Caffeination by Michael Loegering [Comp03]

IFDB page: Caffeination
Final placement: 14th place (of 30) in the 2003 Interactive Fiction Competition

Caffeination (it can’t seem to decide whether it wants the middle “n” capitalized, so I’m going to forego it) starts out like a more intelligent version of last year’s Coffee Quest II. The setup is nearly identical: you’re a depressed, sleepy office worker in an oppressive, cubicle-filled corporate environment, and your object in life is to get a really good cup of coffee. Happily, the literacy level of Caffeination is several steps above that of CQ2, and at first the game seems like it’s going to be a lot of fun. Alas, it quickly sinks into a bog of misguided design, buggy code, and error-laden prose.

In addition, rather than presenting any kind of sharp or thoughtful satire of office culture, the game instead tends towards the hackneyed and the juvenile. For instance, the boss’ name is “Mr. Norom.” Ooh, “moron” spelled backwards! There’s wit for you. Norom is a walking (well, sitting) “evil boss” stereotype, and the clich├ęs continue from there: the bimbo coworker hired strictly for the boss to leer at, the depressing and decrepit building fixtures, even the sleepy, slackerish PC himself. Instead of giving us a fresh, fully-imagined setting, Caffeination just shows us the same dull office we’ve seen a hundred times before. What’s fun about that?

This game also suffers from the same syndrome that seems to plague many entries in this year’s comp: a lack of sufficient feedback and cueing. I struggled against the constraints of the first major puzzle for quite some time before turning to the hints, but these were much less helpful than I thought they’d be. The game has a hint system with great potential — you find a notepad left by a former co-worker who had allegedly amassed all sorts of interesting tidbits about the office. Unfortunately, CONSULTing the notebook about nearly every topic gives you either no information at all (“Bill left some detailed notes, but you cannot find any info on that.”) or no more information than is present in simple object and location descriptions. A hint system in a consultable object is a great idea for integrating metagame activity into ingame mechanics, but to succeed, it must be much more deeply implemented than this.

So finally, I turned to the walkthrough and was astonished to discover that there are no fewer than three different solutions to it. The problem is that all three solutions rely on extremely improbable actions, ones I’d certainly never have thought would work, given the fairly limited implementation of most game objects. For instance, one path involves finding out about a particular bit of office intrigue through dialogue, but even very direct questions about this exact topic elicit no response whatsoever from the NPCs. In fact, most questions to the NPCs elicit the default response, which leads one to stop asking questions in fairly short order. Another path requires discovering an object hidden in a cubicle. However, an object mentioned in the room description that should be in roughly the same spot as the hidden object not only doesn’t lead the player to discovery, it isn’t even implemented, which certainly leads one to believe that searching that area of the cubicle won’t be fruitful. It’s all well and good to provide interesting and unusual solutions, but you can’t expect players to read your mind to get to them. You have to provide cues, feedback, and evidence that will lead the player in the right direction.

In a similar vein, if a game’s coding is focused on its solution path(s) rather than on making a fully interactive environment, it will almost certainly be extremely buggy to anybody who isn’t strictly following the walkthrough. I associate this problem most strongly with Robb Sherwin‘s earlier comp games — if the game was a story, it’d go pretty well, but IF isn’t a story so much as a place, and when an incomplete place tries to be a story, problems ensue — even though Caffeination provides multiple solutions to each of its puzzles, they’re all pretty hard to guess, and exploratory moves towards them founder in a morass of bugs.

This may be a problem that deserves its own category: the “walkthrough-driven game”. These games end up making me feel like one of the time travelers in Ray Bradbury’s “A Sound of Thunder,” terrified to step off the path lest everything around me be screwed up forever. That’s certainly what happens with Caffeination. Especially after the first puzzle, I found myself confronted with one bug after another when I tried things that the game didn’t expect.

Now, like the bugs in Sophie’s Adventure, many of this game’s bugs were beneficial to me, including one that allowed me to win without solving any of the coffee shop puzzles at all. I’ve been surprised by the games in this comp that go to great lengths to explain the obstacles to you, but then don’t bother to actually use those obstacles to prevent winning actions. Still, winning lacks its usual pleasure when it’s done by exploiting a bug. I was happy enough to have the game overwith, but I wish it could have been different.

Rating: 5.5

Sardoria by Anssi Raisanen [Comp03]

IFDB page: Sardoria
Final placement: 13th place (of 30) in the 2003 Interactive Fiction Competition

Okay, I know it’s bad form to start out a game review with a screed about the system it’s written in, but Alan is on my last nerve. First of all, I think I’ve mentioned before that I really dislike its lack of scripting capability. I keep having to remember to call up the scrollback and copy my latest moves into a text editor, and this is irritating in two big ways. First, having to constantly keep an eye on this really breaks immersion for me — I can’t get fully into the game when I’m having to manually monitor this administrative task. Second, if I do happen to become immersed, then I invariably end up forgetting to bring up the scrollback until too late, when the buffer only has half the moves I’ve made since my last copy-paste operation. Bad enough to have to copy from the buffer at all, but the fact that it only seems to hold a minimal amount of the scrollback makes the whole thing extremely aggravating.

The other problem that I just discovered with Alan is the apparent lack of portability in its save files. The way I play comp games often involves a lot of switching machines — laptop on the bus in the morning, work machine at lunch hour, home machine at night. Usually, this is no problem; I just drop my saved game onto a floppy and start up from it at the next machine. However, I discovered that whenever I try to restore a saved Alan game from another machine I get this infuriating message: “Sorry, the save file did not contain a save for this adventure.” It’s very, very annoying to restore “sardoria.sav” and be told that it’s not a save for Sardoria. I don’t know whether these flaws are in the language or the interpreter, but they just drove me crazy with this game, maybe because it’s a little larger or more ambitious than most other Alan games. (Certainly more so than The Adventures Of The President Of The United States, the other Alan game in Comp03.)

Whew, thanks for letting me get that off my chest. On to Sardoria. Well, I’m sorry to say that Sardoria also drove me crazy in its own ways, beginning with the very first room. It’s one of those games that starts with the PC imprisoned — thrown into a locked cellar, in this case — so the first puzzle is getting free. I must have spent 20 minutes trying out different commands in that room, none of which seemed to do anything for me. Finally, I turned to the hints and was told that I “need to find another object hidden in the cellar.” Well thanks, but that’s what I’ve been trying to do for the last 20 minutes, without success. Any more hints? Nope, “[t]here are no more hints for this location.”

Finally, I turned to the walkthrough, and found that the command the game was looking for was something that I would never have thought of doing, because it doesn’t really make much sense. The result is that as I’m doing the nonsensical action, I happen across the hidden item. “No fair!” I cry. It was especially galling because I had specifically checked the area where the item was hidden, and was told in clear terms that there wasn’t anything there. This sort of frustration happened to me over and over in the game; I tried on all of them, but I think I ended up looking at the walkthrough for every puzzle but one or two. Often I’d come close to the solution but failed to find the exact route desired by the game (shades of Gourmet), and sometimes the solutions just flat-out made no sense to me. There were also occasions when the actions required made a sort of sense, but were far too vague or arbitrary.

Many of these problems could have been overcome, or at least alleviated, if Sardoria had provided better feedback. More and more, I’m convinced that this is a crucial element of successful interactive fiction, at least IF with puzzles in it. When a player gets close to the solution, the game should indicate that rather than giving a flat “nothing happens” sort of response until it gets the exact right set of commands. Moreover, if players think of an alternative solution, the game should be able to either let them utilize that solution or provide a convincing reason why they can’t. How can an author provide this level of feedback? It’s all about the testing. Get at least three testers for your game, with a sufficient variety of approaches between them. Then, watch for the things they try. If they get close to the answer, your game should provide some appropriately encouraging feedback.

This is especially important in a game like Sardoria, where many of the puzzles are one kind of combination lock or another, most of whose combinations verge on the totally arbitrary. This is a subject that deserves a more detailed treatment, but I’m unable to do that in a spoiler-free review, so all I can say is that designers must anticipate the majority of player responses and handle them appropriately. It’s a lot of work, yeah, but it can be the difference between exciting and exasperating for puzzly IF.

Rating: 5.3

[Postscript from 2021: In a subsequent discussion on, someone pointed out that in fact, it is possible to create a game transcript in Alan if you start the interpreter with a “-L” switch at the command prompt. Fair enough, but I stand by my response from that discussion:

“You know, somebody probably told me that last year, too. Unfortunately, I immediately forgot because this is such an amazingly clunky and unintuitive way to provide a game transcript. I want to type SCRIPT and have the game start logging a transcript to a file. I do not want to have to open up a DOS window and start the game with a special switch from the command prompt, especially since I will almost inevitably have the interpreter and the gamefile in different directories (something the Arun interpreter barely knows how to handle as it is) and will thus be forced to type out an entire path for one or the other of them.

Anyway, I amend my complaint from ‘Alan provides no scripting capability’ to ‘Alan’s scripting capability is far too hidden and inconvenient.'”]

Temple of Kaos by Peter Gambles [Comp03]

IFDB page: Temple of Kaos
Final placement: 15th place (of 30) in the 2003 Interactive Fiction Competition

Much to my dismay, Temple Of Kaos has nothing to do with Maxwell Smart or Agent 99. I mean, sure, getting permission from whoever owns the rights to the show would probably be an insurmountable hurdle, but legal problems aside, wouldn’t Get Smart make a fantastic IF milieu? The Shoe Phone, the Cone Of Silence… I can just picture it. This Temple of KAOS, however, isn’t the stronghold of a nefarious organization of evil spies whose efforts must be thwarted by the agents of CONTROL, but rather a bizarre otherworld, where rules of time, space, and spelling (it’s called KAOS for no discernible reason) don’t apply.

Basically, the two salient features of this game are its disregard for making sense, and its propensity to express itself in verse, sometimes free, sometimes blank, and usually with end rhymes. As such, it’s a highly experimental game, and while I think both of the experiments fail in this instance, they illuminate some remarkable territory along the way.

Let’s tackle the nonsensical part first. In the ABOUT text, the author states that part of his intention with the game is to “interfere, out of sheer mischief, with some of the normal perceptions / causal relationships of IF space-time.” Mission accomplished, and in some parts of the game, the technique works well. The first section in particular contains a puzzle which utterly confounds standard expectations of how the world ought to work, but it’s possible to figure out the alternate system of reality at work in the puzzle, and thereby defeat it. The process of doing so is really fun, reminiscent of the flavor of The Gostak or For A Change. More specifically, the reason the puzzle works is that even though the PC’s actions don’t produce the expected results, they do produce some results, and from these results it’s possible to deduce what’s really going on. The same can’t be said of most of the other puzzles in the game.

Even for IF set in a much more mundane universe, feedback design is one of the toughest parts of puzzle creation — you don’t want to be so obvious that the puzzle becomes a non-puzzle, but your feedback also mustn’t be so obscure (or nonexistent) as to leave the player shaking her head in confusion even when the solution is revealed. Most of the puzzles in TOK err on the latter side of this line. I think that for every puzzle after the first one, I looked at the hints, and for most of them even the hints were insufficient. (Thankfully, the author provided a walkthrough.)

For some puzzles, the solution made a tortured kind of sense once I’d looked it up, but for many, I found myself just following the walkthrough’s instructions with a shrug. Sufficient feedback is very important in any IF puzzle, but in a world where the normal rules don’t apply, feedback becomes utterly crucial — how are we supposed to figure out the rules without the ability to gather any evidence about them? TOK usually (though not always) fails to provide enough feedback to make its puzzles solvable, which takes a lot of the fun out of playing.

What does provide some fun is the game’s tendency to present its room and object descriptions in a shaky kind of poetry. For instance, the first room description:

In the North Chamber

Chamber of the north, so empty, still, all noise grates
Black as night the chest your thought awaits.
The other chamber southward lies
Cloaked in mystery's disguise.

Most, but not all, of the game’s verse rhymes like this — sometimes the lines lack rhyme or even consistent meter. Moreover, there’s a fair bit of prose mixed in, as conversation, library responses, or descriptions of action, and the presence of these rather ordinary bits of writing juxtaposed with the more elevated verses tends to drain the effectiveness of each. The other problem with the poetry is reminiscent of what happened in Graham Nelson‘s final game, The Tempest. That is, it’s tough enough to craft IF prose that communicates clearly and concisely, and that also provides enough information to the player, but to do so in verse is much, much harder. TOK‘s poetry isn’t at as great a disadvantage as The Tempest, which forced itself to use prewritten lines as room and object descriptions, but it can still be rather opaque.

Usually the lines aren’t pure gibberish, and they sometimes even manage to pack a few clues in, but nevertheless it does take some time to translate, for instance, “Black as night the chest your thought awaits” into “There’s a black chest here.” The poetry technique is ill-chosen in combination with the game’s nonsensical laws of time, space, and causality, since either one by itself is confusing enough but together they can be utterly impenetrable. However, TOK does give some glimpses of how compelling an IF game in pure verse could be, and of how fascinating it might be to play in a universe with a completely different set of basic rules. Play it for these glimpses, but don’t be afraid to reach for the walkthrough.

Rating: 6.5

Color and Number by Steven Kollmansberger [Comp02]

IFDB page: Color and Number
Final placement: 24th place (of 38) in the 2002 Interactive Fiction Competition

Color And Number belongs to that genre of IF I’ve begun to call “pure puzzle games” — oh sure, it’s got a shred of plot, something about investigating a cult that worships colors or something, but that’s more or less overwith before the first move, and from that point forward, you’re pretty much in a pure puzzle landscape. And yes, those puzzles are keyed to a particular theme — you guessed it: colors and numbers. True to the precedent established in Comp01 games like Elements and Colours, the game even names itself after its puzzle theme.

About twenty minutes into this thing, I knew I didn’t have a prayer of finishing it in two hours, so I played until I hit the time limit and then stopped. Thus, in fairness, I don’t know whether the story makes a strong resurgence towards the end or anything, but even if it does, this game clearly belongs to the puzzles. Those puzzles are of the sort that prompts lots of note-taking, charting the correspondences between the various pieces the game teasingly doles out. I enjoyed several of these brain-twisters — they have a mathematical elegance, and some of the best ones suggest their solutions quite organically, which is a pleasure.

Others, though, are a little more imperfect. One puzzle in particular stumped me even though I had looked at all the hints for it, and I think there are several reasons for this. First, the feedback level was too low. The puzzle involved performing a string of actions, but without close investigation, the environment betrayed no particular indication about which actions were successful or useful. It’s not that this feedback was entirely absent, but it wasn’t prominent enough for me to even notice until long after I had looked at the answers.

Secondly, the sequence has a bug in it. It’s just a TADS error (one which oddly didn’t show up in my game transcripts, so I can’t quote it) — not enough to prevent the solution from working properly, but more than enough to drain my confidence in the puzzle’s correct implementation. Between that and the lack of feedback, it’s pretty clear how I ended up looking at hints, but even after I had seen them all, and ostensibly solved the puzzle, nothing happened.

I found out, through trawling Google for hint requests, that this was because I needed to do some other actions in an entirely unrelated area. This is not good puzzle design — at the very least, solving that portion should have yielded some noticeable change so that I could understand that my attempt had in fact worked, even if it wasn’t producing any useful revelations until its counterpart pieces were in place.

Critics like me talk a lot about how difficult it is to pull off combining an arresting story with interesting puzzles, but what’s becoming clearer is that even when IF eschews story altogether and focuses solely on puzzles, it presents considerable challenges to its creator. Little prose errors and formatting issues aren’t so noticeable in a work like this (unless they severely cloud meaning), but even tiny feedback or implementation errors can be devastating. Because there’s no story to distract us from game bugs, they loom very large indeed, and as soon as one crops up, it drastically affects the dynamic between player and game. Suddenly, a struggling player ceases to believe that he’s stuck because of his own inability to solve the puzzle, and starts to suspect that game defects are making the puzzle unsolvable, because after all, if bugs crop up in one place, they can be elsewhere too.

Infocom and its contemporaries had a big advantage in this area — if you bought a game off the shelf, knowing that the resources of a full-fledged company had been used to quality test it and that it had been reviewed by major publications, you could be relatively confident that whatever bugs still might lurk within it wouldn’t be enough to prevent you from solving its puzzles. No such assurances exist for an amateur, freeware IF comp game, and consequently pure puzzle games must be fanatically assiduous about debugging and testing. That’s not an easy mark to hit.

Rating: 6.7