To start off, let me wax poetic about Lemmings Migration again, which is still one of my favourite packs to this date (yes, despite it only featuring classic skills, and that's saying something, coming from me!
). And for a while, I couldn't quite figure out why. Sure, it had great puzzles, but a lot of other packs do as well. Yes, the levels where smaller, which is not as annoying as having to navigate huge landscapes on the majority of levels of a given pack. And yes, the levels looked pretty colour-wise, without depicting anything specific (as in Arty's SubLems or Gronkling's new "Rodents", for comparison). So what was it that made this pack feel so cohesive to me?
And then, while watching kaywhyn's recent LP of it, I realised: I only uses a few different tilesets!However, at the same time, it doesn't just take the easy route and stick to the original ones either. Instead, it uses a clever mix of original, ONML, and custom tilesets that all form a nice sequence of levels.
- orig Pillar
- orig Fire
- orig Dirt
- orig Crystal
- orig Marble
- ONML Brick
- ONML Snow
- namida Mineshaft
- namida Machine
No tileset mixing, except for the very last level.
First of all, in my view, this gives the two tilesets by namida used so frequently throughout the pack loads of "street credibility", because they are just thrown into this mix of original DMA tilesets, and yet they feel like they've always been part of that group.
This is different from the Lemmings Plus Series, for which they were originally created, because in that context, namida always made a couple of tilesets for one given pack and then only used those tilesets throughout. But of course, especially the earlier ones looked a lot more "rough" than original DMA tilesets, and therefore probably wouldn't have blended with them nicely if a DMA and an early namida tileset had followed each other back-to-back in two consecutive levels.
But the much more important discovery for me is here that giving your pack this a-priori limitation of "I'm only going to use these tilesets" can establish the structure and visual identity of your pack much more strongly than using a bunch of different tilesets in a seemingly random mix, as many of the larger packs do (Lemmings United, SEB Lems, my own Lemmings World Tour etc.). Those huge beasts can be pretty overwhelming, and it's easy to perceive them as chaotic if you as the player don't happen to get certain references that make the logic of why those packs are structured the way they are more apparent to you.
There are a few packs for which I've noticed this tileset-restricted design strategy - Lemmings Migration is just the best example for me. But I also remember e.g.
SQRon's Snack Pack mainly sticking to ONML tilesets. Of course, as said earlier, there are the Lemmings Plus packs, each of which has a unique visual identity because each of them has its own group of tilesets. That strategy is just a little "all or nothing", because the styles were new every time, and since they're all that appear in a given Lemmings Plus pack, either the player likes them, or he/she does not.
I wouldn't really consider packs that only use original DMA tilesets (i.e. from classic Lemmings and ONML) part of this group, since a lot of them were created when no other tilesets were available yet. Meaning,
the selection of tilesets to be used for the pack as a whole wasn't necessarily chosen by the pack author. By the same token, I wouldn't count Lemmings 2: The Tribes inspired packs either, like Nepster's "Return of the Tribes" or my upcoming "Lemmings: Hall of Fame", since the L2-inspiration pre-dictates what tilesets you can use (unless you want to create all new tribes of your own).
That said, there don't seem to be that many L2-inspired packs yet, and only using those tilesets exclusively will therefore still give your pack a strong and easily recognisable visual identity.
Which is kind of the point of using this strategy.
Since LOA only uses classic and ONML tilesets (plus two of my own, not custom tilesets by others that I deliberately selected as my "chosen few"), and Lemmings: Hall of Fame "made that choice for me",
I would argue that I personally have never actually used this strategy consciously yet. 
But now that I've become aware of it, I'm all the more eager to try it out myself! I don't think Lemmings Universe, which is my idea of an astronomy-themed pack that got an overwhelming amount of support when I recently put my various ideas up for a poll, will be the right place to try this approach, though. Instead, that pack might become somewhat of a "Lemmings Space Tour", with levels trying to emulate the geology of other planets as much as possible in a visually convincing way. Meaning: Most likely by employing tileset mixing.
In contrast, I think it's safe to say that tileset mixing should be used very sparingly with the approach presented here! Why? Because it blurrs the pack's visual identity - and also that of the individual tilesets - blending them into one. Any piece can now appear on any given level, irrespective of what style it is officially part of.
Naturally, when you're aware of that, you can use this to your advantage, too: For instance, namida has already pointed out that Lemmings 3D doesn't really "think" in graphic sets; you have the different menu screens to establish the "styles" (and the two types of music for each style), but the pieces themselves can be shared between styles much more liberally. (Especially with the Army and the Golf tileset, I often have difficulties which piece is part of what.)
Finally, note that a top-down restriction, given our vast pool of different tilesets, will always feel somewhat artifical at this point.The original DMA designers didn't have this problem; they were probably in a situation much more akin to namida when making Lemmings Plus: Design a couple new graphic sets, then make a new pack using those exclusively: None of the orig tilesets re-appeared in ONML, for example.
Now however, if we artificially limit ourselves to a specific selection of our favourite tilesets for a given pack, it would make sense to "justify" this in-game with a theme that has these restrictions as well.
Think back to the Lemmings 2: The Tribes example: Anything that involves established factions, tribes, people groups etc., anything that is split up into a few, vastly different looking regions - such themes will probably benefit from an a-priori tileset selection.The first example that popped up in my mind was indeed to make a pack based upon the Lego Bionicle franchise, settled on an island with six different climate zones linked to the elements (which are the "standard" elements Fire, Water, Earth, and Air, but Stone and Ice are added to that and treated as if they had always been part of that group - everything in the Bionicle Universe comes in sixes, basically). Then you would have "Light" and "Shadow" as less common elements on top of that. And just like that, one could easily create a "Lemmings of Mata Nui" pack following the philosophy presented here with a consistent visual identity.
Alternatively, the "
Lemmings: The Gathering" approach (which was one of the candidates in the poll that was won by Lemmings Universe) would also work, with Magic: The Gathering's five colours of mana, and/or with certain colour combinations (either the 10 two-coloured ones or the 10 three-coloured ones) being represented by different tilesets.
For
Lemmings Universe, it would only work e.g. within the confines of our own solar system (as I briefly outlined in the poll thread already): One tileset per planet, and then, anytime that tileset re-appears, you know where you are again. But while we do have a Martian tileset, a Tree tileset for Earth, and a Rock tileset for e.g. the Moon or the Asteroid Belt, it's hard to make four visually distinct graphic sets that are all supposed to represent gas giants.
Don't worry, I still want to and I'm still going to do Lemmings Universe - I'm just genuinely intrigued by this approach now and feel like I want to try it.
