Mark, I agree that guessing what the author was thinking, at the low level of thinking of the right verb and imagining what the author would have wanted, can be a very uninteresting way to proceed, not only in IF but in other sorts of interactive experiences as well. The piece should have its own logic that can be grasped and and understood, whether that’s the logic of a riddle, of another sort of poem, of a drama, of a conversation, or of something else.

I think richer simulation can offer good opportunities for better interactive experiences; I just don’t think that it will provide this unifying logic to a piece that lacks it. To try to link this to some of Brenda Laurel’s Aristotelian ideas that Michael has built upon: if there aren’t “formal causes” (roughly corresponding to overall purpose and motivations to proceed) supported by “material causes” (roughly, the simulation; the affordances for action that it provides), making the simulation much richer doesn’t solve the problem. If the formal and material causes do match up, making the simulation richer also doesn’t solve anything, but adds new affordances that don’t really serve any purpose. Although I’d prefer a rich simulation to a thin one, the best thing seems to have one that is suitable for the logic of the piece, whatever logic that is.

My personal bias sort of view of it is that filled-in worlds combined with a richer way for the game to take your actions and weave them into the story would help a lot: there shouldn’t be one action I have to do to advance the story, or even five that I can choose from; there should be a near-infinite number, and they should be woven in somehow. Sort of the “dungeons & dragon” model, where you can do whatever you want, and the gamemaster accomodates it into the story as appropriate. There’s tradeoffs: perhaps you simply want to ignore some useless actions, but you shouldn’t ignore all of them besides a prescripted set you’ve coded the game to accept. Ideally, anyway.

“Flailing” can occur whether you use a text interface or not, and the results can be even more absurd without it. I remember clicking around the screen randomly in the first Legend of Kyrandia and being extremely surprised when my character stopped, laid down a trail of jewels, hid behind a rock, and waited to kidnap a passing gnome. Ah yes, just what I was thinking.

The essential problem with “flailing” is not at the interface level. The problem is that the interactor doesn’t have any idea of what meaningful action is. (Unless they’re just failing around to see how the program reacts or for some other reason unrelated to enjoying the game in the usual way, of course.) For instance, why not root around in my room examining and toying with things at the beginning of Future Boy!, since there’s no reason to do anything else? I don’t think that just “implementing” everything in the world solves this problem by itself, although richer simulation is probably better if it fits well with the overall riddle of an interactive fiction. If you just fill out the simulation without changing anything else, though, you’ll just be allowing the user to do 50,000 meaningless things, never getting any closer to understanding the principles behind the strange world that the player character is in.

Instead, it makes sense to make the simulation adequate while also providing scenarios that compel action, and while supporting the interactor in puzzling out the IF world, making this challenging, but at an appropriate level of difficulty.

I think “figuring out what you need to do” in the more general sense is essential to good IF. If it really bothers you at this more general level, it’s probably a sign that IF isn’t for you. Similarly, if you don’t enjoy learning about a character’s experiences, encounters, and perspectives on the world, novels aren’t going to do much for you. “Figuring out” is what makes Varicella, Savoir-Faire, Bad Machine, Spider and Web, Curses, Christminster, A Mind Forever Voyaging, and a bunch of other IF work very powerful and effective. In the best IF, though, the interactor is given a rich and enjoyable way to figure out a strange world, and isn’t stuck trying to operate a interface.

As you suggest, one approach is to tell the player the full list of commands available, which has its pluses and minuses.

Another approach, which some of us are working towards, is to design the IF characters/world itself to robustly handle non-understood or partially-understood text from the player, to never simply reject the player’s input with “I don’t understand”. This probably easier to accomplish when text is only required for speaking dialog, and not for manipulating objects or navigation (i.e., a graphical world, with language solely for dialog). Dialog is inherently messy and ambiguous, and its interpretation can fail gracefully, whereas object manipulation is less so and probably requires stricter, non-ambiguous commands. (Although ideally textual object manipulation commands could be made a lot more robust as well.)

As far as this particular game goes though, I rather like what it’s trying to do, although admittedly I haven’t played it for more than 20 minutes either.