Review: Lighthouse

From Sierra On-Line

Review written by Andrew Plotkin

Dialog and writing
Mostly very good, except for some contrived timing
Fairly hard
Forgiveness rating
You usually get several tries, and several alternate solutions, but you can get stuck if you try hard; you can die in exactly one place.

Lighthouse was awful in some very enlightening ways.

Your weird neighbor, the Mad Scientist, leaves you a message warning you of great danger. Or something. So off you go, and what do you know? His lighthouse is full of Mad Science! And what's that? A gateway to another world! And an evil being who steals the MS's baby daughter!

(I must admit, the baby daughter is a new twist. They get bonus points for not kidnapping your lover, father, or attractive young cop partner.)

<click> Setup complete. But what's interesting is that, almost immediately, the plot diverges from the norm of graphical adventure games. To be specific, it does diverge. You can leap through the after the evil being, or stay and explore the lighthouse. The plot works through slightly differently each way, but it does work through. If you fail to get your hands on a particular item, you can encounter it in two different places later on. Events can occur in different orders; you can learn things in different orders. There are several ways to get around, which is how a lot of this works; you can get back to a place if you missed something there. Sometimes the around-getting ways are one-way, but that's ok, there are several of them.

Multiple endings, as well, lest I forget. There are at least two satisfying solutions to the overall problem of the evil being. (I only found one -- the other was hinted at by the hints, about which more later.) There are some number of less complete victories, each of which is a consistent outcome of a particular set of actions. The documentation says that there are sixteen possible endings. Dunno about that, but the plotline is considerably more flexible than the last game you played.

And this is impressive stuff for a graphical game. In fact it highlights one of the big truths of graphical games as opposed to text games, which is that alternate paths in graphical games are expensive. Every scene, every variation of a scene, is a tremendous amount of graphics work. A scene that the player doesn't have to go through is -- to certain accounts -- a waste of effort. But this is exactly what Lighthouse has; scenes that the player doesn't have to deal with, variations that he may never see, even puzzles he doesn't have to solve -- because he took a different path through the game.

This is very gratifying. I can't think of any other graphical game that puts this much work into alternate plotlines and possibilities. (One might count Bad Day on the Midway, but that doesn't have a fixed plot at all, and is more like alternate possible events than alternate storylines.)

Unfortunately and in the midst of all that, I didn't like Lighthouse. So what went wrong?

First problem: good though their game design was, it's more of a good try than a success. Some pieces fit into place very badly. In particular, other characters are a real problem. They show up when they feel like it. "When they feel like it" often is meaningful in terms of the plot -- but the player doesn't know that. If you walk into an empty room, there's no way to tell that if you've learned what you need and if you stand in the right spot and if the appropriate witness is present, an enemy will appear who you can defeat. There are irritatingly many of these little plot contrivances. I don't honestly see how you can get them all right without resorting to the hints (about which more later.)

Then there is the interface.

The navigation hotspots are just terribly placed. A small rectangle in the center of a staircase, instead of the entire staircase. Strange borders for left and right turn spots. Strange ways to get to what should be an accessible spot. Over and over again, an arrangement where "down" is in the center of the screen, and the bottom edge is "turn around". It's a tremendous headache pretty much all the way through the game. In one room I literally didn't realize there were two doors, even though I turned around several times; no possible position faced the second door, except for one which was so close to it that I couldn't tell it wasn't the first door.

Plus, the cursor-changing and highlighting is buggy. The highlighting is optional, by the way -- why they thought it was good to leave out "push here" cursors in a game full of dials, levers, and switches, half of which are purely decorative, is entirely beyond my comprehension. Turn highlighting on as soon as you start the game. But even when it's on, the cursor often gets stuck in a particular shape or highlight level, until you yank it around the window a couple of times. Try searching for hotspots when that bug hits. It's the tiny little problems like this that can really wreck a gaming experience.

Plus, the "push here" highlight is a barely perceptible lightening of the grey arrow. Yeesh.

But more fundamentally, there's an interface problem which took me a while to figure out. In the end I decided that it was the curse of graphical games trying to go where Infocom went before, only without the clue.

Here's the deal. You encounter an object. You pick it up, and it enters your inventory. You click on it to use it. For some objects, you get an object-shaped cursor, indicating that you can apply it somewhere on the screen. For others, you get a pop-up window showing the object, which you can click on more precisely.

Sounds good, right? A key should be applied to a keyhole; a machine should be popped up so that you can manipulate its dials, switches, and buttons. But the problem is, this conveys information. And the authors didn't always think about what information it conveys.

For example, I found one object which belonged to the first category, even though it looked mechanical. Ok, I thought, this is a plug-in unit and I can ignore it until I find the machine it needs to be plugged in to. Well, not quite. Turns out that it was a broken machine. I had to fix it, which meant that I had to "apply" it to a particular empty workbench (and then fiddle further.) That, my friends, is the classic problem which cannot be solved without telepathy. How can I know it's broken? And the problem comes from the "optimization" of only allowing valid commands (in this case, no pop-up window for a non-functional machine.) Which comes from the limitation of a one-click interface. In a text game, all commands are (by definition) available to be typed; I could have typed "push button on machine" and gotten a "nothing happens" response. Or a "tired buzz", or a burst of sparks, or something. Here, I was never even given the option. Or for fixing the damn thing -- "open machine", or "unscrew panel", or something of the sort -- maybe the work could only be done in a particular workshop, but there would be some kind of failure message clueing me in that success was possible. In this game, the option of fixing the machine is only visible in one place (the workbench), and then it's not obvious until you stumble across it (by putting the machine there.)

Or the context-sensitive toolbox. I opened a box, found two tools, used one immediately and kept the other. Everything else in there was untakable; I assumed it was decoration. Later on I wanted to do something. Couldn't make it work. Had to read the hints (about which more later) to realize that I had to open the toolbox again and take different tools, which were now valid hotspots. Argh.

Another example, more closely related than it might seem. I saw something on the ceiling. No hotspot, so I figured it was out of reach. I tried poking at it with a long item I was carrying; no good. Ok, I said, ignore and come back later. Again, not quite. The hint for this one said "Aha, isn't that an antenna sticking out?" Oh, it is? This is a game of Mad Scientists; everything is covered with knobs, poles, valves, pipes, wires, tubules, and suchlike excrescences. In a text game -- pardon me for harping, but this is really the point -- in a text game, there is a well-understood semantic structure for pointing out significant items. You say "look", and it says "There's a [...] on the ceiling", and you say "examine [...]", and it says "...there's an antenna on it." If it's important the author will mention it, and if the author mentions it, he at least wants you to ponder whether it's important.

(I'm not saying this is impossible to reproduce in graphical games. It can be done, largely with careful use of "examine" zoom-ins and cursor highlighting. But in Lighthouse it doesn't work.)

Enough about that; let me talk about the hints. As a result, I assume, of the problems I encountered, the current version of Lighthouse (2.0) has extensive on-line help. Well, this is good; I very quickly wound up going to the hints for nearly every problem, because it was easier than telepathy and faster than thorough experimentation. (That's another rant about graphical games, the speed problem, which I won't even get into here.) This enabled me to mostly finish the game myself. I had to go to the Net for one hint, for a puzzle-box, which I can't call unfair because puzzle-boxes are supposed to be obscure, but I do sort of wish they'd put a less obscure puzzle-box in this game. Slightly less obscure. I liked it up until the last lock. But anyway --

-- the hints were apparently written in great haste by someone without a clue. If it was the game author, gosh, I apologize -- but, um, what were you smoking? Many typoes ("sight" for "site", "its" for "it's", come on, proofreading is part of product testing and a spellchecker is not a proofreader). But much worse, the hints repeatedly give away the upcoming plot. Not just how to solve the puzzle, but why to solve the puzzle, and in fact what will happen next in the plot after you solve the puzzle. In a few cases this is helpful -- particularly the plot contrivances I was complaining about earlier. But it sure does take the tension out of the game. I knew pretty much everything that was going to happen in Lighthouse long, long before it happened; and I don't think there's any way to avoid this problem. Which means that, however nicely assembled the plot is, it's a lousy story. Why care when you know what's going to happen?

I seem to have talked about the hints without talking about the puzzles. The puzzles, once you get past the techical complaints, are pretty good. It's all machines, of course. Mad Machines. I was able to get into most of them; you play around, watching the results, until you understand the point of the machine. (There is a point; these are machines with a reason, not usually arbitrary puzzles.) Sometimes you screw up, but there are usually spare parts. Save anyway, and remember to try pulling up as well as down.

Once again this is getting long. Other random complaints (you know there's a problem, I'm afraid, when the review is a list of complaints):

The Macintoshness of this game is zero. No menu items, or rather there are some but they don't work. You have to click on a button to get to the save/load/quit/hint screen. Particularly annoying when you (i.e., I) want to check the hints very frequently. Also, there was some incredibly annoying problem which would cause the game to freeze up entirely for twenty seconds at a time, or forty, or a solid minute. It was some kind of disk access problem -- at least, the freeze always ended with a burst of CD activity. More pain. Saving and loading games was always slow, and the saved game files were immense (up to a megabyte each near the end of the game. A megabyte!?)

Due to the varying order of the plot, you sometimes have to do a lot of running around. And running around is slow. I thought of a new approach to winning just now; it could produce a pretty impressive ending. Am I going to try it? Hell, no. It would take at least half an hour to set up, all of which would be travelling via pathways I've been through a dozen times before. (Plus the occasional freeze-up.) I've already won; it's not worth the time. So that ending, however cool it may be, means nothing to me.

And there's exactly one spot in the game where you can die. It's near the end. Not a climactic scene, just a place where if you're not paying attention, oops, sudden death. Ha! Ha! Bleah.

So the conclusion is... It explores a lot of ground that I want other games to explore. There was some very good design work. But due to large problems and small, Lighthouse just isn't much fun. On the other hand, it's half the price of more recent games. And it'll take you a while; it took me several days. A lot of that was travel time, but a lot of it was exploring time. It's a pretty big world in there.

So is it worth it? Er, well, I'm not making any promises. Sorry.

Availability: You know, Sierra is advertising a joint Lighthouse/Rendezvous With Rama bundle for the Mac. But I called the phone number on their web page, and the guy I talked to denied the existence of any such thing. Whatever. In any case, make sure you get version 2.0 of Lighthouse -- it's a hybrid Mac/IBM format, so it should be easy to find -- and then get the 2.0.1 patch off of Sierra's web site. (There's a critical "out of handles" bug near the end which the patch fixes.)

System requirements: Powermac, 16 meg RAM, System 7.1 or later, lots of hard drive space (especially if you're saving games), 640x480 256-color display. I must say that they could have gotten it to work on 68040 machines if they'd really wanted. Other games have done 640x480 8-bit video. But it's the PPC age, I guess. On the bandwagon with ye.

Game Reviews