Yes, I've just stopped after 5 levels so far, due to the anticipation of changing physics. For example, I just noticed one of my replays broke due to the introduction of the "Looker / Dangler" state for the Slider. (No problem, I just had to redo the last two assignments - a Basher was now being assigned slightly too late, only after a lemming had turned around.)
As long as only few levels are affected, these things are easy fixes; they become much more annoying however if they need to applied to an entire pack retro-actively.
Hence, I'll continue to focus on NeoLemmix for the time being - specifically, Lemmings: Hall of Fame. The successor to that (working title "Lemmings Band Camp") will then be started officially once you announce a somewhat stable state for SuperLemmix.
In the meantime, feel free to try out different physics for whatever you like. For blocking Swimmers, for example, I can imagine lots of potential uses - also lots of potential backroutes for existing levels.
This is where I agree that, especially when it comes to the development of a new engine, existing levels should receive a lower
priority. Since what's possible in the future shouldn't be hindered by what has been done in the past, especially if the primary argument for the past behaviour is just "this is the way we've always done things, now it's too late to change it". Such as the Stoner being a suboptimal skill when compared to the Freezer. (Added side benefit, "Slowfreeze" in SuperLemmix makes much more sense when it actually turns the lemming into a Freezer, rather than a Stoner
).
So I guess Nepster's statement that existing levels were flat-out
"unimportant" came off as just a little too radical, since it implies the relevance of existing content is 0, rather than, say, 50% to 70% of the relevance of potential new packs, which would then be created specifically for SuperLemmix.