Posted 25 May 2001 to rec.games.int-fiction

I wanted to jot down some short comments about these games and transcripts, but found that for certain reasons I had a lot to say, perhaps because constrained creativity is a topic that is near and dear to my heart, and also because being I'm not very good at being both brief and clear, so I try to favor clarity. And since I wrote about all of them this came out pretty long.

=
= Jigsaw 2 (Adam Cadre)
= Lighan ses Lion (Andrew Plotkin)
=

The thread running through all the games in this competition is "game which both would never have been written if not for this competition and yet does its best to convince you otherwise". This is less common in other competitions; many mini-comps have simply restricted games to "genres" we have seen outside that comp (one-room games, inventory-less games, CYOAs), and games written for SpeedIFs rarely have the time to try to convince you otherwise.

These two authors (and I think it's no coincidence that it's two of the most accomplished IF authors) made the most conscious decision to leave out the "do its best to convince you otherwise". Instead, the authors saw the opportunity of a new medium, the opportunity to write games that unabashedly would never exist outside of the competition.

Adam's game uses the competition walkthrough to humorous ends; I can even imagine this as him taking his "instead of saying 'it would be cool if', just go write the game" advice (see 9:05) to an extreme: noticing something amusing about the competition walkthrough, he chose to make the joke into a game rather than say it on ifMUD. I don't think it's worth saying more about it; suffice it to say I enjoyed it thoroughly for its entire length.

Zarf's "game" goes to a different extreme; rather than focus on the walkthrough, he takes the opportunity to write a transcript for a game that could never really be. Of course, it does use the walkthrough, but I suspect the value is in its entire construction, not in how it cleverly builds on particular properties of the walkthrough. More I cannot say, since delving into it in detail did not attract me personally.

Of course, both of these authors might disagree with my analysis, and whether they do or not, my analysis might be right or wrong; sixteen possibilities all told, so I suppose the odds that I am both right and that they would agree with me are low. I'll spare you "A Dark and Flamey Posting", my choose-your-own-followup game exploring that space, with its easter egg location "Replying About Jigsaw 2 (as Adam Cadre)".

=
= Constraints (Stephen Granade)
= Deus Ex Machina (Gunther Schmidl)
=

I have chosen a rallying cry for my criticism of this comp. That cry is:

WHY A DUCK?

A better cry would be "WHY A LION", since almost all of the games have one of those, whereas half the games choose to have something flying at your head from out of the corner of your eye instead of an actual duck, but I prefer the sound of "WHY A DUCK".

The point of the cry is that few of the games met the challenge full-on, answering the question "why is there a duck in the game" plausibly. Constraints answers it a little, but it quickly invites more questions.

To address the broader question implied by "Why a duck?":

Constraints answers the challenge of the comp by piling on nearly every strategy known to mankind. Where Dreams uses a surreal setting, Constraints uses four different settings, three of which allow "strange" actions: a world of magical realism, surreal dreamworld, the "real" world, and a fantasy world. The use of multiple disparate sections is itself a strategy for combining the multitude of actions found in the transcript.

DEM does something similar, although without all the variations, and steering more in the direction of the absurd than the surreal; the author seems well aware of the implausibility of the whole thing, given his willingness to title it as he did. In DEM and Constraints, the real world setting is used to bring all the chunks together, and they are perfectly realistic and plausible in this way; and yet I can't help feeling that if one's goal is to create a game with a plausible independent-of-the-comp existence, surrealistic actions combined with "it was all a dream" is a cheap way out.

Then again, it worked for The Wizard of Oz.

[Perhaps I am being unfair to Constraints because I liked the first room of the game so much I was disappointed that the game didn't actually deliver that "promised" experience, and am unfairly holding that against it.]

Constraints gets points for connecting the story in its "real world" with the theme of the comp, as reflected in the title; DEM's title also connects to its story beyond simply reflecting the mode of its plot.

One more nice thing about Constraints...

[Gameplay spoilers for Constraints in the next paragraph.]

There is a level on which the author is not concerned about plausibility that is I think one of the crowning moments of this game (much like in Jigsaw 2). The author manages to introduce at least one puzzle *for* players who *are* playing from the walkthrough, and if you haven't played it and are planning on playing it, just stop now, okay? Anyway, when a confronted with the sequence "TAKE NEXT TURN SMOOTH DUCK...", a player might type "TAKE", which picks up some random object. The player quickly discovers that "NEXT" is not a verb, so UNDOes and tries "TAKE NEXT". Instead of taking an object, this appears to mean "WAIT", which is a bit odd, until the player realizes that "TAKE NEXT TURN" is a "plausible" synonym for "WAIT" (not that it would ever appear in a real walkthrough, so plausibility founders a little bit, but that's seemingly unavoidable in this comp). There is, at this point, a duck, one who can be seen to smooth his feathers, so it is with some confusion that the player finds that "SMOOTH DUCK" is not accepted, because "SMOOTH" is not a known verb--which means adding even more words on the end isn't going to help, and suddenly the player (well, I did) has a confused moment: a puzzle even though there's a walkthrough. I was *very* impressed with how Stephen pulled this bit off, and I don't think anybody else in the competition came close (probably because they didn't try), although he achieves a similar effect with "SWITCH PLOVER...".

[Also, in the process of playing Constraints, I discovered that Tads is rather arbitrary about what effect UNDO has after a parser failure; UNDO would step back before "I don't recognize that sentence" properly, but after "don't know the word" it would seemingly undo the previous command as well, as if it didn't count.]

=
= Dreams Run Solid (Caleb Wilson)

Where Constraints is all over the map, Dreams runs straight with a very consistent surreal tone. I think of all of the entries that were actual games, this one seemed to me to be the most plausible "could exist outside the comp", although that degree of plausibility is still not very high. But I think it works better as a game than Constraints, despite the fact that I had 10x the number of things to talk about with Constraints.

=
= Twilight in the Garden of Exile (Alexander Spiridonov)

This game seemingly hung WinFrotz rather early on. (I say seemingly because maybe there was some magic keystroke I was supposed to press--for example, for a while I thought Walk Through Forever had crashed, until I thought to look at the status line. I know Pytho's Mask is great and all, but I really wish people who want to do this sort of thing would switch to a development system that allows the "status line" at the bottom (for Inform authors, that would be Glulx). Input does not belong at the top of the screen. Heck, neither do status lines, according to the human factors literature; it was just a convenient hack for traditional scrolling outputs, which would scroll it off the screen and then you just redrew it. We can do so much better today; note where the status line in most GUI programs is.) So I can't really say much about it, hence the huge parenthetical comment about a different game.

Well, and one more thing. Twilight offers the archtypical example of what I'm calling a "non-puzzle." There's an object. To open it, you need the password. True to walkthrough form, you can just type in the password without learning it in-game. But, just for the sake of thoroughness, what is the puzzle? If you go off-walkthrough, how do you discover it in-game?

By examining the object.

Throughout most of the comp games there were an awful lot of non-puzzles like this.

=
= A Walk Through Forever (Duncan Cross)
= A Venture (Denis Hirschfeldt)
=

Why a duck?

Both of these games try to cover up their absurdity by asserting that they are "bad" games--WTF supposedly incorporating a goofy game the protagonist threw together quickly, and Venture supposedly being an "old sk00l" game.

They are both kind of fun in their own way. WTF's bad game seems kind of pointlessly bad, though, bad to the point of not too fun, although the author makes up for it with things like "A voice from the audience calls out, 'OPENING SQUARE BRACKET'"; but the author seemingly ran out of steam and decide to put in an interminably long unfun sequence which, at least, demonstrates that Graham Nelson was wrong: a crossword can BE the narrative. (Although the section is vaguely amusing in a sort of meta- way, but Jigsaw 2 does it better.)

A Venture has some really beautiful touches--like its use of "ON"--but it also doesn't go to any effort to hide some absurdity that is beyond absurd, like the "WAKE FISH" sequence. Well, it does, through its meta-commentary, which is basically to point out just how hard and unforgiving these "really old skool" games are--they're not your father's old skool games.

The walkthrough sequence "DRAW SWORD WAVE FAN DANCE ABOUT PAINT" was apparently a very hard one for most of the games in the competition to address plausibly; both of these games use it as essentially a time-killer, although Venture makes it a bit more interesting than that--but the point is that the way in which the commands advance the game has very little to do with what the commands actually accomplish.

=
= Bollywood Hijinx (Jamie Murray)
= A Very Strange Day in the Life of a Maid (Mel Brittingham)
=

One of the joys of fiction is that you get the opportunity to experience a new world, something you've never seen before. So I was interested in Bollywood's setting--especially after reading Emily Short's comment on it--and was a bit disappointed with what I actually got; I'm not clear that anything in the game would change if it were rewritten as a wacky game set in Hollywood instead.

A problem I have with these two transcripts is I can't believe in them as transcripts. To some extent that they hew to the form of adventures, by being present-tense second-person with prompted imperatives that specify the actions of the protagonist; but Bollywood lacks room descriptions and any sense of "place" as it unfolds, and both have an awful lot of people just walking by at random at appropriate moments in the transcript, without trying to establish their existence previously; they don't have a real feel of being fuses and daemons. Ok, actually, Strange Day doesn't seem to have that many (the muffin is a notable example), but I definitely came away from it feeling that way.

Bollywood at least made up for it with some cute moments, like the context in which "LION" and "PRAY" are brought together.

=
= Time Bastard (Matt Francisfordcapollasdracula)
= Persistance de la Vision (J. Robinson Wheeler)
=

Of all the games in the comp, I found these two to be the most plausible, the most convincing. I say that about Persistance with severe hesitation. I played Time Bastard last, so my expectations were very low by the time I got to it, and I was pleasantly surprised to see how successful it was. [As I go through this fixing grammar and spelling, I note that I wrote "played" instead of "read", and I think that reflects just how convinced I was.]

All through my reading of Persistance, I apparently misunderstood the tone of voice the author intended. There's clearly some conscious irony: TAKE ALL, or the exchange
   "You have free will, Henri," says Jean-Claude.
   "Not hardly," you sigh.
(which might have one ironic meaning in IF, where the player is in control, not the PC, but which takes on a different meaning in a game so obviously on rails). Because of that, and the implausible inference from the parser at the beginning ["THINK (about Stephanie)"] I had a grand old time reading the whole thing imagining the author chuckling over my shoulder at the sheer implausibility of the whole thing *as a game*; the fact that nearly every command is effectively triggering a cutscene, the fact that most commands are spelled out to the player in the last paragraph before the prompt. But on reading the author's notes it seems he was serious about the material, at least as a story, so I suspect that I was reading this tone of voice into it.

And so I have a lot of trouble imagining that anybody would actually put this thing into a game: except that its author is Rob Wheeler, and I recently read his essay about his "movie player" technology used in Centipede; this transcript seems sort of like the culmination of that strategy, and hey, at least it (perhaps) makes the player type in the commands instead of typing "Z". So, if any other author had written it, I would write it off as "Implausible. A story written in second person present tense, passed off as a transcript." Instead, in the face of its author's other work, I have to consider it plausible; at which point its ranking leaps enormously, because it manages to be the sole game of the comp to avoid surrealism, absurdity, or wacky silliness. [Except when it gets to the dream sequence. But so close! I'll choose to believe that the author did the dream sequence for the fun of it, not because he was too constrained.]

Time Bastard, on the other hand, combines the rather strange elements of time travel, Lovecraftian mythology, and a small amount of wackiness into a whole that is vaguely plausible. It's perhaps not hugely plausible, but it has a relatively coherent world-logic without appealing to "surreal logic" the way something like Dreams does. Moreover, it pulls out a twist ending which gives it a grounding in the real world just as much as Constraints does; so these two aspects together--a comprehensible logic for the weird parts, and an ending that puts it all together in a much-less strange way--combine to create a game that I *could* imagine being written independent of the competition.

For example, I think the game section in Time Bastard for "STAND ON EAST SWING KNIFE LION PRAY" is the most convincing of the comp: a multi-stage puzzle set in a single scene, with all of the components evolving naturally; and while there is a certain silliness to the narration, the situation itself is quite believable. (Time Bastard also used a single technique to work around the odd "ZRBLM" and "XYZZY" commands, which was an example of a *good* way to use the same strategy to get around two problematic commands.)

Time Bastard is (I think) the only transcript which chose to include off-walkthrough commands, which is crucial to creating this believability. It's *possible* that Persistance wouldn't need so much leading text if Rob had done this, but I have my doubts, since its narrative seems to need to run continuously the way it does, without slowing down for off-walkthrough turns.

=
= Fit for a Queen (Margia McPolti)

Actually this game was written by me. I hadn't planned on admitting to it--my oeuvre of released games now comes to three games written under time pressure, none of which are at all representative of my WIPs or what I *want* to be doing on a large scale, and which play less to my strengths as a programmer and more to my weaknesses as a writer--but I feel it would be dishonest to make the previous comments without admitting to it.

So I will take this as an opportunity to talk a bit about rails. First let me say that I find it ironic in the Alanis sense that the author of Galatea and Metamorphoses (and perhaps A Dark and Stormy Entry?) should run a competition in which every submitted game is on rails, often on huge, enormous, blatant rails.

[Every game, AFAICT, except Margia's, the first two-thirds of which (six rooms) can be done in mostly any order, and the last third of which was on rails as a sort of parody; had I known what the other games in the comp would look like, I wouldn't have bothered parodying rails. Then again, I guess what I discovered, struggling with the added constraint "not on rails", is that this was too many constraints; I stayed off rails at the expense of "continuity", hence my choice of genre to try to mask this.]

Admittedly, given the relative successes of "Being Andrew Plotkin", "Shade", and "Rameses", being on rails may not be such a bad thing.

I guess the issue here is that they're not just on rails, or maybe it depends on how you define "being on rails". There seem to be two major additional issues in my mind: the degree to which the player character takes actions without the player choosing them, and the degree to which actions taken have unpredicted consequences.

The phrase "on rails" means many things, but the overriding idea that it is normally used for seems to be a heavy linearity: at each point in the game, you can only do the thing that the author allows you to do; you can perhaps bash around in place trying other things, but nothing is going to happen until you do that one action.

The thing is, games "on rails" often still try to give the player the illusion of control--the actions taken by the player character are the actions specified by the player; you are simply constrained to taking the linear sequence of actions that the author has come up with for you to follow along.

So let me define some terms. I'm going to use "on rails" to mean what I just said above: strictly sequential sequences of allowed commands. I also want a term that refers to the fact that the player character takes actions without the player requesting them; this is common in cutscenes, but I don't want to confuse this with proper cutscenes (which need not even involve the player), so I'll call this "uncontrolled". Finally, there's the fact that commands entered by the player can have unexpected consequence; I'll call this "unpredictabile". (Of course, a certain level of unpredictability is always required in every game, or it's not really a game.)

Let's look at the three games mentioned before. I just picked those three off the top of my head as representative games from the last comp, so they may not be an ideal set, but what the heck. I'm also going to characterize them from memory, despite the fact it's been more than half a year, so I may get this wrong.

"Being Andrew Plotkin" has some major cutscenes, but I don't recall it as being that uncontrolled. The player still chose to sort the files, open the file cabinet, open the door, enter the tunnnel, etc. Similarly, sorting the files sorted the files; opening the file cabinet didn't, since it was stuck; opening the door did; entering the tunnel did, although it led to strange consequences; so I'd deem BAP neither "uncontrolled" nor "unpredictable" in the senses I've defined.

"Shade" was not at-all uncontrolled. Nor was it unpredictable, exactly; the actions had unpredicted consequences at first, but they quickly became predictable--and this unpredictable consequence was really the whole point of the game in the first place. So I'd deem it neither.

The famously ineffective parser in "Rameses" I would tend to deem "unpredictable"; clearly in some cases the player character's actions are not the choices of the player, but that's not really my point of "unpredictability"--Rameses is a special case here, since it's not merely the consequences of the action that are unpredictable, but the action itself. I can't remember how uncontrolled it is; somewhat, certainly, but I think the game tended to give you a parser prompt when it was time for something major. You could argue that that whole parser refusal business is more a matter of "uncontrollability" instead of "unpredictability", I guess.

Why are these issues? Well, uncontrolledness just robs the game of being interactive at all. Unpredictability tends to lead to non-puzzles and anti-puzzles and unmotivated gameplay. Here's a fictitious example Adam Cadre invented in his comp00 reviews:

You're in a cell. You want to get out. The door won't budge, and there's a guard posted outside. You have a gold coin.

GOOD DESIGN:
Get the guard to open the door and let you go free in exchange for the coin.

BAD DESIGN:
Swallow the coin. This randomly causes the door to fall off its hinges onto the guard, allowing you to make a break for it.

Unpredictability can lead to bad design. Fortunately, most of the walkthrough comp games avoided this flavor of unpredictability; instead, the action had its intended consequence, solving the puzzle, but would lead to a further unexpected consequence; let me expand Adam's example with

SURREAL DESIGN:
Get the guard to open the door and let you go free in exchange for the coin, but when he opens the door, the door turns into a duck which eats the guard... and now you have to get past the duck.

Is this good or bad design? The unpredictability doesn't make puzzles unsolveable, but it tends to mess with the sense of overall motivation and progress toward a goal; you're constantly overcoming obstacles that you'd never seen until the previous turn. This is clearly connected to being "on rails", yet not strictly the same thing.

[And none of this even addresses the "leading" I discussed previously, where the puzzles are non-puzzles because their answers are explicitly spelled out in the output from the previous command.]

So how, IMO, do the games stack up in forcing the player forward, in terms of the three jargon terms I'm using here, "on rails", "uncontrolled", and "unpredictable"?

Time Bastard:	on rails, unpredictable at times
Persistance de la Vision:	on rails, uncontrolled
Dreams Run Solid:	on rails, unpredictable
Constraints:	on rails, unpredictable
Walk Through Forever:	on rails, unpredictable
Bollywood Hijinx:	on rails, unpredictable
Deus Ex Machina:	on rails, unpredictable
Very Strange Day:	on rails, unpredictable
A Venture:	on rails, unpredictable
Jigsaw 2:	on rails

So, ok, there wasn't that much use to that exercise, except that if you go along with my analysis you can see that these games are a lot more unpredictable than the comp "on-rails" games. I think "Time Bastard" and "Being Andrew Plotkin" are actually probably similar, but I'd need to replay BAP to be sure.

For comparison, in the main section of Fit for a Queen, there are only a few unpredictable moments, if you stop to examine things first--for example, the consequences of "SEAT ZRBLM" and "WAVE FAN" are clued--but there are exceptions; TURN SMOOTH DUCK and the final outcome of BOWL (although the latter was for tone, not constraint). The only uncontrolled actions are stepping through the door when you unlock it and falling off a platform, which are very minor. To me, all of this makes the game a more plausible actual *game*: you have a goal, you can try to achieve the goal, and you can see how most of the actions you're taking are getting you closer to your goal (if you don't stay strictly on the walkthrough). And achieving all that was the goal I set myself for writing this game.

===

A few comments about common threads to the games:

Almost all of the games avoided the "obvious" interpretation of DRAW SWORD and as a result picked roughly the same interpretation as each other... leading me to question what it means to be "obvious". The game Constraints rather cleverly combined both meanings.

Sleeping fish form another theme; some games avoided it by reading the walkthrough differently, but Time Bastard addressed it head-on in a very clever way--one that I kicked myself for not seeing coming.

Despite the lack of motivation for it in the walkthrough, two games make reference to tentacle porn, and that's not counting Time Bastard.

A few games made use of parser followup questions, despite the way this is unnatural in a walkthrough; SWING KNIFE (at) LION, SWITCH PLOVER (with) EGG, LOOK UP DRESS (in) BOOK [it was interesting to see the more clever ways of avoiding the "obvious" interpretation of "LOOK UP DRESS"]; admittedly, my game seems to have done this more than any other.

Nearly every game ended up with some unused commands, although I think nearly every author was working under the aesthetic that including unused commands was inappropriate (Dreams Run Solid has so many that I don't think it was intended to, and there *is* a school of walkthrough that includes extras). From my notes, it appears that (ignoring the alpha and omega of the comp, Adam and Zarf's games) none of the actual games used every command; of the walkthroughs, Time Bastard is the obvious exception, although I don't didn't take notes about several of them, so A Venture may also have succeeded on this front. (If someone wants to debate this, I'll post the specifics.) There is an overall pattern here which is, I guess, not very surprising: the transcripts manage to meet both the "use every command" and "plausible comp-independent existence" more effectively than the actual games; not very surprising because I imagine that the transcripts would require a lot more effort to actually turn into games than any of the actual games took (although, actually, I'm not sure Time Bastard would be that much bigger than Fit for a Queen).