack was originally created for NeoLemmix (plus, several of its levels were for other engines, and have been converted to new-formats NeoLemmix), but has a SuperLemmix-specific conversion.

As far as I can tell from a detailed look at each level, its intended solution, the available skillset, and my existing experience with finding and creating DD solutions, none of these levels can be backrouted using DD.

I'll have a look at GeoffLems next, and possibly NepsterLems if I get around to converting that one (currently likely).

[WillLem]

Coming back to this:

Consider the actual visual representations of objects, i.e. how they are presented to the player, not how the game sees them internally.

At present, it's possible to design a swirling vortex and have it be an exit. It's also possible to design an exit with steps and a doorway. Which design should the physics match?

In an environment that allows a range of custom styles and endless design possibilities per object, it very quickly becomes absurd to make decisions like "exits shouldn't be accessible in midair", because someone might design a mid-air exit. Conversely, we also shouldn't only allow exiting in mid-air, because someone might design a ground-based exit (as in the OG styles). And to flip the argument itself, if someone designed an exit that was mounted to the side of a wall, we wouldn't suddenly decide that only Climbers should be able to exit!

The point is that it seems far better to make decisions based on the effect of the trigger area, rather than any one of an infinite number of visual design possibilities. So, we decide "for each interaction, exits remove the lemming and increase the rescue count by 1"; no further decisions about exit behaviour need to be made. We have a simple, easy to grasp, consistent concept that's applied gamewide and doesn't rely on a list of factors.

Does it matter, then, what exits look like? I guess I'm arguing that it doesn't, but only because they can look like anything in the context of an engine that allows custom style creation.

But, perhaps that doesn't really answer Dullstar's comment. For that, let's take it right back to the OG styles, and tackle the question of "physics = appearance" head-on.

The OG exits are all grounded, all have a doorway, and most also have steps. Can safe interaction in midair and via DD be justified with these designs?

For midair situations, the steps themselves provide something for the lemming to "land" on - Jumpers, Reachers, Fallers, Gliders, Floaters, and indeed any other lemming state should be able to physically interact with the steps.

For DD, unsafe fall distances are trickier to justify, admittedly, although my instinct remains that safe interaction is correct. I guess I imagine that the exit itself provides anti-splat properties (in the same way that an anti-splat pad or an updraft does), or that the doorway pictured is some sort of vortex which draws the lemming in (or at least resets the fall distance at the point of contact). Maybe exits are just "magic", and allow the lemming to defy gravity simply by being in the immediate vicinity!

[Turrican]

Here is, some additional input , about the direct drop topic. First of all , I am not against direct drop by itself. I had used it in the past for various solutions in games and engines , that supported it, like Lemmix. And I have posted also , some of these solutions in my youtube channel , and also, I am completely fine, with the current situation of having engines that support it like Superlemmix , and engines that don't support it, like Neolemmix or Lix. The reason , that I am opposed to this feature , on this thread, is because , I don't want it , to become, the new community standard , for the reasons , that I have mentioned in my previous post , in this thread.

So, the following days , after that post , I did some testing , on various levels from various levelpacks ,and I  created , a number of "what If" backroutes, which are practically , how these levels, would be played , and solved, if Neolemmix allowed direct drop. I also saved , the replays of these. Initially , I didn't want to post these replays , but I've decided to post them here, because they cover various types of backroutes, that can become possible , because of direct drop, and also, these replays , will be useful , for anyone , that will decide at the future,  to convert a pack , that contains any of these levels to Superlemmix.

Here are the levels. Comments about their backroutes are in the spoiler tags :


Revenge of the Lemmings 2022 - Armaggedon 5 - Getting Stuck in a Cave :

Spoiler

Completely different to the , intended solution , and it completely ignores the lemming at the lower part of the level.

Revenge of the Lemmings 2022 - Armaggedon 40 - Stroke at Retiment Age :

Spoiler

This is a characteristic example , of the type of backroute , that I don't want to see in custom levels. It commpletely skips , the whole puzzle.

Lemmings Plus IV - Insane 11 - Tower of Confenction :

Spoiler

Another characteristic example . It skips the entire puzzle.

Lemmings United - Pasifism 13 - Vertical Voyage :

Spoiler

Another backroute , which skips the entire puzzle

Lemmings United - War 21 - Center of Attention :

Spoiler

This solution skips , the entire part of the puzzle on the right part of the level.

   
Lemmings United - Bonus 32 - Shadows of Ourselves :

Spoiler

This backroute skips the last part of the intended solution , and also a very clever trick at the beginning of the solution.

Lemmings Plus Omega II - Spiky 6 - Plytime :

Spoiler

Another characteristic example . It skips the entire puzzle.

Lemmings Quartet - Quartet 7 - Is the beach day cancelled then ? :

Spoiler

It skips the last part of the intended solution.

Lemmings Reunion - Nightmare 11 - The Chosen Lemmings of Ra :

Spoiler

It skips entire parts of the intended solution.

Clammings - Insane 13 - Over and Under and Out : 

Spoiler

This backroute skips the part of the intended solution , that takes place on the lower part of the level.

Casualemmings - Hyper 2 - The joke's on You : 

Spoiler

This backroute skips the part of the intended solution , that takes place on the lower part of the level. 

Links for the packs :

Lemmings Plus Series : https://www.lemmingsforums.net/index.php?topic=1922.0
The linked post , also contains links , for the replays of the intended solutions.

Revenge of the Lemmings 2022 : https://www.lemmingsforums.net/index.php?topic=5963.0
The linked post , also contains a link , for the replays of the intended solutions.

Lemmings Reunion : https://www.lemmingsforums.net/index.php?topic=2175.0

Lemmings United  : https://www.lemmingsforums.net/index.php?topic=4131.0

Lemmings Quartet : https://www.lemmingsforums.net/index.php?topic=5796.0

Clammings : https://www.lemmingsforums.net/index.php?topic=3642.0
The zip file of the pack , also contains replays of the intended solutions.

Casualemmings : https://www.lemmingsforums.net/index.php?topic=4710.0


About my pack : When I converted my pack to Superlemmix , I had already , an idea about which levels had even a small possibillity to be affected , and I took the needed meausurements. There is also , a possibility , for these levels , to be impossible to be backrouted with the use of direct drop, even without these measurements , but in practice , you never know!

[namida]

At present, it's possible to design a swirling vortex and have it be an exit. It's also possible to design an exit with steps and a doorway. Which design should the physics match?

Probably the one that was used in L1, was used in the L1 spinoffs, was used in L2, was used in L3, was used in L3D, was used in Revolution, was used in the overwhelming majority of fanmade content... the fact that it's technically possible to make a completely different graphic (as is the case for any possible graphic; you could make a fire trap that looks like a pickup skill technically if you wanted to, doesn't mean it's a good idea, let alone that it's an argument for giving fire objects pickup-like behaviors) doesn't override this huge base of tradition.

I say this as the creator of one of the very, very few styles that did actually at one point use an intentional, designed-for midair exit (namida_sky).

Based on the analysis, it's very, very clear at this point that direct drop was a glitch in the official games. It's a bit less clear whether it was intended or just overlooked in Cheapo, but Cheapo is not an official game, even if it did have a very significant impact on the direction of fan content, and at any rate it was generally treated as a glitch among the community in the Cheapo days (and from a progamming point of view, this makes sense - I can definitely see how, if there wasn't already the existing background of direct drop having been discussed and regularly exploited in challenge solutions, a lemming falling onto an exit trigger is something that wouldn't cross a dev's mind). It really feels like your constant revisiting of it is based purely on "let's make more stuff happen, and justify it later" (which yes, is something I was also very guilty of during the earlier days of NL) than actually considering whether it adds value and/or makes sense.

[WillLem]

TL;DR (this is a long one!)

Thanks everyone for your patience and for indulging me in this discussion so far, I understand that it must be frustrating to go over old threads again, and I appreciate the opportunity to add my voice to the discussion at this late stage.

@Turrican - I highly doubt the community standard will change anytime soon! And, changing the accepted standard is certainly not the point of this thread. I'm just interested to discuss and research the possibilities of DD and offer an alternative perspective on it, that's all. Thanks for your valuable input and for the replays, please can you also post a zip of the levels themselves?

I am completely fine, with the current situation of having engines that support it like Superlemmix

Honestly, I think I needed to hear this. As much as I want SLX to be a true representation of my own ideals, I also do want people to like it. It's impossible to please everybody, but I'd also be disappointed if something like DD put people off even giving it a try.

Not that I have to worry about that in your case; you've converted a pack of levels, which I very much appreciate!

The reason , that I am opposed to this feature , on this thread, is because , I don't want it , to become, the new community standard

Maybe I underestimate my influence on the Forums, but... I highly doubt that DD would ever become the community standard, having been so comprehensively rejected previously. What's far more likely (and what I hope) is that it will be accepted specifically in SuperLemmix as an alternative, whilst continuing not to be a feature of NeoLemmix, Lix and others which are far more likely to continue to follow the community standard.

This topic is more about simply researching the mechanic itself, exploring both the possibilities and potental problems of it, and re-opening the discussion I didn't have a chance to take part in previously, rather than an attempt to redefine the community standard.

Perhaps your concern is that I mentioned in our Discord PM that I don't want to suggest potential backroute fixes for DD (even though I can think of plenty!) because I don't want to perpetuate the idea that DD is something undesirable that needs to be fixed. I should be clear that I was specifically referring to its implementation in SuperLemmix, and more specifically in response to Jeremy's suggestion that we make a special exit for it that doesn't allow DD. Implementing such a feature reinforces the message "DD is bad, here's a way around it if you don't like it". Instead, I'd rather continue to promote the idea that DD can be an interesting game mechanic, particularly for challenges, talismans and alternative solutions (all of which are game elements I hope to bring to the fore in SuperLemmix). I think that there is very little chance of SLX replacing NL as the community standard, but it may certainly attract speedrunners, challengers, and players who prefer the more traditional gameplay style. This represents a small proportion of players in an already very small community, so the threat to the status quo is relatively low, if existent at all.

NeoLemmix is very likely to remain the community standard for the forseeable future - someone would have to come up with something pretty darn special to outright replace it, and discussions around game physics, etc would likely resurface and be re-evaluated at that time anyway. Or, in the scenario that I'm called upon to help with NeoLemmix development after its final release, I absolutely would not make controversial changes to the physics, I would respect its development history and maintain the status quo regardless of my own preferences.

It seems worth making the above clear at this point, since there has been talk of me making further contributions to NL development, and I really don't want the community to worry that I would do anything to upset its gameplay mechanics, which I'm aware result from years of careful tailoring. As long as namida is active on the Forums, I would likely never do anything without his specific approval anyway, and if he's unavailable I'd go to Icho and Simon next, as these are people I know namida trusts.

Hopefully, with that said, we can continue to discuss DD without any such concerns.

So, the following days , after that post , I did some testing , on various levels from various levelpacks

Thank you for sharing these! You've been quite thorough - am I correct in assuming that these are the only levels you found in those packs which have DD backroutes? My own research has turned up 3 in 120 (or 1 in 40) levels that can be backrouted using DD; I wonder if your research also reflects this figure.

Also, could you add a zip of just these levels to your post for quick access? OK if not, I can download the full packs if needs be. Again, thanks for doing this valuable research!

the fact that it's technically possible to make a completely different graphic ... doesn't override this huge base of tradition.

Fair point, but the fact is that we can now make exits look like anything. Whilst that doesn't necessarily mean that physics must be reconsidered, or that any such reconsideration is necessarily a good idea, it does at least provide the opportunity to reconsider them with arguably good reason.

And I say that as someone who almost exclusively uses the OG styles in levels, and is a strong advocate of traditionalism in the Lemmings game!

Also, I did specifically address the OG exits in my previous post (i.e. the visual designs of these exits make sense for midair exiting (definitely, because steps) and direct drop (in my opinion)).

Meanwhile, the decision was taken to not allow Jumpers to exit in NeoLemmix, so I'd ask why not, if the images clearly show steps that can arguably be "landed" on? Or, to state more explicitly what I'm getting at here, why should visual design be a consideration for Direct Drop but not for Jumpers?

Based on the analysis, it's very, very clear at this point that direct drop was a glitch in the official games.

I have very little doubt that it was, but we aren't always necessarily concerned with following the blueprint set out by the original games (including what should or shouldn't be considered a bug), that's something we can in fact all agree on.

Having said that, I'm aware that DD has been discussed at length on its own merit, which brings me to your final comment:

It really feels like your constant revisiting of it is based purely on "let's make more stuff happen, and justify it later" (which yes, is something I was also very guilty of during the earlier days of NL) than actually considering whether it adds value and/or makes sense.

It's fair to call me out on this because I do bring it up a lot, but I was late to the party and never had a chance to add my voice to the conversation. Honestly, I'm grateful that people are responding to this topic and indulging me. Other Forums might have shut it down as a dead topic, so the fact I'm being given allowances here isn't lost on me, I assure you.

And yes, it seems that DD has been discussed extensively, and its bug status wasn't the only deciding factor in its eventual removal from NeoLemmix - this is a good thing. However, I still think there's room for challenging the narrative that it's definitely a problematic game element; I believe it both makes sense and adds value, and that's really what I'm hoping to demonstrate with this topic.

Hopefully Turrican posts these levels he's found, and I'll definitely give them a look and assess the damage for myself. I absolutely do trust experienced Forum members, engine developers and level designers in their opinions on what's right and wrong based on prior research, and there's lots I've conceded over my few years here on the Forums so far. But, there are some things I need to actually see for myself before I can fully understand the other side of the argument, particularly if it's something I'm invested in and feel strongly about.

Full disclosure, it's probably an ADHD thing - I get super excited about an idea and, as you've quite rightly pointed out, I just want to see it in action first and then probably do feel the need to justify/discuss it later. I understand if that's sometimes frustrating for people who've very much "been there, done that", and I appreciate your patience - for me, this is an interesting and stimulating discussion!

[Turrican]

I have attached a zip file with these levels.

As , to how I found them, I just browsed these packs for levels that had open spaces above their exits , and if the skillset looked suitable , I tried to backroute them. It was relatively quick browsing, which didn't cover the entirety of these packs. Also , some of these levels , I had played them at the past , so I knew , that they would be able to be backrouted , this way.

[namida]

Yeah, I should be very clear that I have zero intention of forbidding such discussions, regardless of how I personally feel about either the subject itself or the repeated discussions on it. If no one wants to engage, they're always free to not respond or even read it, after all.

[Simon]

Drowning lemmings exactly during transition from falling to drowning.
Burning lemmings, exactly during transition.

You can't have two different trigger areas on the same pixel in DOS.
inject a change exactly between when it checks for each type of trigger

Ah, thanks, it's enough for my research to know that you can't become drowner and exiter in one frame (without injecting things with a debugger). No need to study the code to handle a mid-update change of trigger areas. I'll allow the code to rely on the normal invariants of the program. Reason: My angle of research is less to understand DOS L1 instruction-by-instruction, but to guess what the devs thought essential.

Mindless already told me last month that drowners and exploded lemmings can't exit. You reminded me that DOS L1 has only one trigger area per location, therefore you can't become burner-into-exiter within one physics update either. This covers most direct drop potential around dying lemmings. The single possible known one is faller-to-splater-to-exiter within the same physics update.

I don't know if a burner might exit on a later physics update. You'll have to replace the fire trigger with an exit trigger, using the bug where a falling ohnoer restores his saved trigger area elsewhere. I believe that this won't turn the burner into an exiter. I don't have time to research this; if I catch Mindless, I'll ask him.

-- Simon

[namida]

To further elaborate on that - you can't have a mix of different trigger types (or two objects of the same type, in cases where the individual object matters, ie: traps) in the same 4x4 area (as in, 0,0 to 3,3 must be a single type; 4,0 to 7,3 must be a single type, etc; etc; not all trigger areas must be 4 pixels away from other ones). This includes "none" - you can't have a grid-aligned 4x4 area where the top half is a trigger and the bottom half is not - and blocker fields, including the "neutral" area inbetween the force-left and force-right, also follow these rules + exclude other triggers (the way that blockers near objects is handled is the cause of the moving-trigger-areas bug you mentioned).

This is the source of the bug where exits would sometimes appear to not function depending on their exact vertical position in custom levels - the trigger area gets grid aligned and moved upwards slightly (and as such, if you place a builder near the exit, you can actually still use it).