The "free" of it definitely made a difference, but there was also a certain lack of concern about the shape of the resulting product that made ZZT feel differen

The "free" of it definitely made a difference, but there was also a certain lack of concern about the shape of the resulting product that made ZZT feel different and more flexible than the preceding tools when you tried to work with it. It had built-in, non-configurable stuff for action gameplay with some puzzle elements. It did very little with regards to assets, unlike contemporaries - no animation sets or hitbox definitions or music importer or similar ideas, really just a level editor. It was just not good in terms of shelf-space bullet points - nothing flashy about it. If that was all that it had, it would not be a hugely differentiated game since the level editor concept had been around since Lode Runner. But it also had code, and code integrated in a way that was well documented and tutorialized. It had scripted behaviors, functionality to manipulate and transform the board, and hypertext dialogues with event flag toggles. Having the hypertext and flags may have been the one most critical part of all this since it really enabled choice-based interactive storytelling techniques, and that's a huge deal even if you didn't have enough of an algorithms background to try abusing the other functionality to make a platformer or whatnot. And so I think the longevity of ZZT really rests in the way that it guides you towards code, because it has so little else to offer.