The following interviews originally appeared in Computer Games Magazine #185 in shorter form. The interviews are reprinted with permission.

Andrew Plotkin Interview, Part 2

Lara Crigger: What do you think distinguishes a good puzzle from a bad one?

Andrew Plotkin: From the player's point of view? Trivial answer: a good puzzle is solvable but not obvious.

From the author's point of view? A good puzzle is good for all players. This means that all puzzles are bad puzzles. Wait. Let me try that again...

I wish I had a general answer, but there's no secret to good puzzles beyond "try to get it right for as many players as possible".

If you want the player to have an insight about a game element (a command or object), introduce it beforehand—both in the background and interactively. Game mechanics always trump background description, from the player's point of view. Don't tell or show; make the player do it.

Remember that players know less than you. If the player is missing an object he needs, he doesn't know what door it's behind. If the player is having trouble with one possible solution to a puzzle, he doesn't know that alternate solutions exist. What the player believes is important is what the game has emphasized so far—not what you put in the ending.

How did your approach to writing interactive fiction evolve over time? What about your approach to the programming aspects?

My first couple of (serious) games hewed to the 1980s adventure model: you will get it wrong the first ten times you play. You'll die; you'll have to start over; you'll have to keep a file of saved games, and juggle them around as you explore the game. Then you'll have to start over some more.

I evolved out of that pretty darn quick. I don't think it's forbidden territory—many games still use it—but these days it's a model that you have to justify, to yourself and to your players, whenever you uncork it. Over time, I find that fewer and fewer of my game ideas truly require it.

As for programming, I don't think that's changed much. I'm doing the same kind of work I did in 1995. Most IF coding work is shallow: you decide when you want particular messages printed, and you juggle counters and flags until they come out right. Most of the hard work is handled automatically by whichever IF development system you chose.

(The remaining bits of hard work are generally in grafting your ideas into the library code that comes with the IF system. The modern development systems make this as easy as possible, but that doesn't always mean easy.)

I have ideas about better ways to structure IF programs—a new evolutionary step. But I haven't battened them down completely, much less implemented them. So they remain mere hot air, at least for the moment.

The very interactivity of IF makes it difficult to design what most writers consider plot and narrative (since plot tends to hinge on linearity). As an IF author, how have you managed to reconcile this problem and craft a compelling story?

The key insight is that the flow of a story is not a single line. It exists on many levels. There's the overall flow of events. A given event ("Bilbo beats Gollum in a riddle-game, gets the Ring") occurs as a sequence of chapters or scenes. And a scene is a sequence of actions (or perhaps lines of dialogue). There may be intermediate levels as well.

Within each layer, yes, there is a tension between interactivity and scripting. The author has to decide whether to permit one outcome, or two, or a handful, or a vast range. (Does the game allow for the possibility of Gollum keeping the ring? If so, can Bilbo return later and try again? You can imagine straight lines and branches, but also loops, parallel paths, reorderings. And so on.)

However, the author can make these decisions independently for each layer of the story structure. Perhaps the order of game chapters is fixed—but each chapter offers multiple pathways to the (single) outcome that leads to the next chapter. Or perhaps each scene is a linear puzzle, but the way you act while solving that puzzle affects the overall story, leading to many possible scene-orders and outcomes. And, in either case—or any other variation—there is plenty of room for the player to make choices on a move-by-move basis.

The player will always have a sense of which scales allow him freedom, and which restrict him. In other words, a sense for which of his actions are important to the game. (If he doesn't have that sense, the game has already failed.) His play experience will be focused on the interactive levels—that's game mechanics trumping background. But the story can be focused by the other levels, with strong guidance by the author. As long as all the levels come together into a unified work, you can have both the story and the interactivity, without contradiction.

Despite the distinct lack of ghosts, zombies, vampires, or demons, Shade is the creepiest game I've played in years, including Resident Evil, Silent Hill, and Fatal Frame. What was your inspiration for the game, and how did you approach crafting such a suspenseful game?

Thank you, first of all. I'm not sure I remember the original inspiration for Shade. I know I had that core image, of ordinary things falling into sand. And I'd tossed around a form of the story idea, which I jokingly called "Four Endings and a Funeral" -- but that idea changed wildly as I developed the actual game.

The crafting was just—"just"—a matter of keeping that image front and center, and then conveying the mood. The sequence of moods, really. It wouldn't have worked without that passage from ambivalent expectation, through dislocation, through mania, to—whatever it is that characterizes the ending. (Heh.)

Was it suspenseful? I never thought of that term specifically... but I was trying to keep the focus very close: next action, next action, next action. A narrow spotlight; beyond one step ahead is darkness. That's suspense, if you like, but I wouldn't have taken that approach in a non-surrealist game.

(But then, the commercial horror games you name—and I admire Silent Hill and Fatal Frame greatly—also move into surrealism, particularly at the end. I'm just saying that the narrative genre we think of as "suspense"—Die Hard or Sleuth—are a very different thing.)

(I think.)

The $64,000 Question: What makes a good IF game?

Again, I have only obvious answers. A good IF game floods the player with his environment—all the senses, all the time. It has a story, or variations of story, which turn on the player's actions. (Not necessarily his choices.) It invites the player to think inside the game world, by requiring deductions and combinations of game elements which are interesting in the game world's terms. It ensures that the player always has at least one thing he wants to do; and ensures that whatever the player wants to do will always advance the game.

For a short game, pick a few elements and push them hard. (Shade is bright light, gloom, and dry sand. Hunter, in Darkness is dark wet caves and blood. Keep hitting those notes.)

Also, get your spelling, grammar, and punctuation right.

What is it about interactive fiction that has inspired such a community to rally behind it, even after any commercial aspects from the games have disappeared?

People like IF.

What answer did you expect? The Internet is building hundreds of thousands of vital communities about everything. There are at least three IF communities that I know of—they don't overlap much -- and that's not even getting into hypertext, game narratology, and other areas that I'm entirely ignorant about.

Do you think IF titles these days are better than the ones put out 20 years ago?

Yes, certainly they're better these days. The old games had a limited view of what a game was for. There was an exponential expansion between 1980 and 1990, of course, but it was inevitably bounded—both by the requirement to write for a commercial market, and by the small pool of designers working at the adventure companies.

Today's community is diffuse, but it has vastly more viewpoints and ideas in it—simply by virtue of having more people. And then there's the idea of the short IF game, which was popularized in 1995 (quite deliberately) for the first IFComp. That let people experiment with new IF ideas without spending eighteen months to write a "full-sized" game.

What's the best thing about the online community for you? The worst?

There is a huge store of accumulated knowledge, experience, theory, ideas about IF.

However, that does make for a community which is somewhat stagnant; and it's hard for newcomers to join. If you pop in and start posting enthusiastically about your IF ideas, you may not get a lot of response. Either we've seen it, or we've seen something like it, or we're waiting for you to implement the ideas so their strengths and flaws become obvious.

(To be fair, that "write it and get back to us" attitude is not wholly the result of stagnation. It was around in 1995, too. And it caused a lot of people, including me, to write stuff. So I won't say it's a bad climate for a community to have. But it does lead to daunted newbies, sometimes.)