Regarding the animation: watch this clip of direct drop in WinLemm (most likely the platform where I discovered it was possible - through trying it out, I hasten to add). To me, this looks like they're bouncing in midair - they never actually touch the floor. Thoughts?
Based on the video, it looks like WinLemm's version of direct drop is different, and so maybe in WinLemm the lemmings can even exit with the exit completely in mid-air, unlike other versions like DOS where the lemming still has to land but then exits instead of splatting (or if you can see the DOS game's programming, it is first turning the lemming first to splatting and then later to exiting, all before updating the display so you only see the exiting).
I'd love for you to test what happens with the mid-air exit in WinLemm. Based on the video, it does look like WinLemm can suck falling/floating lemmings in mid-air, unlike most other versions out there.
It still doesn't change the fact that the game never made it any remotely a core part of the experience. The fact that it never worked on so many other versions, including Amiga being one of the original versions, also shows that it wasn't something DMA intended. Lemmings 2 also adds emphasize to the land being core to exiting by adding so many tribe-specific elaborate exit animations that only makes sense on land.
Direct drop is a source of backroutes in well-designed levels and its puzzle potential is limited to misleading levels, because there is no reason for a player unaware of direct drop's existence to even think of trying it.
See this post, above. The only reason I'm even aware of direct drop as a game mechanic is exactly that: I tried it out, and it worked!
That's also how most glitches are discovered. Often even by accident as the player was doing something unrelated but managed to luck into some combination of circumstances and moves that were sufficient to trigger the glitch. "Someone has discovered it" is by itself not a good argument for making something part of the game.
And in the case of direct drop, it's not even true--the people who only played the versions of Lemmings that do not support direct drop, would actually have come to the opposite conclusion had they tried out what you tried out.
And furthermore, if SuperLemmini makes these things optional rather than forcing players to adopt a particular philosophy, then it becomes an even more attractive option to yet more players.
You are oversimplifying the situation. For what you are proposing to work out, SuperLemmini still needs to have enough levels that cater to the people firmly on the NeoLemmix philosophy. If instead most levels are just about, say, timed bombers, without other redeeming qualities for example, then making them untimed just leaves boring levels. In other words, you want to encourage levels that include things which are a positive for fans like WillLem but strong negative for the NeoLemmix fans. Allow play assist tools like making bombers untimed, merely turn the negatives in the level into zeros, it doesn't automatically turn them into positives. The net sum of the level's worth/enjoyment is still lower for the NeoLemmix fans compared to the WillLem fans. And if the levels tend to lean heavily towards the anti-NeoLemmix stuff, then that resulting lower sum of enjoyment might just be too low for NeoLemmix fans.
Another thing to look at is how NeoLemmix and Lix fared. In this case, both actually align a lot on most of the philosophies. You don't end up with equal usage of both.
It's fine that you want SuperLemmini to go in a different direction than NeoLemmix, but it's naive to think that merely having settings to add back the play assist tools would suddenly make SuperLemmini also attractive to NeoLemmix fans. It's going to be mainly down to the levels, and we know at least some of the levels will not be appealing to NeoLemmix fans, which doesn't exactly encourage them to commit very deeply to using SuperLemmini. There's also the fact that NeoLemmix now has a greater variety of skills and elements, which may well be a turn-off for some, but it can also add more impediment for people creating SuperLemmini levels if some designs can only be implemented in NeoLemmix. (Well, there's also the ridiculous fact that SuperLemmini still doesn't even have its own level editor, but let's forget that point and assume it will soon have one, for sake of its long-term survival.)
Realistically, if you want to use SuperLemmini to promote the stuff that turns off NeoLemmix fans, the outcome is that SuperLemmini will attract a different set of people than NeoLemmix. I think that's fine. It's good to have variety and just like Lix nowadays serves the niche of multiplayer, it's sensible to have SuperLemmini serves those who desire some of the gameplay aspects that NeoLemmix had moved away from. Regardless, the people who are currently very happy with NeoLemmix will likely stay with NeoLemmix given that it is more familiar, supports more skills/elements, and the things you suggested so far to add to SuperLemmini are at best don't-cares and at worst turn-offs for them.