Textcraft: Alpha Island is a parser survival game made in Java. The protagonist is marooned on an islet for a bet, and has to survive for seven days without calling for help; in a standard twist, communication breaks down and you have to survive for real.
It’s a homebrew, which makes a certain amount of sense for a game with mechanical concerns this specific. It uses real-time effects (you can pause, though, and once I figured that out I did it constantly), recolouring the background and text depending on whether it’s sunny, raining, or night. None of the things it’s doing would have been impossible in Inform or TADS, but they’re not the focus or strength of those platforms; I can understand wanting to roll your own for this fairly specialised purpose. My main issue here is that it means the player’s stuck with the same retro-looking monospace font, and that there’s no way to select and copy text or to store transcripts – irksome for the reviewer, this, and not great for testing either.
There’s a materials-based crafting system; you can make a hat out of woven reeds, but you could also make it out of the cloth from your shorts. Materials-based IF has been around since at least Metamorphoses, but it’s rarely-used because it takes quite a lot of work to implement, and requires a pretty extensive game to justify that implementation. (There’s no point in writing rules for buoyancy, say, if there’s only one puzzle in the game that requires it and two objects that might be used for that puzzle.)
Textcraft is not, at present, really playable without keeping a wiki open, largely because its crafting requires obscure guesswork. Two examples:
The island is mostly exposed, and it’s sunny a lot; I figured I’d need a hat. One of the basic resources you can find is a stand of reeds, hats are often made from woven materials, seems reasonable. CRAFT HAT FROM REEDS gives the message ‘Reeds is an insufficient fabric to craft a hat.’ OK, maybe that ‘an’ is just an infelicity of default responses (like, say, ‘reeds is’), and I need more reeds? No. OK, maybe I need to make cloth first: but CLOTH and FABRIC are both unrecognised. What I actually should have gone with is WOVEN MATERIAL. That’s… not great. Guess-the-noun is not a whole lot more fun than guess-the-verb, and Textcraft has its share of both. (You can’t CLIMB a tree or ENTER a shelter; it’s GO UP and GO IN. Until I checked the wiki I assumed the trees were unclimbable.)
A second example: I know I’m going to need food at some point, and there’s not a lot here but ocean, so some kind of fishing seems reasonable. Things I thought of: SPEAR, NET, FISHING LINE. What the game wants me to come up with: FISH TRAP. Which… OK, I guess traditional fish traps are made out of willow or bamboo or whatnot, so reeds aren’t that much of a stretch, and Crafting Game conventions can let us handwave the time and fiddly work it’d take to construct one, but… it’s just not a thing someone’s likely to guess. I’ve played a kajillion crafting-survival games and I am substantially more woodcraft-outdoorsy than the average person and it did not spring to mind.
The other thing is that position matters in crafting; a shelter requires three parts, and you have to declare them in the order (vertical support, structural element, covering). Position of nouns matters in parsers, but most parsers only do so as far as normal English grammar suggests: PUT BREAD IN OVEN doesn’t mean the same thing as PUT OVEN IN BREAD. But normal English assumes that in ‘make X from A, B and C’ A, B, C can be arbitrarily ordered. It doesn’t make sense to expect people to learn that system rather than writing a crafting function that handles that for them.
Sidebar: it’s long been kind of odd to me how consistently ‘crafting games’ are uninterested in, or unable to interestingly show, actual craft. Crafting games show, in fact, the opposite of craft: they assume production to be a factory process, with the same inputs and outputs, consistent and replicable once you have the formula. Crafting is usually instantaneous and reliable once you’ve procured the materials; The Long Dark makes crafting take time and have a failure rate, but the actual process is still invisible. Skyrim crafting produces potions, enchantments, weapons and armour of varying quality depending on the relevant skill, but the process is still abstracted away. This abstraction isn’t inherent to computer games – you expect a walking simulator or an FPS to resemble the experience of the real thing with at least degree of felicity, because walking around in space and shooting at people are what those things are about. But crafting games elide craft; it’s as though every FPS faded to black the first time a gun was drawn, and re-opened in the aftermath of the fight. Crafting games are about gathering, or forming collections, or production workflows.
(I don’t have a comprehensive explanation for why this is, or of its implications, or any real suggestions on how to do otherwise. But it’s still weird.)
OK. So. The indie game market is glutted with survival games, to the point where the basic formula – crafting, fending off hunger and thirst and extreme temperatures – isn’t all that compelling. Survival is a perfectly fine gameplay cycle, and it can provide the central action and main interest of the game – but it doesn’t really stand up on its own.
Games like Don’t Starve, NEO Scavenger, or The Long Dark are motivated as much by exploration as survival – indeed, the two go hand-in-hand, because you exhaust nearby resources pretty quickly. Most involve some kind of unpredictable, non-routine challenge beyond the base food-sleep-shelter cycle – usually that’s combat, although that’s mostly because combat is a thing that games know how to do. And most have a strong aesthetic style, something to hold your interest in the early game when you’re still figuring things out. The Long Dark is all about a moodily-lit, naturalistic Northwoods, the texture of trudging through snow. NEO Scavenger draws heavily on the apocalypse-as-homelessness vibe of The Road; marginal existence, cobbled-together trash, suspicion. Don’t Starve has its macabre-comic vintage children’s-book thing, with echoes of Gorey and Lemony Snicket.
By contrast, Textcraft seems to be aiming at a consciously minimal style: an AFGNCAAP protagonist, a setting that riffs on a stereotypical one-palm-tree desert island, very simple descriptions and a very sparse setting. The setup – you’re here on a bet – emphasizes that this is an artificial challenge. Even the title is carefully generic, an announcement that this is aiming to be just another item in the vast profusion of Minecraft successors. Together with the subtitle, this gave me the ultimate impression that this was intended as a proof-of-concept, an experiment with a story bolted on. But aside from this meta-judgement, it means that Textcraft is a pretty bland experience – and that’s a big problem, because survival games are difficult, and you need players to have some immediate motivations to stick with it through that difficulty. I found it really hard to re-motivate myself to go back to the game after I’d failed at it a couple of times, because it wasn’t offering that much in terms of immediate interest. One of the ways that Textcraft is accurate about survival situations is that it’s got a lot of boring sections where you just have to kill time and put up with discomfort.
There’s weird stuff going on at the edges of the story; on the first night there’s something that seems an awful lot like a nuclear explosion, but other clues suggest that it’s a catastrophic time-travel event. So I was fully expecting to be eaten by a megalodon or something on the eventual swim back to the island. But all of this is delivered through flavour text that’s separated from the main game by quote boxes; it doesn’t really affect what you’re doing. What’s going on may be a mystery, but it’s not really a mystery I can investigate.
I ended up abandoning TextCraft on the third time I ruined everything – I had looked up fish traps and caught some fish, and finally figured out what the game thought was acceptable to build a fire with (you have to use beach grass rather than reeds as tinder, but dumping an entire log on top of your smouldering ember is perfectly fine). I tried to cook my fish, but apparently cooking isn’t a thing, you’re meant to eat sashimi, and PUT X IN FIRE assumes that you’re using X as fuel. (There’s also no UNDO, which I was generally OK with, but… man.)
It kind of takes a long time to die, though, and I left the window open while I worked on the review on a second screen, and hey, I made it through the night and survived up to the morning of Day 6, before getting a twofer death while trying to check fish traps:
You die because you are too cold and you drown because you black out while you are in water.
…which implies an interesting system of event tracking, if nothing else.
Anyway. The amount of work required to get this all set up is pretty impressive, but to my mind this isn’t really done yet. More work needs to be done on the parser and on synonyms; ideally, there should be an in-game way of suggesting things to craft, rather than relying on guesswork and a wiki. And I think it needs something more to hold the interest, something the player is surviving for, both in the short term and the long. I can’t really say what that ought to be; there are lots of possibilities.