Parser-less text adventures: where is everybody?

2018-04-11

When Detectiveland won the Interactive Fiction Competition in 2016, almost everyone praised it. Not just for the excellent atmosphere. Not just for the excellent presentation. But you see, what Robin Johnson achieved was the interactive fiction equivalent of finding the Holy Grail. (Actually, he'd already done it in his previous game Draculaland, but anything released outside of a competition might as well not exist for the community.)

But what exactly did R.J. do? You see, for the longest time interactive fiction was synonymous with a command parser, and that made the genre less and less accessible to potential audiences as time went on. There had been attempts to drive a text adventure through picking from a list of numbered choices, the way gamebooks had worked so well in the analog realm, but that proved a poor way to manipulate the complex world model that had become traditional. And fans of interactive fiction were very fond of their puzzles, you see. Even the runaway success of Twine, which showed that story games didn't need puzzles to be compelling, only served to embitter old timers, who saw their efforts to refine the formula over literally decades suddenly made much less important.

And one game brought the two worlds together seamlessly.

I'd have expected people to start cloning that style of user interface right away. My first prototype, called Adventure Prompt, was ready after only three weeks, including a small but complete game called Secret of the Starry Depths. After all, the hard part when making homebrew games is the parser... and that had just become a non-issue. The world model is easy to code. Not so much to design, but that's why I built on the TinyMUD database schema, a brilliant piece of work I happen to be familiar with.

Nobody else could be bothered to try.

Here we are, 16 months later, and in an ironic twist the very same R.J. just entered the Spring Thing 2018 with Zeppelin Adventure, his third game made with Versificator 2, as the engine appears to be called now. Ironic because I also just revived Adventure Prompt, expanding the runner and adding another way to make games, in addition to a suitable new demo. And it wasn't even on purpose. My intention had been to make an engine for text-based roleplaying games based on the same overall concept, and leave the original prototype behind. It's just that the more I worked on it, the more ideas popped up as to how the various enhancements could be backported, validating the original vision and giving me two engines for the price of one.

We're still the only two people to try advancing this new style of game.

Oh, we're not doing the exact same thing. While Versificator 2 still preserves the principle of having a text stream, with the clickable buttons generating actual commands for a (two-word) parser, Adventure Prompt dispenses with that entirely -- cutting out the middleman, if you like. It also relies on a database, but one that gives all objects the same properties, interpreting them differently based on context. That in turn allows for a variety of simple behaviors with very little code -- 900 lines in the runner right now.

A more subtle difference is that I'm making a proper authoring system, not just a reusable engine. It's a couple of tutorials away from being good for public consumption, but that's just a matter of persistence. Humble I am not.

As for similarities, what both engines have in common is the elephant in the room: for every action to be just one click away, grammar must be limited to the equivalent of a two-word parser, placing any game made this way firmly in the 8-bit era. Oh, you can have quite a bit more text... if you can make it fit on screen without too much scrolling. And of course you can have more of everything: rooms, scenery, items, whatever. What you can't have is the feel of a parser game: since all top-level objects must be shown on screen before you can interact with them, hiding stuff works a whole other way.

Then again, is that such a bad thing? One of the most common complaints you'll hear in reviews of parser-based interactive fiction is, "I got stuck at one point for failing to realize a certain object was present / could be interacted with". As for graphical adventures, don't even get me started about hunt-the-pixel situations...

No, I don't think it can ever quite recapture the feel of a parser-based game. But that's an unreasonable expectation. What it can do, should you choose to see it that way, is enable the creation of text adventures that play more like graphical adventures -- this time the right way. Or like digital gamebooks: someone even compared Secret of the Starry Depths to a Fighting Fantasy look-alike, and that particular game doesn't even have spells or combat in it.

Looks like an underserved niche right there. In other words, an opportunity. So where is everybody?

One possible answer is contained right in the second paragraph: it may well be that nobody cares because bringing back a full world model was a red herring. Authors can do just fine without one in many cases, and that's because games don't have to be complex to be games. Not even if you cling to a definition of the word that looks more and more quaint as time goes on. And embracing that simple idea is a big part of what made Twine revolutionary.

Old timers, on the other hand, couldn't care less, because you see, they love the parser for what it is. Not just because it enables intricate puzzles -- though solving those makes people feel smart -- but also because nothing else can give that feeling of carrying a conversation with the computer (at least until playing via virtual assistants takes off).

Which is perfectly fine, of course. But it does create a gulf nothing can bridge.