Let’s say that a “modern” commercial parser interactive fiction gets released. It’s perfectly implemented, understand every sensible commands, is not cruel, has beautiful typography, graphics, sounds, or whatever bells and whistles you ever dreamt of.
But it has a two-word parser. As a player, would you care? Would it show a lack of effort? The game has been designed around that, it still has disambiguation, pronouns and whatnot. But it cannot accept more than two words.
Or, from the other side: a new perfect authoring system is released, it has all the features and bells and whistles from above, but it can only be used to create two-words parsers. As an author, would you care? Would you dismiss it right away?
I think I wouldn’t care as a player since most of the commands I type are already two-word. It could get clunky for some actions like putting things in/on other things, but then the game has been designed around that.
As an author, I wouldn’t care that much for the same reasons, except if it makes players avoid my game.
I haven’t played many two-word parsers, but it wouldn’t be an immediate dealbreaker to me. Some of my favorite parser games limit the player’s options, like Eat Me, for example. In a way, I think I favor limited parsers because they generally come with less ambiguity about the correct combination of words to type in.
For an authoring system, the lack of stuff like “give A to B” would be annoying, but I believe you can still design a good game without it.
A lot of my games involving actions with two nouns, like showing something to someone, putting one object in or on another, or using one object to pry another or unlock another.
Allowing two objects to be used in one command makes it harder for people to just try every command, so it makes information-based or inspiration-based puzzles a bit easier to write.
I’ve seen two-word or even one-letter parsers that just use the appropriate item if you have one or that redirect SHOW ITEM to show it to the most relevant available person.
If I had a two-word parser, I would have to stop using that type of puzzles and start using others. I’ve seend a lot that re-introduce that complexity by having you do actions in the correct sequence in different rooms, like the Twine game You that gives you different powers depending on which room you were in, or the very recent game Super Halloween Horror Show where you have to utilize the room layout to use your two-word parser more effectively.
So, a two-word parser system would naturally (in my mind) move away from dense, highly detailed rooms and have more of a spread-out playstyle with each room having few details (like Wizard Sniffer).
Yeah, there’s a lot of potential with puzzles that involve space and time, going from room to room in the correct order at the correct times and so on. Having different rooms change the player in a way that’s necessary for solving future puzzles. I remember Eat Me had a puzzle where you had to perform one action in a specific room and then hurry over to another location, on a time limit, to solve the puzzle right. I got stumped on that one and had to consult the walkthrough. It was the puzzle with the onion, if I remember right, and the weeping ghost?
I’m also remembering another parser game I saw a long time ago, where you had a time machine and needed to use it to do some very specific stuff in a limited area, but couldn’t ever be in the same room as your past self or the resulting time paradox would immediately kill you. And make you lose the game, of course. I don’t remember the name of that game, but I remember it was brutally hard and required intense planning and trial-and-error to get it right. I gave up on it after a few deaths since it wasn’t my thing.
Edit: Though now I have a hankering to watch a real genius solve a complex time-traveling puzzle perfectly. Something like Primer if it actually made sense. Wouldn’t that be something?
Multi-objects actions are still possible, they could be made in several steps for example:
> give money
To whom?
> merchant
You give the money to the merchant.
That what I meant by “clunky”.
I have the feeling that what are commonly called “limited parsers” are not exactly the same thing? They limit the actions a player can take, but not the commands – in a technical kind of way – like a two-word parser? But I might be wrong.
I haven’t checked all of them, but from a quick glance a lot of them are from the 70s-80s, so it doesn’t say much of player’s or author’s expectations today.
It’s interesting you mention consequences in the game design, when my mind was mostly worrying about the technical side.
I didn’t think about that until you brought it up. It would work, and I think the clunkiness could be easy to get used to after a while. “give money [enter] merchant” is actually less keystrokes than “give money to merchant”. It could make things easier for beginners, who know they can only type two words at most in a command. The disambiguation would be annoying to implement, though. I feel like two-word parser engines would still tend towards producing more generalized puzzles instead of highly complex ones requiring very specific actions.
You definitely couldn’t have a complex one-move game like Rematch with a two-word parser, at least not without drastically changing it to be more inconvenient, or ruining the puzzle.
I think a two-word parser is perfectly acceptable, so long as its nature is communicated to players, and puzzles are designed to work within its limits.
As an author, obviously I would only be interested if I was already in the market to make a two-word game, or had an idea that could be made to fit into that limit.
That would be All Things Devours, one of my favorite parser games in high school! It took me ages to plot it all out. The battery puzzle is really a stroke of genius.
I have no issue with two-word parsers if they’re executed well. I have a lot of respect for the Scott Adams games and there’s a lot you can do with a two-word parser—look at the Phoenix mainframe games, with some of the most elaborate puzzles of the mainframe era!
At the same time, though, there’s a lot you can’t do with a two-word parser. If someone really wants to be building a standard parser game (or a choice game) and is stuck with a two-word system, there’s probably going to be a lot of chafing at the limits.
So it’ll all come down to how it’s presented. I’d love to see an elaborate two-word game in IFComp, but I’d want it to be made by someone who’s enjoying the limitations, lipogram-style, not by someone straining against them.
There have been thousands of games with two word parsers. All* the Quilled games (850+) use a two-word parser.
(**As mentioned earlier, many Quilled games used their two word parser to parse four word instructions, back in the day…
TIE ROPE
To what?
TO TREE )
Two word parsers might not cut it for modern IF players, but it’s an interesting restriction to work with. A previous Text Adventure Literacy Jam, took things a little further… Treasure Hunt Jam (Text Only) - itch.io
I’m fine with two-word parsers as long as they’re implemented well. Accept lots of synonyms and ways to say things so that there’s less friction, and I’ll be happy.
I agree with Mathbrush here: I think if you want to use a two-word parser, your best bet is making a limited parser, like Eat Me (where the only thing you can do is EAT X or move in the normal directions). If not, I’m not saying I wouldn’t play it, but it’d be harder to make it work. Not that it’s impossible, but it’s harder.
To be honest I think my main concern about playing a game with a two-word parser wouldn’t be directly to do with the parser at all but would be about whether that signals that it’s a Scott Adams-style retro game aimed at people who have nostalgia for that, which is not me, and implies both design decisions and aspects of writing style/focus that are unlikely to appeal to me.
Beyond that, though, sure it could be implemented clunkily, but so could any other parser game. It’s all about how the game is designed and how the author works with the limitations of whatever system they’re using.
I will admit that if it frequently used that “splitting a >2-word command into multiple commands” workaround mentioned previously, I would be sitting there going “why didn’t you just write this in a system that could handle it without the awkward workaround?” But if the answer was “because I’m nostalgic for the older games that did it this way and I wanted to replicate them as closely as possible including the more janky or awkward aspects, because I’ve come to be attached to those too,” then we’re just back to “this is probably a perfectly fine game that just isn’t for me.”
I’m largely in agreement with EJ here. The one exception would be clever writing. Like a self-consistent diegetic reason that your commands would be limited to two words that adds to the story in some way. Say you were sending a robot outside (the spaceship, the bunker, whatever, outside of safety) but you couldn’t get the AI running the robot to comply, so you removed it’s AI module and it can now only understand and execute short literal commands. Then the inherent difficulty of trying to work around a limitation becomes part of the challenge of the story itself.
I cynically suspect that you could release a game with a two word parser today and nobody would notice for months.
The only commonly taken 3+ word command is PUT object IN container, and you can work around a lack of object containment manipulation puzzles by not make a game where the protagonist does anything that boring.
It’d be a fun experiment to run. Although, if you go through with it, you might have to submit it under a pseudonym after posting that to keep players unsuspecting.
Looking at the high-ranking parser games from last IFComp: Wise-Woman’s Dog has putting spells on things as its core mechanic; The Semantagician’s Assistant is all about putting objects into devices. Phobos has a two word command (HACK thing) for putting your hacking tool on things but I’m fairly sure requires you to GIVE or SHOW something TO someone at least once. So I’m not sure that puzzles requiring multiple nouns are quite as boring as all that.
I should point out that although SHHS has an extremely limited verb set (you can finish the game using only direction commands, EXAMINE and SHOW), it’s “core” verb (SHOW something TO someone) does take two nouns. And the design definitely requires more than one character in the same room some of the time, so you can’t always rely on it inferring who you want to show things to.
@HanonO said he tended to design games which have a lot of two noun commands and so don’t suit what we’re calling a two word parser.
Even though my WIP is a behemoth and also sports what I think is major synonym and alt phrasing support, I feel like it’s ’two word’ at heart. I accept the player’s expression of most ideas that way, but obviously it will also take two noun commands when needed - and in a ton of ways. I personally play by typing as little as possible, even though I cater for everything as an author.
For me, the nicety of the powerful parser is that it can understand the most common ideas, which aren’t the two noun ones, in a zillion ways, with or without synonyms, articles and pronouns.
As people noted above, two-word (often graphical) parser IF was the “European style” of IF during the 80s and early 90s thanks to Quill/GAC/PAW, and if I’m not wrong, the Spanish community here has a modern two-word compiler (DAAD)
Personally, I managed to reduce the inform source of Adventureland to a working skeleton (IIRC that I have posted about it here) and occasionally fool with it (with what result, I dunno…)
On the GIVE MONEY (to MERCHANT) debate, IMVHO can be solved with care in designig: the main reason of giving money to a merchant is buying, so can be reduced, to, say, BUY SWORD.
Side point, reducing an indirect-object action to a direct-object one is an interesting challenge in creativity… taking the buying example above, having only one merchant/seller NPC in scope, e.g. placing a swordsmith and an armorsmith as static NPC in two different locations allows the simple BUYing above.
On the extensive, sparse map instead of a dense, highly detailed map, personally I think is a core design point; for example, frankly, I think that Isekai (which is both extensive and dense…) is utterly incompatible with this design concept, so everyone should endure the long time needed to tackle the bad english issue… but OTOH, back in the 80s was also the issue of too many repetitive rooms (a different issue from the maze issue), so having a sparse but interesting map is a major design/narrative challenge.
OTOH, IIRC that the “adventureland skeleton” compiles fine in .z8 (and its 512 objects…) so techincally IS feasible doing a spread-out sparse map, as defined by Mathbrush in a .z8/adventureland format.