The Stack

One of my favorite things: a UI experiment! It’s a little like Texture and a little like Robin Johnson’s games. The screen is divided into two panels. The left panel shows the story text and the room illustrations. The right panel contains all the interactive elements: inventory, environmental objects (including characters), locations you can go to, even an inventory of “thoughts”, which you can use as conversation topics or just reflect on. All these things are represented as card-like icons, and all actions are performed by dragging one card onto another. It takes a while to get used to, especially for actions that feel like they should take only one card, like examining an object or moving to a location. These are performed by dragging the relevant card onto the card representing Luis, the player character.

Luis is a recent arrival at a small village in post-war Spain, where he’s started suffering from recurring nightmares and strange feelings of elation and nausea. Even though there’s an entire tabbed panel for inventory, I spent a good long time without any physical inventory items at all — most of the story is instead spent asking the locals about your Thoughts, investigating the village’s mysteries: suspicious deaths, possible witchcraft, darkness in stark contrast to the sunny illustrations. There’s one character that I’m pretty sure is a hallucination, a reflection of a repressed guilt that I never fully plumbed. I’ve seen an ending — a bad one, where you wind up in prison — but haven’t yet brought the story to a satisfactory conclusion that answers all the questions.

That’s because there’s a lot of guesswork involved. The physical puzzles are pretty obvious, but when it comes to conversation, the interface produces an element of combinatorial explosion, and only sometimes gives you the guidance you need. Sometimes it emphasizes particular cards by slowly pulsing them (a nicely subtle effect), but a card is only half of an action. Sometimes there are multiple approaches to advancing the story — for example, you can break into an abandoned house by force or find out who has the key and get him to trust you. That helps, but it doesn’t help all the time.

So, those problems might be inherent to the format, or they might be avoidable in a completely different game, with different content. There are also some minor noticeable problems that could have been fixed here without changing anything fundamental. It’s fairly easy to make the game logic inconsistent by doing things more than once, or by triggering both sides of a plot branch. The game seems to have been translated from Spanish, and while the translation is generally good, there are occasional bits of awkward or unnatural English, possibly the result of translating Spanish idioms too directly. The UI could definitely be improved. The tabbed interface highlights the selected tab with a pink nimbus, but doesn’t put it in front of the other tabs. Some of the cards have a dark background that makes the text hard to read. Most of the cards are taller than they are wide, while the text on them is wider than it is tall, meaning there’s a lot of wasted space and the text still gets cut off at the beginning and end sometimes. There are some flashy blur and ripple effects that seem out-of-place while the basics are this underdeveloped.

Still, this is a Worthy Experiment, and probably worth the effort of cleaning it up a bit. I also hope to see the experiment carried on further in other games.

It’s been a busy day, and I only have the energy for a few brief comments on a similarly brief game.

The Surprise is an autobiographical story of a woman self-administering a couple of pregnancy tests in an office bathroom, learning that she’s pregnant, and notifying her husband through joyous tears. That’s it. Just a little episode from the author’s life that meant so much to her that she wanted to share it with the Comp judges, to whom it will probably, in most cases, mean not quite as much. It’s made of Twine and minimalist, functional prose, with a little freely-roamable map of sub-locations.

There’s some use of real-time delays, which is something I’m going to complain about in every game where it occurs. The longest such delay here is used to reproduce the 20-second wait for one of the tests to display results. Can I just say that I do not find this to be an effective way to build the sense of anticipation that the author probably intended? I can’t speak for others, but my reaction to a deliberate 20-second pause in an interactive work is pretty much always going to be annoyance.

Anyway, that’s not what I found interesting about this game. What’s interesting about it is what it shows about the state of Twine. This is, as I said, a short and minimalist piece written in Twine. And yet, it’s implemented adventure-game-style, with rooms and inventory and objects that can exist in multiple states. Apparently this style is now easy enough, or perhaps just expected enough, to do in Twine that people will choose it over stateless branching text for tiny personal games like this one. This was not the case just a few years ago.

I know I said I’d be posting about just one Comp game a day, but today, the randomizer just happened to give me two thematically connected games one after another.

Winter Break at Hogwarts is obviously Harry Potter fanfic, but if Harry Potter himself is around, I never saw him. It seems to be set before the books, with the player as some other student stuck at school while all the other children are away. The result is an almost barren environment. The map is large — so much so that the author saw a need to enclose a multi-tiered map with the game. The map is as attractive as it is necessary, but it doesn’t change the fact that Hogwarts is sparse, composed mainly of corridors and rooms without interactive detail. A handful of locations house a familiar character from the novels, but they’re only minimally responsive to player actions. It all feels not so much like an interactive Hogwarts as a Hogwarts museum, with dioramas you can look at.

Ah, but what about the magic? After an hour of play, I still hadn’t cast any spells. You start the game with a notebook with various familiar-from-the-books magic words written in it, but your wand is missing, and you can’t cast spells without a wand. After extensive exploration, I managed to locate the wand, but it was out of reach, requiring more puzzle-solving — I’m guessing I could have gotten it by flying on a broomstick, but I failed to locate a broomstick. (The cleaning supplies closet was locked, and I didn’t have a wand to alohamora it open with.) And so I gave up. It’s entirely possible that the game really picks up once you have the wand, but I think that if a game locks the good stuff behind a dedication barrier, it’s fair to judge it on that basis.

Coming off that, I was immensely pleased at how quickly Remedial Witchcraft got going. It’s sort of a Sorcerer’s Apprentice variant: you’re a witch’s apprentice suffering through a series of mishaps that you fix with magic, which then causes more problems. The ending is pretty unsatisfying — it leaves you at the beginning of the biggest mishap yet, with no real indication that you’ll be able to get out of it at all. But other than that, it’s great fun. The map is small but dense, consisting solely of a ramshackle cottage and its garden. The magic you deal with is mostly in the form of enchanted objects rather than magic words, but that’s excellent adventure-game-puzzle fodder. In particular, it gets good mileage out of a stone that you can teleport to, wherever it is. I recall Zork Zero had something similar.

Mainly, though, Remedial Witchcraft has a lot of character. The protagonist is an adorable tyke who’s capable of working wonders but struggles to carry a crystal ball around. Her master, the Witch of the Howling Woods, is absent for most of the game (the better to leave the kid to her own devices), but makes an impression as a comically disorganized and overbearing foil, and the entire unkempt place reflects her personality. Her familiar, a cat, is wiser than you are, but still a cat. Certain objects become animated and have to be caught, and their actions are described like they’re small animals that got loose.

So, basically, this is the Hogwarts experience I wanted. I notice that the comments about it in the intfiction.org forum are mainly complaints about bugs, but twenty days into the Comp, they seem to have all been fixed.

In past Comps, particularly the first few I blogged, I made some complaint about the people who devote their effort to creating new parser-based IF authoring systems and ending up with worse results than if they had simply used the established systems like Inform and TADS. It’s rare to see people doing this any more. I’m not sure why. I’d blame the Twine Revolution, but it’s hard to imagine that really affecting the people who build their own engines for fun.

Island in the Storm is the fruit of someone building their own IF system, and it’s actually pretty okay! That is, I can make some complaints about the game’s content — it’s got too many filler rooms for my taste, and I thought one of the puzzles hinged on an unclear description. But the interaction is quite comfortable. Maybe it’s partly because minimal command sets are in right now, so I don’t expect the range of action that I once did. But this game provides a solid core of what I expect, from common abbreviations to pronouns to decent handling of command history, and seeing that is a relief when it’s not guaranteed. I suspect that part of the reason it works as well as it does is that the author chose to create a library for making IF in Python rather than inventing a domain-specific language. That’s got to be the saner approach nowadays. I don’t know if the library is at all usable by anyone other than its creator, but it hardly matters to the player.

A few notable peculiarities:

All the output text is contained in graphics-character rectangles, usually one per turn. These vary in width to fit the output, so that shorter responses have narrower enclosures. And I don’t just mean one-line responses; it varies how soon it wraps to the next line. I suspect this would look much weirder without the boxes.

If you use a verb with syntax that the game doesn’t recognize, like if you provide an object and it isn’t expecting one, the game suggests that you ask for “VERB HELP [VERB]”, which gives you a list of all the accepted ways to use that verb. This is a really good idea! It feels a lot like asking for a man page.

Conversation with NPCs takes the form of choosing suggested topics. You’ll TALK TO somebody, and the response will contain things like “(You could ask about the four altars)”. We’ve seen things like this before — Lost Pig is a memorable example. But the peculiar thing here is that you can pick a response be means of any of subset of its words, without any other verbiage. Just typing “ALTARS” at the command prompt would be enough to pick the topic I mentioned. I’m not convinced this is superior to a numbered menu, but it’s definitely different.

There’s a limited light source that slowly recharges while exposed to light. This is an interesting idea: it gives you a similar sense of urgency to a limited battery, but with lesser risk of permanently ruining your game. It essentially turns the caves into something like an underwater section in an action game, where you have to periodically resurface to survive. I’m not sure anything of value was added by making the recharging gradual, though. I wound up entering a lot of “Z”‘s when I was impatient to start my next sally.

As for the story: You’re shipwrecked on an island, and have to effect repairs to your boat in classic adventure-game mode. It strikes me that the desert island is one of the less-imitated classical adventure game environments. This island isn’t deserted, though — it has a town and even a shipyard, but no one leaves, and all but one of the inhabitants worship a goddess that controls the storm that shipwrecked you, and which exerts a sort of semi-mind-control over them. It’s a bit Doctor Who. After you make all but one of the necessary repairs to your boat, the second act begins: you can only get the final component you need by delving into the trap-and-puzzle-laden dungeon in the ruins on the mountain to kill the goddess. Progress in the dungeon still involves popping out to talk to the locals to get needed materials via newly-appeared dialogue options. This took me too long to figure out.

I didn’t get all the way there; I managed to crash to the desktop first, and didn’t feel like replaying the part since my last save. This was unexpected — IF is usually more resilient than that. But I guess that’s the risk you run when you write your game in a general-purpose programming language.

The Legendary Hero Has Failed is a last-moments-before-the-apocalypse piece, like Anna Anthropy’s Queers In Love at the End of the World and I think one of the Fingertips games in Apollo 18+20. I’m pretty sure it’s also Legend of Zelda fanfic. I’ve only played a couple of the Zelda games, but the end of the world here is a result of the moon crashing into the earth, and furthermore, it’s described as grimacing while it does so, and isn’t that something that happens in one of the Zelda games? Reinterpreting things from popular media and investing them with more meaning is of course a fine thing, but you can easily wind up with something that’s only of interest to people sufficiently familiar with the source material, and I think that might be happening here.

It’s basically about human reactions to impending doom. You’re sitting around a campfire with four friends, watching the world end, and have the option to talk to them, comfort them, rage at the moon, play Fizzball, get high… not a lot of options, really, all told. Maybe these are characters from the original game. Maybe not. All I can say is that I’m not really very attached to them even after talking to them in this game. I think I was inclined to be uncharitable toward them because the game decided to annoy me at the very beginning by delaying the appearance of the link that lets you start the game. There are a couple of other moments like that within the body of the game, and it’s even more irritating there, because the whole thing is on a real-time timer. The moon crashes into the earth after 5 minutes of play no matter what, and the in-game pauses waste precious seconds of that. It’s a pernicious combination.

And yet, inconsistent hobgoblin that I am, I also wound up feeling like the timer was too long! Compare it again to Queers: the timer there is a mere 10 seconds, because it’s trying to make you feel like there’s nowhere near enough time to do everything you want to do. Whereas five minutes is more than enough time to exhaust your options in Legendary Hero if you don’t dawdle. The approach of the grimacing moon becomes less of a constraint on your actions than the mere fact that there isn’t much to do. And I really don’t think that’s what the author was going for.

Turandot, by Victor Gijsbers, is a romance flavored with peril, only very slightly connected to the opera of the same name: it tosses the exoticizing orientalism out the window and expands the princess Turnadot’s test for her suitors from a mere three riddles to a full-on dungeon with pitfalls, traps, and a ravenous crocodile. The protagonist, Calaf, is a self-centered hedonistic playboy who sneaks into the princess’s garden on a dare, and, on seeing her, is immediately smitten with an irrational operatic love that persists even when he finds out that she’s the sort of sadist who would build such a dungeon in the first place. The bulk of the piece is spent in that dungeon, where Turandot essentially plays GLaDOS, following your progress from various balconies and observation posts and bantering with increasing familiarity as you go. Most of the interactivity here is dialogue; actions like dodging traps and wrestling the crocodile are peppered throughout, there being no particular reason to separate dialogue from action in a choice-based work.

There’s something very comfortable about the texture. By which I mean, the amount information you get before making your choices, and the amount of text between them, seems just right to me. This has definitely not been the case for every Choicescript work I’ve encountered, and when they err, they tend to err on the side of giving you too much to comfortably process at a time. One way Turnadot avoids this is by just throwing in fake, inconsequential choices once in a while to break things up into manageable pieces. In other games, I’ve sometimes found this annoying, smacking of dishonesty. But here, it’s generally clear what’s going to have real effects and what isn’t, and the author frequently uses these points as what Emily Short has called “reflective choices”, where you’re choosing the contents of your character rather than the progress of the plot. Figuring out what it means to be in love with someone who’s trying to kill you.

Because that’s really what this work is about: a character being forced into confrontation with himself. Two characters, really, including the deadly princess. The dungeon is a metaphor for the terrifying work of self-knowledge, and of being known by someone else. That goes for the princess as well — the whole thing is basically a huge, externalized defense mechanism to keep people from getting too close. At the same time, she has the power to make things easier or harder for her prisoners; ultimately, the real reason all her previous suitors failed the challenge is that they annoyed her. She’s almost a vigilante, really, avenging the sins of male privilege, of all the princes and potentates who dared to try to claim her, treating their desire as more important than her will to self-determination. Most of all, her experiences have made her detest powerlessness, and so her tests cull out both those who will try to take away her power, and those who will allow her to take away theirs.

Even at the story’s most serious moments, when Turandot and Calaf discuss their guiltier secrets, the tone is fairly playful. And this works into the choices as well — there’s some formal experimentation that reminds me of the old Lucasarts games. On several occasions, you’re given a set of choices with the sole sensible choice grayed out, clearly indicating that Calaf knows it’s an option but is determined not to do it. The ending indulges in this so hard, it feels like he’s trolling the player.

Sugarlawn, by Mike Spivey, is built on essentially the same conceit as Ryan Veeder’s 2013 game Captain Verdeterre’s Plunder: the whole thing is one big optimization puzzle, giving you an environment rich in treasures to collect, and not enough time to collect them all. Some of the treasures are just lying out in the open, others require puzzle-solving. The cash value of each treasure is unknown until your evaluation at the end, so it takes multiple playthroughs to learn which items are worth spending the time to do whatever you need to do to collect them. (This is extremely distinctly characteristic of Verdeterre, and eliminates any possibility that the game wasn’t made in conscious imitation of it.)

That said, Sugarlawn does innovate on the formula. For one thing, it’s a lot bigger. Verdeterre was set on a sinking pirate ship, Sugarlawn in a sprawling southern plantation house turned tourist attraction. This alone has an enormous effect on how one approaches the game; the chief thing that consumes your limited time is simply walking around.

Now, a sinking pirate ship imposes a time limit fairly naturally. To get one in Sugarlawn, the author adds the premise that the whole thing is a game show. This also excuses a variety of other unnatural rules. In addition to simply retrieving treasures for money, you can return them to the locations where they belong: books to the library, for example, or a toy steamboat to a river-themed bedroom. This earns you a bonus which, in many cases, is larger than the value of the treasure itself, but it makes for a lot more running around. You can double this bonus by refusing to use the sack provided for you, which imposes an inventory limit. Such things make the question of optimizing your earnings a lot more complicated.

Another complication: You’re only allowed to carry one key at a time. There’s a box in the foyer where you can exchange a key for another one, but again, this means going back to the foyer, which uses valuable time. Apparently there’s a shortcut: each locked door also has a voice code, a magic word that opens it. I still haven’t figured any of them out. Finding the passwords seems like it would be a major breakthrough in the game, a point where your experience of the thing is utterly transformed and you can really start thinking about optimizing. Before that, maybe you shouldn’t bother.

I did anyway, of course. My first playthrough, which occupied the majority of my time during the judging period, was spent taking the scenario at face value and trying to do as well as I could within the time limit even though I didn’t know anything yet. I was under two time limits, really, the one in the game and the one imposed by the Comp. And I found this quite stressful. Going back to it afterward was much better.

That’s a lot said about the game’s structure and gameplay. The rest of the content — the descriptions and the like — is pleasant and, I suppose, faintly educational. In addition to the historic displays (explicitly somewhat altered for the game show), there’s an announcer who provides additional information whenever you enter a room for the first time, describing its relevance to the history of the house and to the history of Louisiana. Slavery is mentioned, but not really addressed beyond acknowledging that it happened, which, I’m told, is somewhat daring for a historical plantation house. I strongly suspect that all this additional data works into the game’s riddles, the passwords for the doors and certain display cases, but can’t say for sure.

The thing is, even without solving any of the game’s real puzzles, there’s a lot to do here. You can spend a lot of time just running around picking up loose treasures for the pleasure of easy reward. I certainly did.

Truck Quest is a social satire, almost a political cartoon. But it’s served on a substrate of choice-based trucking sim. You know, the same sort of thing as in Euro Truck Simulator 2 and the like, but greatly simplified. You get to choose missions of varying difficulty, the harder ones paying more, then you make some kind of arbitrary choice about your approach (go fast and risk accidents, or go slow and risk being late, sort of thing). Make enough money, and you can pay off your loans and upgrade to a bigger truck that can haul larger loads for more money. Independence. Entrepeneurialism. The American dream.

Except it quickly becomes clear that just trucking will make you money at a dismally slow rate, and you’ll make much better progress if you take on shady side missions from a series of caricatures: a hedge fund manager, a paranoid cyber-warrior, a big-government technocratic politician. Unlike the trucking, these missions are non-interactive and never fail. What’s more, they’re exaggeratedly effective: completing them alters the zeitgeist. Do missions for the hedge fund manager, for example, and people in general become more profit-minded and greedy. Do missions for the privacy advocate, and people become less engaged with their community.

Thinking about it afterward, I think the game missed an opportunity by not having the cultural mood affect the details of the trucking missions you’re offered. Or maybe it did, and it just didn’t make it obvious enough for me to notice.

At any rate, the piece makes its points mainly by showing you what doesn’t change. Between missions, you can talk your trucking mentor, Joan, who had to retire for medical reasons. No matter what changes you wreak on the political climate, Joan can tell you why it’s bad for her and her neighbors. Then there’s Smilin’ Dan, who runs the truck dealership. He sets the terms, gouges you at ruinous interest rates, and then, when you’ve finally made your last payment and own your truck free and clear, he secretly sets fire to it. (It’s pointed out on a few occasions that his name isn’t “Honest Dan” or “Empathetic Dan”.) Smilin’ Dan doesn’t care what the reigning ideology is, because none of the options challenge his power to run roughshod over you.

At one point, a waitress at a truck stop proudly announces that she’s decided to become a trucker herself. The sensible option at that point is to warn her not to do it, to get out while she still can.

So is it a South-Park-like nihilism, where every possible option is equally bad and all you can do about it is point out how bad it all is? Almost, perhaps, but the good ending, which involves exposing the grisly secrets behind a large tech company’s self-driving truck technology, has society reforming around new cooperation between all the factions. The real problem, the author seems to be saying, is imbalance.

It’s just about the least convincing part of the game. The entire scenario is basically an argument against capitalism, but since the author doesn’t trust governmental power, they don’t see socialism as a viable alternative. So instead we get a kind of redeemed capitalism, capitalism purged of its flaws. But the capitalism we see for the rest of the game, capitalism with its flaws exaggerated, looks a lot more familiar.

It’s easy to get the impression that all of Interactive Fiction is based on just two models, the explorable environments typically seen in parser-based text adventures and the branching stories typically seen in hypertext. So it’s good to get occasional reminders that there are other alternatives. Skybreak! is built out of randomized storylets, kind of like Fallen London and Reigns — but more like Reigns in the way it denies player control.

The premise is one of space-opera exploration, zipping from star to star, but the destinations are chosen by your ship’s AI, which claims to be in love with you. When you arrive at a location, you get a series of menus that let you choose what to do there. Once you’ve followed a path through the menu tree all the way to a leaf, it’s time to get back on that ship to another randomly-chosen destination. Often the choices are, unfortunately, meaningless: you arrive at a star with multiple planets and choose which one to explore further, on the basis of nothing more than an orbit number. However, on the basis of a jpeg map included with the game, I think it’s likely that the contents of each star system are fixed, not randomized at runtime. That you could theoretically take meticulous notes about every destination and use that information on subsequent playthroughs. It would take some time, though, because you can’t control where you go, and the galaxy is large.

Large, and uncooperative. At one point I seemed to be stuck in a morass of planets where there was basically nothing to do but mining, which I was terrible at. See, in addition to inventory, there are character stats and skills in this game, and a fairly elaborate character creation process; success at your actions is frequently contingent on where you put your skill points. I had optimized as a scholar/storyteller, figuring that finding the lost history of extinct civilizations would be the most interesting path. But I never found much of that during the Comp’s two-hour judging period.

Character creation involves choosing a race from a short list and two “backgrounds” from a longer one. The races notably include elves and goblins, and the backgrounds include sorcerer. Which is at least honest, I suppose. Backgrounds give you both victory conditions (which I never got anywhere near) and special abilities, some of which involve actions that aren’t in the menus. See, even though the game is basically menu-driven, it’s the sort of command-line-based menus where you type in a number at a prompt. Other commands can be typed in too, such as the ones that display your stats or inventory. This system feels a bit old-school, like Hunt the Wumpus, especially if you play it in a browser, where clickable links are more natural. But I kind of dug it. The game is written in Adrift, which is usually used for second-rate parser-based text adventures. I don’t think I’ve seen an Adrift game on this paradigm before.

The whole thing is just impressively baroque. There’s stats and skills you’ll never use, things you’ll never find. There’s a whole section in the inventory for exotic beetles. It may be best appreciated as an art object rather than played as a game.

So we’re two weeks into this year’s IFComp’s month-and-a-half judging period, and I haven’t posted anything yet. I intend to remedy that, but I’m definitely not going to be reviewing every single game. It’s almost cliché by now, but: This year’s Comp is the biggest ever. There are 82 entries. Last year, I mentioned how Cragne Manor had more authors than the Comp ever had. If I’m not mistaken, that’s no longer the case.

I was actually considering skipping the Comp this year, as I’ve done sometimes in the past. I haven’t posted Comp reviews since 2016, and anyway, who am I to be judging people’s creations, really? But then I looked at the list of titles and got excited. Such variety, such creativity! Browsing the Comp games list gives me much the same feeling as walking into an art supplies store: so much potential! And yet, in both cases, so much eventual disappointment as well.

But why dwell on that? My intention is to do what I did in 2016: post about one game per day for the rest of the judging period, picking just the ones about which I have something to say. I already have a few selected.