Review: The Crystal Key

Official web page; Earthlight Productions (creators); Dreamcatcher Interactive (publishers).

Review written by Andrew Plotkin

Minimal, but what's there is acceptable
Very good
Dialogue and writing
Very good
Very hard, due to interface problems
Forgiveness rating
You can die. In a couple of places, an action which ought to be completely innocent turns out to be a fatal error, and it's not obvious why it's a fatal error.

Oh, dear. What a mixed blessing.

Here's a game been in development for a while. Well, they all are, I suppose -- a graphical adventure that ships only two years after inception is in a heck of a hurry. The Crystal Key, by its documentation, has been growing for five.

That tells, I think. The game feels... a bit old. I remember what games were like in 1995, 1996. Lessons yet to be learned. Okay, lessons learned, but yet to be pounded into the heads of commercial game designers. This game makes some classic mistakes. Then again, it gets some things classically right.

But I get ahead of myself.

A radio telescope intercepts an alien message, from a besieged world to its colonies. Bare days later, the aggressor (Ozgar to you) appears in Earth's skies. Global cataclysm! Doom, doom, alas and crap. You, the intrepid hero, are stuffed into an experimental hyperdrive ship and sent off for help.

Well. And so the game itself, which (predictably) involves exploring the ruins of the civilization which this Ozgar turkey already attacked. They drove him back once, so it seems. How you don't know, but they have at least one technology shared by none else: a portal that steps from world to world. ("This is not your father's hyperdrive.")

Good stuff, yes, but the story really doesn't go much beyond that. You hear more bits of story about Ozgar's assault, and the fall of Arkonia -- but the game itself is entirely a matter of "explore more; get to more places; get more tools." Eventually you have enough tools in enough places to win. So you do that.

Plenty is involved in doing all of that. The game has lots of subgoals, subsubgoals, and clever ways to achieve them -- puzzles, one might say. And very well integrated. Some rough spots at the beginning, perhaps. (Look, the aliens have locked their portal chamber with a combination lock made of soup cans on a shelf! Okay, I'm sorry, they're books. They sure look like soup cans--) But through most of the game, your actions make good sense in terms of the worlds you explore. You use machines that can reasonably be expected to do what they do, and be where you find them. Locks protect places and objects that you'd expect to be protected. Not too much is excessively contrived. (The contents of the boss's desk, that was a stretch. But not too much else.)

And the game is surprisingly dense. ...What a confusing word; I shall explain. The plot moves back over the same territory. You go through a location once; and then again, later, for different reasons. An inventory item is used for two different things, at different times. You are snatched from one world, and then return via entirely different means.

That's unusual, if you think about it. So many games invoke the trope of "multiple worlds to explore" -- but most use it to simplify the game design. Enter a world, deal with it, move to the next. No going back -- or if you can, it's because they're all independent, and no element of one affects another. The plot is a set of self-contained scenes, parallel or serial.

The Crystal Key does the opposite. The story may physically span several planets, but the plot is a single interwoven map, which you gradually gain access to more and more of. (It is significant that though each Arkonian world has a portal, all the portals are the same. Each leads to any place you have acquired access to.) The last exploration game whose layout was as carefully tangled was Riven -- which puts this game in high-noble company.

(Footnote: I say "exploration game", to include such multi-world games as Riven, Morpheus, and Timelapse. I'm not comparing this plot to event-driven games like The Last Express or Titanic. Those games have very complex plots, but not in the same way; they are sequences of events, branching over time. The Crystal Key is a sequence of essentially static places which you explore, which nonetheless form a tangled, unlinear game. That's what brought on the comparison to Riven.)

(Footnote deux: I just realized that I typed "Plot: minimal" up in the header. I don't mean to contradict myself. By "Plot:" I meant the underlying story, which as I said is mostly "Here's some background, now explore." When I then talked about the tangled, complex plot, I meant the design of what happens in the game. I should, perhaps, have been saying "Gameplay". I should, perhaps, pick new terminology. --I should, perhaps, get on with the review.)

That, all, is hindsight. I'll tell you my first impression of the game (after soup cans): I got in a dune buggy and died.

Worse than that, really. I got through the first few scenes, solved the two most contrived puzzles in the game, and found a vehicle. I boarded the vehicle. It drove itself to another location. Well. I tried to get out. It drove back to the first location -- and then, still under the game's control, I got out of the vehicle, was captured by an evil guard, was told to switch CDs, was thrown in a cell, and then got the message "Game Over".

Remember, the last action I had tried to perform was "get out of the vehicle".

I nearly gave up on the whole game right there. "Open the door and die" scenes were dropped like a hot pincushion in the Eighties. And when the game forces you to stick in a new CD for the privilege of seeing your death scene, it's time to get out the pitchforks and torches and go up the mountain stalking the designer, who will most assuredly be played by Boris Karloff.

Fortunately for Boris, I had no other graphical adventure games in the house, so I had to keep playing this one. The truth, as it turned out, is that being thrown in jail is only "Game Over" if you lack the tool needed to escape. If I had succeeded in leaving the vehicle, I would -- probably -- have found that tool.

A small problem, you might think. A trivial bug: the hotspots for getting out of the vehicle were not marked by cursor changes (unlike every other hotspot in the game). And a small design oversight: the return vehicle ride automatically triggered the jail-capture, whether you had the tool or not.

Fixing the cursor bug would be idiot-simple. And rigging the plot, so that the soldier only appears if you're carrying the tool, is just as easy. It's a plot contrivance, yes; it makes no sense in the world's terms. But it's the sort of contrivance that a player never notices -- at least in a game like this, where there's no "drop" action. It makes the story work; the plot advances when you've explored enough (i.e., found the right thing). And it would have prevented a horrible play-session train-wreck, which, please believe me, brought the CDs within inches of a sparkly microwave demise.

Let me complain more about the interface. A standard panning display. The cursor indicates movement hotspots (bar that one bug). The cursor also indicates action hotspots. Some action spots are "click to make it happen" -- examine, pick up, push, pull, whatever. Fine. Other action spots do nothing when you click on them. These, you are told, are places to use objects. You first select an inventory item, then click the hotspot; if the item can be used on the spot, it will be.

Fine again; an admirably simple design. Problem is, the game is sometimes too complex for it. A few things have to be dragged around the screen, even though dragging while clicking is usually the "pan" command. The cursor change that indicates this is sufficiently unobvious that I didn't figure it out until after I'd finished the game.

And the final puzzle, while a brilliantly clever piece of game design, just absolutely doesn't fit into the interface. I'm sorry. I knew what I wanted to do, but I stumbled through doing it by pure trial and error. Hotspots appear and disappear at random. Navigation stops working. You have to use an inventory item on a room feature, but a closeup view of that feature turns out to be the one view you can't use the item on. A wall panel stops working and then starts again. One action drops an inventory item (did I mean to do that?) and then another action must be done on that dropped object (apparently because the designers couldn't think of any hotspot that made sense. I couldn't either.)

In fact, none of that sequence really made sense. I clicked everything on everything else, and hoped that the protagonist would carry out the Clever Plan that the designers had come up with. Great plan; lousy implementation.

(If it makes you feel better, I know exactly how I'd implement it in a text adventure. A text adventure has a "drop" command. You can drop things whenever you feel like it, and if it makes sense to position them usefully, the game can cue you by telling you you've done it. But, of course, no pre-rendered graphical game can afford to have a "drop" command.) (And note that "drop" would have messed up my proposed solution for the jail-cell problem... complications, complications.)

I also had great trouble with the game's sense of focus. To get good focus, you must design a game so that every important part of a room is obviously important, and the background detail (however painstakingly attractive) doesn't waste the player's time. In The Crystal Key, that failed pretty miserably, I'm afraid. I wound up panning around every room, waving the cursor in a fine grid over every steradian. It was the only way to find every hotspot. Pixel-hunting isn't realistic, it isn't good pacing, and it isn't fun. It didn't even work. A couple of tiny, hidden-in-the-margin objects steadfastly resisted discovery. I had to literally search through the CD's script files, find the hotspot definitions for certain actions, deduce which rooms those spots were defined in, and then go back and search those rooms in even more minute detail. Cheating was mandatory. I could not have searched every room in the game with the attention necessary to find those objects.

Focus can be large-scale or small. A detail in the room can be drawn to draw attention -- or there can be something about the room itself which draws attention, leading the player to search it. One problem with a back-and-forth, go-everywhere game like this is that a missing object can be anywhere in the game. Searching it all, as I said, is a daunting prospect.

This is a hard problem. Contemplating a puzzle should point you at its solution, but when you're staring at a safe, how do you know where to search for the tool that breaks in? Once you have the tool, its use may be brilliantly logical -- but that does no good at all if you've never seen it. It could be lying in any corridor you've passed. (That, therefore, is an archetypical case where the detail must draw attention to itself. If the tool is hidden in its random corridor, hard to see, the player may get stuck without any idea how to proceed. Like I did. That's one place where I searched the game scripts.)

Contrariwise, if you have the relevant objects, the simple interface lets you try every inventory item on every hotspot, fairly quickly. That lets you brute-force a few of the puzzles. But not too many; more often it cues you that you're on the right track, but still leaves the meat of the puzzle to be solved.

Okay, that's deeper into game design theory than even I usually get. I'll hurry forward. Let's see, I can leap lightly over the navigation problems. (As usual for this game, the worst examples were at the beginning -- four viewpoint-positions in a square room that were not all connected to each other. C'mon, suck it up and render those extra transitions.) (And then there was the lost-in-the-jungle scene, where I had to play hunt-the-pixel just to count exits from each room. Focus problem, again.) (And I should warn you about the final puzzle area -- the Clever Plan depended critically on its geometry, but that geometry was so unclear that I had to take on faith it would work. I didn't understand the details until it was all over; both my attempts at mapping that area were dead wrong.)

No, no, I said leap over that...

I can happily recommend the artwork. Yes, it's a bit simple, compared to some of the insanely-modelled extravaganzas we've seen. But the artists get their point across. Lighting, sound, atmospheric effects; the overall sense of place is excellent. My biggest complaint about the first alien world was that it looked too realistically an Earthlike, grassy hill. Despite the overprocessed anamorphic display effect. Not shabby at all.

The authors fell a bit into the failures of imagination that I complained about in Lightbringer. Once again, the boring old human color wheel. The language-translation disc you find is, once again once again, called a "decoding disc". And it translates into English, even though it was made by one alien civilization to "decode" another. The two civilizations share the same written numbering system, for that matter. The authors seem to have just invented one and then said, "If they're not from Earth, they use it." ("A completely parochial attitude," Spock commented disapprovingly.) Worse, there's a single navigation-coordinate system used by all three civilizations, including Earth.

The ending of the game is pretty darn anticlimactic.


Conclusion: Some very good material, obscured by some very annoying problems. Have hints or a walkthrough handy when you play. The Crystal Key took me only about two days; it's smallish for a modern adventure game. Then again, the list price is $20, which is half what most adventure games cost. Consider it a worthwhile experimental short story, rather than a disappointing mass-market novel.

Macintoshness: Awful. No menu bar. No way to prevent the game from switching resolutions (and 640x480 mode looks terrible on some machines I've used). To get to the save/restore screen, you use ill-chosen and ill-implemented keystrokes. The save system doesn't use Mac file dialogs, substituting its own storage system, which is limited to ten saved games. That's just annoying. (I know some of you are reading this on the PC games forum. I hope that bypassing the Windows file dialogs, and being limited to ten saved games, is equally annoying to you.)

System requirements: Box says MacOS 7.5, 120MHz PPC, 32 meg RAM. (For you PC people, that's Win95/98, Pentium 133, and 32 meg RAM.) It claims to want 70 megs of hard drive space, but I found it actually needed 115 megs. And it also says "8x CD-ROM drive", which isn't an exaggeration at all. My middle-aged PowerMac 9500 was fine for CPU horsepower, but CD loading delays made navigation intolerable. Switching to the kick-ass laptop with the 24x drive was a vast relief.

Availability: Dreamcatcher's on-line store. It's a hybrid Mac/PC package.

Bugginess: I managed to corrupt the save-game file once. (All ten slots.) I don't know how, or whether it was a Mac-specific problem. Fortunately I hadn't gotten very far; I wiped the file ("") and made a clean copy (from ""), and started over. Thereafter, I followed the following superstitions: Always label saved games with alphanumeric names, no spaces or symbols, no more than eight characters. Always save by hitting "x", not "s". Quit and back up the save file every few hours, just in case. I don't know whether any of that really helped; but I had no more problems.

Game Reviews