So, as I said, I want to reply to the general argument, rather than focusing on specific levels. It's true that a couple of levels don't work with fixed SI (
The Hotel in Hell can probably be fixed even though the fixes suggested so far have been shown not to work;
The Continuum Hypothesis certainly doesn't work without drastic changes). Still, I'm not convinced by the general argument that removing VSI is a good thing and the individual levels that don't work are the main problems standing in the way.
The main "anti" argument seems to be that VSI isn't the "core" of gameplay. But in the first place, it's not easy to say what the core of a game is. And many games have supporting features that are not part of the core, but make the game more enjoyable, whether it's by adding variety or aiding usability or both: bombs in Touhou, powerups in Super Mario World, bonus stages in Sonic, etc.
As for
Lemmings and
Lix, while giving skills to lemmings is certainly part of the core, so too is the fact that they wander back and forth when not performing skills. That immediately makes spacing important.
This article, recently linked on the forums, suggests that the makers considered traps part of the core, and setting SI to compress two lemmings so only one is killed is a major mechanism for getting past a trap.
But really, the main question should be whether keeping VSI makes the game, overall, more enjoyable or not. I do agree that fiddling about with the SI to get a particular spacing can be a bit irksome. But Nepster makes a good point: fiddling about with skills can be as bad or worse. And it's not something you can really get away from, because altering spacing is an inherent part of the game, whether it's to get one lemming past a trap, give a builder enough space to finish building a bridge, get enough lemmings into a dig pit before it goes all the way down, and so on and so on. If you take away VSI as a tool for these things, players are going to have to do the same things with whatever tools they have left. Focusing the discussion on levels that
cannot be solved without VSI misses the point that it aids on a lot of other levels. Changing SI is a relatively clean way to alter spacing, because if you can work out the spacing you need, you can just enter it as a number. While it's true that this is drastically different from interacting with lemmings by clicking on them, so too is framestepping to move the assignment of skills to an exact time, and that was judged to be a good thing to add to
Lix to aid execution.
* * *
I also want to talk about discoverability. Simon's first point is correct: as-is,
Lix's UI does not make it obvious what the SI buttons do. That is a reason for improving the UI, not (in itself) a good reason for removing the feature. I have suggested an icon of two lix with a double-headed arrow between, a standard symbol for "spacing".
Simon's second point, "They aren't needed in 98 % of the levels, and don't scream "use me" once they are", is thinking in the wrong direction.
Lix is a puzzle game, and telling the player "this feature is needed for this puzzle" is providing a hint before they've even started thinking about the puzzle. It's as if we proposed to have every level start with the skill you need first already selected. Sure, it makes the program easier to use, and might sound on paper like a good idea, but it hurts the enjoyment of starting from a hintless puzzle and a given set of tools and working out for yourself which of those tools will help you solve the puzzle. If we solve the issue of making it more discoverable that VSI exists, then the issue of players being stuck because VSI is needed for such-and-such puzzle and they don't know it exists automatically goes away.
For the same reason, the suggestion to make fixed SI the norm and set the SI to variable only when needed for a level is absolutely terrible. First, it necessarily provides an unwanted hint; secondly, it takes away a tool that is often useful even when not necessarily required. The ability of level designers to fix SI on individual levels for backroute prevention is fine, but other than that, ultimately we have to decide whether we (as a whole, on balance) think the pros of VSI outweigh the cons or they do not; in which case VSI should either be fully allowed or never allowed.