CE 1.1.2

During play, in the panel at the bottom of the screen, above the skills:

Number of lemmings turns red when it's not sufficient for winning. Okay.

Clock turns red at arbitrary time (1 minute? 15 seconds?) and turns purple when it's not sufficient for winning. Instead, ditch the arbitrary threshold. On different maps, different thresholds will feel like too short anyway. Consistently use red for insufficiency.

Save requirement is printed in blue? Blue makes it hard to read; the digits 5, 6, 8 look similar. And the blue makes the panel look more colorful than necessary. This removes the meaning of other colors. Practically every number is now a funky color at whim.

Clock is yellow when it is a time limit. Okay, we must pay attention. And it matches the yellow hourglass. (As said, don't put a blue save requirement; it makes the yellow blend in, the yellow becomes yet another irrelevant color next to the blue.)

Lemming count is yellow when there are neutral lemmings. Why is this important? Why do we have to pay attention? I would understand it becoming yellow when there are only neutral lemmings left.

Lemming count also becomes cyan in some cases? When and why?

-- Simon

CE 1.1.2
NL 12.14

After playing one level, NL displays the postview screen. It tells you:

Lemmings saved in your recent attempt
Lemmings saved in your record
Time taken in your recent attempt
Time taken in your record
Skills used in your record

It should also tell you the number of skills assigned in your recent attempt.

-- Simon

Welcome to the forums!

Thanks for sharing the level pictures. It's always lovely to see new ways to use these old tilesets. Nice artwork! Every author has his own unique style.

With "WinLems", do you mean WinLems by Warren Gaylen (from 2007, which looks like it accepts .DAT files) or do you mean Lemmings for Windows 95 (whose directory is also called Winlemm)? It's a nastily similar name. I believe neither of these two games has active players here. From early 2025: Giga's topic about Warren Gaylen's WinLems

You can attach your levels. Or you can upload .DAT packs to Mindless's Lemmings Level Database.

-- Simon

Congratulations to the winners! Many solid levels.

Epilogue had the most interesting solution of the three. During the poll of the final 3 levels, it seemed a clear winner. Now it tied. If I had to guess a downside ... maybe it's less elegant than the other two because it relies on more complex mechanics (permanent skill removers)? Maybe the solution is a tad too hard to find? Doesn't matter, lovely level, and a deserved top place.

It's Crazy Sports Day: At first, I was surprised to see it do so well. I had already seen all its tricks elsewhere. But in hindsight, it makes sense. It's a backroute-free showcase of several different tricks and it needs both planning and insight.

Industrial Incident is a classic case of hiding a simple solution in plain sight. Nothing is distracting, nothing is unfair, and yet it takes time to solve.

* * *

I had soft spots for these:

Armani's Underroot. I may have a soft spot for the rock tileset in general, but there is more here. One technical trick, and you feel that the entire level is made to support it. Before you know the trick, everything looks beautiful -- the visuals reminds me of Utopia by Pieuw -- and after you see the trick, it's beautiful and functional.

Icho's Tower of Sacrifice. The closest relative may be Geoffster's Parking Garage, doing everything with terrain removers when it looks like you need blockers and builders.

The Split by Dexter. Looks beautiful. Has a nice mix of easy pattern recognition and serious planning to tie the solution together.

All 4 maps by Giga. After you memorize what gadget does what, you get interesting puzzles every time. Lawful Neutral I liked the most.

Building, Fencing by Dexter. A well-hidden fair solution.

Crane's A Step Down With Difficulty, kaywhyn's Organic National Park, Armani's Let's Go Spelunking (after removing the horrible glare from the torch).

-- Simon

Hi! I believe that the first Github link should point to VorticonCmdr/lemmings, not to the github.io.

Networking mode against myself works in Chromium after enabling the HTML-in-Canvas global option, which is disabled by default.

On your readme, you explain how you support efficient rewind in singleplayer by taking savestates and recomputing. This is the most performant solution that we know. Lix was the first to do that in multiplayer, then NeoLemmix implemented it for singleplayer, and then Lix offered it in singleplayer.

Your multiplayer assignments during netplay have significant delay, over half a second. I think that such input delay is a problematic way of combatting lag. Clones (another Lemmings-like) and Starcraft do it like this, too, but it's more acceptable in Starcraft (where commands should happen ASAP) than in Lemmings (where it's a problem when commands happen too early). Have you considered instead to rely on the efficient recomputing from singleplayer? When the local player receives a delayed command from the opponent, the local player inserts it into his own history, then recomputes. I suspect that recomputation makes it harder to catch cheating, but it makes for better flow.

There is a third, novel way of combatting lag: Neon Swarm (Lemmings-like in Lockstep Arcade) makes terrain modifications and blockers immediate for the local player, and has a ~1-second delay before they materialize for the opponent. From the local player's view, his own young builder bricks are semitransparent until they're old enough to affect everybody.

Do you strive for exact 1991 Lemmings 1 two-player physics, bugs and all? Then you can't use such immediate local impact with delayed remote impact, and I'd suggest savestates and recomputation instead. It really depends on what exactly you envision. E.g., the 1991 physics have a possibly unwanted first-player advantage because his lemmings spawn before the second player's lemmings.

-- Simon

Thanks for trying Lix in these emulators! Happy to hear that it's running smooth in Gamenative.

For Winlator: Allegro 5 (Lix's graphic/multimedia library) allows for a choice of backends (OpenGL, Direct3D, ...) by supplying a config file allegro5.cfg, but I haven't done that myself yet. Maybe that's useful for Winlator. But you don't have to investigate because you've already found that Lix runs well in Gamenative, and you've chosen to focus on that.

I've never designed for phones nor tablets. Some parts of the Lix UI would need a serious redesign for the world away from mouse and keyboard.

-- Simon

When you merge insert mode and regular replay mode, consider changing the default for the user option that keeps/deletes the garbage from keeping the garbage to deleting the garbage.

Definition: In this thread, with garbage, I mean assignments to the same lemming in the future of the newly inserted assignment. Usually, these will desynch because of the newly inserted assignment. In rare cases, they remain useful. (In other threads, it might mean any assignment in the replay that fails to resolve, or that resolves but does something else than originally intended, but those meanings are a different concern out of this thread's scope.)

Reason for changing the default to deleting the garbage: When you push the merged mode to users, this will behave more closely to how assignments behaved before merging the modes. This will be a new additional reason. Another reason is the same as before, the garbage is practically never useful, and you can play around fringe cases (assign climber early/late to preserve/lose it). Another reason is playtesting feedback from across 2 years now: Lix insert mode deletes the same-lemming garbage, I like the deletion there, and I've never gotten feedback that the garbage should remain.

Do you think you want the option (to preserve garbage) for yourself still after you make airclick replay cutting optional?

-- Simon

Hmm, we all found different habits around the LMB clash, i.e., around the two meanings (cut replay, advance) of left click:

• Icho clicks to advance if the click won't affect the replay. Icho presses the hotkey to advance otherwise.
  
• WillLem always clicks to advance. If that click would cut the replay, WillLem first enters the cut-preventing insert mode.
  
• Simon: I always press the hotkey to advance. I only click to affect the replay. I fear clicking.
  
• Guigui, I haven't asked explicitly yet how you prefer to advance by 1 frame. Do you click to advance? Do you press the hotkey to advance?
  

The click to advance seems popular.

WillLem's proposal still reconciles this. WillLem proposes: Create a user option for whether air click cuts the replay, and have only a single mode during play that always inserts. I don't know if this is ultimately best (well, when do you ever know that); it's the best I see for now. It's not clear if Icho is going to like this, but Icho is happy to try.

I agree with Icho that conveying replay mode in the UI is important. We're still in the early bronze age with this, even Lix, which displays "Replaying. Click to play from here." Post what comes to mind. If it's ultimately separatable from merging the modes, we can always split it to separate topics later.

-- Simon

Quote from: WillLem on July 22, 2026, 10:11:55 PM

Quote from: Simon on July 22, 2026, 04:17:27 PMThe default shouldn't remove a useful shortcut (global-future-cutting air click) that does no damage for them: They won't LMB into the map for other reasons.

Not so! There is at least one more reason to airclick, and that is to advance gameplay by 1 frame whilst paused. Sometimes, this is done after rewinding, and with no intention to erase the global future. This is one of the main reasons I engage insert mode, and don't want airclicks to cut the replay.

I strongly suspect that those who don't sympathise with this only use either non-mouse hotkeys or the dedicated panel buttons to framestep.

Ah, I didn't have this LMB-stepping in mind. Yeah, I single-step by hotkey.

It's an odd design history. The usual meaning of LMB into the map to cut/assign. After assigning, it makes sense to advance one step to show the physical result of the assignment. I don't remember who/which program started to advance on airclick in general. At the end of the replay, the airclick cuts (nothing to cut) and advances. And this is common enough to make it natural to rely on it. I believe that namida relies on it, too.

And finally the two meanings of airclick clash in the new mixed mode. After decades.

I'd like to hear namida's opinion.

-- Simon

All right, no undo for now, and the difference between the two surviving modes (airclick cuts global future vs. airclick does nothing) becomes a user option. There will be no mode toggling during play. Good riddance if it turns out well. You get to remove at least one hotkey (the mode toggling) and some iconography.

I think I like it.

Quoteclicking air cancels the replay (by default, it shouldn't)

No, I recommend that you keep the default that airclicks cut the global future.

Argument 1: Even though most users know about this, nobody has complained that it does too much damage in CE 1.1.2 or before. Replaying will become more powerful, but, at the same time, it becomes less destructive because assignments will cut only the assignee's future, not the global future.

Neutral observation: Inserting is advanced. Let's say 30 % of users will insert. If you anticipate that you will insert a lot and you fear about destruction, you'll activate the option for yourself. If everybody were an advanced user, I'd be 50:50 about the default. But that's not the argument.

Argument 2: Another 30 % of users will continue to play with replay mode, but not with insert mode. They'll never LMB into the map until they want to cut. The default shouldn't remove a useful shortcut (global-future-cutting air click) that does no damage for them: They won't LMB into the map for other reasons.

Argument 3: The final 40 % is that invisible portion of users. You rarely catch a twitch stream of them. They don't want replay, they'll always solve from scratch. Replaying annoys them. They always run default options. Don't remove the main way of cutting the unwanted replay from them!

QuoteMore input from other users would be good.

Yes, please post. Even (and especially) if you never use insert mode!

Do you click air to stop the replay? Or do you use a hotkey or click the R in the panel?

-- Simon

One crossroads in the design is: Should cutting the global future be undoable?

You can decide against this for code complexity. Or for difficulty of making understandable UI for it. Or you can allow it, but undo has only depth 1, i.e., you have the most recent previous replay in memory for restoration. Or you can make a full undo-redo stack, either by storing full replays, or by storing the modifications.

Reason behind undo: Powerful actions with powerful undo is nicer than making it hard to reach the sharp knifes that you need all the time. But it remains hard to say how commonly the knife is needed.

UI concern: Undo is one extra button, and it's fundamentally different from rewind. Undo/redo would even be two extra buttons. Rewind is travelling along the 3rd dimension, replay editing is working in the 4th, and an undo stack will be the 5th. It feels only one level below the strength of git rebase.

* * *

If cutting the global future won't be undoable, then I sympathesize with your preference to guard against misclicking air.

I don't think moving the airclick to RMB is good. Reasons as yesterday: RMB airclick clashes with priority invert (where a lone RMB click/hold affects lemming filtering only), it clashes with scrolling (where a lone RMB click/hold does nothing), it's undiscoverable, and it's unconventional (in Lemmings; many other games have RMB canceling).

One idea is to still have two modes, switchable during play:

• Clicking a lemming first cuts the future of that lemming, then, if no clash with different lemming, inserts the assignment. Airclicks cut the global future. This is can be the default mode.
• No cutting whatsoever. Clicking a lemming never cuts, and it inserts an assignment only if it doesn't clash with any lemming, same or different. Air clicks do nothing.

UI concern: Modes are hard to discover. Modes annoy the user when he doesn't understand them completely. Hard to make icons for these modes. I feel like colored Rs won't cut it; the symbol for no-cut mode would rather be struckthrough scissors.

Require something really explicit to cut the global future? Comes at cost of breaking consistency with many Lemmings games, with Lix, but the idea isn't entirely off the table. Hard to sell this to users who rarely insert, but who rewind often and already know not to LMB. The ice is thin.

* * *

Abstract background reading:
Preventing User Errors: Avoiding Unconscious Slips, 2015
Preventing User Errors: Avoiding Conscious Mistakes, 2015
Undo vs. Confirm, when to use what, 2025
Forgiveness Principle: Designing Undo, Confirmation, Error Recovery, 2026

-- Simon

I like the examples from reply #13 more than the examples from reply #12.

I think the author line should be dark blue like the pack and position, to not stand out as much. It's as relevant as the pack and position.

Pack name is nice for livestreaming. People will see excerpts without stream metadata (stream title). Good to orient viewers.

Re long pack names: When you have to trim long pack and position because of long directory names, cut from the left. Or cut from the center. Don't cut from the right. People tend to be happy with the position.

Spacing is a concern for later. Now, if author is missing, there is gap. I don't mind the gap much.

-- Simon

Little time, the hunches are:

The shift in design will be away from {different modes for different tasks} toward error prevention/recovery. E.g., you value how blue R cannot accidentally cut global future. Error prevention may be via modes again, or via undo, or via something.

{LMB into air to cut global future} is accident-prone and obscure. But RMB fixes neither of these problems, and it clashes with many RMB habits where click is no-op: priority invert, scrolling, ...

More when I have time.

-- Simon

Ah, directly edit those in the file. Yeah, I should try that. Thanks!

-- Simon

Quote from: lemchip on July 09, 2026, 12:55:02 AMRegarding vertical scrolling: even Lemmings Redux has been scrolling up and down just slightly lately

You can click-and-drag the border of the NeoLemmix window to resize. It will remember that size across program restarts. The dragging is the only way I know to resize.

Does your monitor/resolution limit your dragging because the NL window is already as big as it can be?

I don't believe there is a user option that nails the window to a given size. Do you think that would help? I've loosely considered to make a topic anyway about this because I prefer a perfectly sized window for livestreaming.

-- Simon