The Stack

Spoilers follow the break.

After a brief introductory chapter or two, Escape from Summerland‘s central gimmick turns out to be that it lets the player switch at will among three characters who exist in the same environment, but perceive and interact with it differently. So, kind of like Suspended, except set at a funfair, and the three controllable characters are a ghost, a robot, and a monkey. The robot describes everything in functional, jargon-y lists, and its ability to interact is limited by what parts it has installed. The monkey’s text is excited and full of monkey emoticons (irritatingly so, at times), but it doesn’t know what anything is, and it refuses to obey commands that it finds boring. Neither of these two gives a clear picture of the world. For that, you have the ghost, who describes things more or less the way you’d expect in a text adventure, but who can’t interact with the world directly at all.

All this is lovely stuff, and furthermore, it’s the sort of thing that’s best done in text rather than graphics. So I really look forward to playing this game when it’s finished. It definitely isn’t yet, and I’m really disappointed.

There are disambiguation problems, as when the game asks you which of two objects you mean, and then lists the same thing twice. One room is missing a description when you’re the monkey. One very important object — a balancing pole — is undetectable without using the word “all”. I gave up when I was unable to exit an elevator. There was a button that was supposed to open the door, but although it was still mentioned in the room description, it seems like the actual button objects got left behind when the elevator moved. One other person reported managing to get past that point, but couldn’t remember how. And that’s the state of the game after being patched once.

In addition, there’s one puzzle that I’m pretty sure I solved by accident, and didn’t understand at all until discussing it with others afterward (and possibly still don’t understand really): a matter of opening a mirrored cabinet with a spring latch by making the robot attack it, which it’s not willing to do when it can see its own reflection in it. This might be just a matter that would be clarified if I had seen all the text I was supposed to see — in particular, any text at all indicating that the cabinet was open would have helped — but this strikes me as a fairly intricate puzzle that would be very difficult to solve intentionally without a great deal of feedback, which the elliptic perceptions of the robot prevent us from getting. So there may be difficulties here beyond the bugs. But I still say it’ll be a grand game once it’s incorporated all the user feedback it’s getting from the Comp.