This is a spin-off from the post about Jon Blow’s remarks on the IF parser, but it goes in a different direction, so I wanted to take it back to the front page.
I’ve been having a comment exchange with a commenter named Veridical Driver, who suggested a number of possible improvements to the IF interface (automapping, journaling events as they happen, bolded words to show what’s interactive, etc.). I pointed out that there are games that try most of those things; Veridical Driver responded that it’s not enough because IF should be standardized on those features.
So this post started as a response to Veridical Driver’s last comments, especially these bits:
These are things that the IF community may have experimented with, but not things that are any way standardized in the IF interface. The standard IF interface has barely changed from the Infocom days.
Adrift may have mapping, but Inform and z-machine is the standard for IF and do not. Some games might have custom note systems, but this is really something that should be standard, just inventory is standard in all IF. Sure, there is a keyword interface extension… but this kind of functionality should be a standard part of all modern IF.
…The problem is, you are thinking as an IF author, not as a gamer. You don’t like the ideas/features I mentioned, or suggestions other have made, because they constrain your artistic vision. But as a gamer, I don’t care, I just want some fun.
Nnno, I don’t think that’s quite it. Two of the examples I pointed to (Floatpoint, Bronze) are my own games; other projects of mine (especially Alabaster and City of Secrets) include graphical sidebar content that’s nonstandard but is designed to ease player experience and communicate game state better. So it’s not that I dislike these features categorically.
Where I’m pushing back is on the idea that we can or should enforce these features as a standard.
There I’m speaking not just as an artist, though I can think of several of my works for which the features you describe would be a bizarre and awkward prosthesis on the text — what’s automapping for in a one-room conversation game? what’s journaling for, in a game that runs for five minutes and is designed to be replayed?
But setting that aside, I’m also coming to this as someone who’s handled a lot of feedback on one of the most-used tools in the IF community for the last five or six years. People want to do a lot of different things with their interactive fiction, and they should have the opportunity to try their various visions. Some specific use cases, other than the artistic concerns I already mentioned, where your suggestions might be an active hindrance include
- games intended for mobile platforms or small screens, where screen real estate is at a premium
- works for the visually impaired, which need to be simply accessible with a screen reader
- works written with a heavy narrative focus, which may put aside the concept of “rooms” entirely in favor of a different style of presentation; these aren’t always even intended for a gaming audience at all
These aren’t hypothetical; they’re things that people are actually working on and are the basis of real support requests.
So the issue is, tools that force too many features run a big risk of narrowing the creative range to just the projects that work well with those features. Inform has tried to err on the side of making a lot of things optional — through extensions — while not imposing too many constraints through core library decisions. This is always an area of compromise, and there are some features we’ve added that have made Inform games larger, to the chagrin of those optimizing for very small, low-processing-power machines. So these things are always on our minds.
I’m happy to say that a lot of progress has been happening, and continues to happen, on the extensions and interpreters side. The desire to foster collaboration, conversation, and creative thinking about IF interfaces is a major part of the impetus for the IF Demo Fair we’re putting together for PAX East.
Still, this opt-in stuff is obviously more work, and it’s not going to force authors to include the features you’re looking for — and the novice authors are the ones least likely to put in the extra work if the tool doesn’t make them do so. I typically consider it a good sign — not always but often — if I start up a competition game and find that it has cover art, a splash screen, a non-standard status bar, etc. That’s not because I think those are universally important, but because it means the author put some time into generating non-default content. Which means he thought about it. Which is good.
From a game consumer’s point of view, I think what would help the most is curated collections and branding.
What I mean by this is not that IF necessarily has to go commercial to re-find its audience (though there are a lot of people who *do* think that).
Rather, the community does not currently have a good way of saying to outside players “this game meets quality standards X, Y, and Z, and it provides a gameplay experience of an established type that you’re already familiar with.” We’ve got reviews and awards and rankings and star counts on IFDB, and that’s all really great and useful. But we don’t have a way to
- create and enforce a standardized-yet-not-bare-bones look and feel
- build brand identification and loyalty in a way that would make new materials recognizable to prospective players
What we’d need is an editor or editorial body that would
- define a clear brand identity, such as “NPC-rich games with a romantic theme” or “traditional puzzle games with attractive automapping and hints”
- solicit and accept game submissions from authors
- try them out to make sure they conformed to the brand guidelines and do some QA
- provide support to polish the game in accordance with branding, especially in those areas where individual authors might have a hard time meeting expectations. This might include hiring artists for in-game assets and the cover, or supplying a standard automap extension that everyone had to use
- publicizing the games of this series in a standardized way, on a website or similar, with marketing teasers, press releases, author bios, interviews, and probably a support forum
This editor would need to offer identity definition, graphical design, website development, understanding of user experience, organization skills, sufficient tact to deal with authors, and sufficient clarity about the project to ignore or reject angry complaints and community backlash along the lines of “this is so elitist!” or “why didn’t you accept my masterwork, you clueless buffoon???” Which, I’m reliably told, is the constant lot of editors in other fields, except that they typically get money.
If this were a commercial concern, they’d also need to provide community support and customer service to players, a web store, contracts and royalties for authors, etc.
The kind of game you’re describing — IF for a gaming audience, with polished UI trappings for movement, compass information, journaled quests, etc. — would be an appropriate direction for one such brand. Having a brand focused on this could also increase the quality level on those projects.
I really liked The King of Shreds and Patches, for instance, which does have automapping, journaling, and a bunch of other features similar to what you’re describing, and was definitely developed with an eye to what puzzly, large-scale IF could learn from RPGs and large-scale commercial games. That said, I think the features that were there could have been presented more attractively with a slightly better interpreter toolset and the services of a graphic designer. I don’t mean this to knock Jimmy Maher’s work at all — I don’t know that I could have done better, and the features he provided really make the game easier and more fun to play. But when I look at a screenshot, I see something that doesn’t suggest the involvement of a graphical designer: the proportions are not quite right, the contrast is sometimes hard to read, I’d tweak the typography here and there…
…and these things make a huge difference to user experience and perception. A lot of people, myself often included, would rather entirely skip adding certain features to their games than add ones that look painfully amateurish. (And some of the things I did years back looked okay to me at the time, but now look wrong.)
I can’t blame users for making judgments on this basis: I constantly judge things this way. If I have two websites offering me apparently identical products, I guarantee you I’m going to buy from the more attractive, more polished site, because competence in that area is a shorthand signal for professionalism and competence in a lot of others. For very low-information transactions, this signal is the only way I have to judge quickly whether the sale is likely to be legitimate, or whether I’m being offered eyeglasses by a dodgy con artist operating from his garage. With IF, we’re not trying to make people pay, but we are asking for their time, which also has value. So quality signals are hugely important, and an editorial/curatorial body would have the job of making sure those signals were clear and consistent for the brand they were promoting.
If the next question is “why doesn’t an editorial branding body exist, then?” …well, I suspect because it would be an enormous but largely thankless job to be the editor, publisher, and support staff for this project.
Possibly there does exist somewhere a person who has both the skills and personality to take on such a job and who would enjoy doing it for its own sake.
If there isn’t, though, then there has to be some kind of incentive to be that person. A commercial venture is one possibility, but a risky one, and it’s probably conceptually harder and riskier to start a publishing company than to write some games yourself and try to market them. (Ha ha. As if that were easy to start with. I realize that’s like saying “it’s easier to fly to Mars than to Saturn.”)
Failing that, I’ve occasionally thought about whether authors would support such a service up front: I might be willing to pay someone that I thought had the skills to help me polish and present a game of mine, if I thought it would ratchet the production values up into where they looked like a commercial project.
But there are problems there too. I’m probably at the extreme high end in terms of my willingness to spend money on developing freeware, and even I doubt I’d be willing to spend the amount necessary to make this really worth the time of the consultant and his subsidiary artists and designers. I can’t imagine it would cost less than about $10,000 to roll out the WIP I’m working on now, and possibly quite a bit more depending on what we wanted to do with it: say 100 hours of QA; artist time to provide cover art, an attractive map, and a poster art series for the teasers; editorial time to assess and comment, create website content, generate a jacket blurb and write press releases. And that assumes that the publisher website has already been completely built and has a reliable CMS.
This is not a business model for freeware.
The other problem with the artist-hires-editors model is this: if part of the point is to develop a brand and brand identity, then the editor is partly responsible to the players, which is not compatible with being financially dependent on the authors. I don’t want a vanity press here, where the richest authors get glitzy packaging and game quality becomes totally decoupled from presentation quality.
So I don’t know what the solution is. But forcing features into the IF creation tools is absolutely the wrong way to standardize, and community consensus doesn’t (and isn’t going to) exist about negatively reviewing games that don’t have automap, for instance, the way people tend to negatively review a game with mazes or a hunger puzzle.
We’d need a third route, something involving an editorial voice and vision.
Emily Short's Interactive Storytelling 
I want to respond to one small part of this post, the idea of the need for some way to explicitly highlight things that can be interacted with, with two points.
First the very strong downside to this approach is that it works against the natural feel of a seamless world that text is able to present. Obviously such a world is not truly seamless, but to emphasize the seams by highlighting all the interactive bits explicitly, weakens any atmosphere created by the prose.
Secondly it is patently unnecessary. I’ve played IF games where figuring out what is and isn’t interactive is a complete non-issue. They simply make it clear from the response whether the object you’ve tried to interact with is not something with which you need to interact. I seem to recall a phrase like “I don’t think you need to do anything with the ___” but there are many possible variations. It really is not such a burden to the user to now and then eliminate objects this way. It only becomes a burden when the number of possible objects in room descriptions is multiplied by the number of ways one might interact with them without eliminating them as interactive.
First the very strong downside to this approach is that it works against the natural feel of a seamless world that text is able to present. Obviously such a world is not truly seamless, but to emphasize the seams by highlighting all the interactive bits explicitly, weakens any atmosphere created by the prose.
Yeah, I agree. Some people like this, but to the extent I want it at all, I want it as an occasional opt-in feature, a way to find things in the prose that I somehow stupidly missed (or that the author missed drawing attention to).
Secondly it is patently unnecessary. I’ve played IF games where figuring out what is and isn’t interactive is a complete non-issue.
I have some anecdotal evidence that novices find this somewhat challenging, because they haven’t clued into some of the basics — like the expectation that a thing described by a concrete noun is there to interact with (and *should* be interacted with, often), while something whose presence is merely situationally implied, or described abstractly, is not.
So the problem is partly figuring out what is and what isn’t an object, and this is a piece of world model construction that most veteran IF players do automatically without thinking about it. But I bet if you asked a novice player to make a list of the things that were in the first room of, oh, I dunno, Anchorhead, he’d come up with a very different list than a seasoned player.
The second issue is that having the back-up — having a way to force the parser to reveal affordances — means that it’s easier to slightly hack your way past those points where the author’s design let him down. I would still prefer games to be written in such a way that I can play smoothly without this kind of help.
But I’ve also given up on a lot of games where I could have scooted past one or two sticky places in the design if only I’d be able to force the game to tell me something that the author had neglected to signal clearly. Like “what verb can I use on this explosive to mean I want to stick a fuse to it?” or “where are the exits? this room description is too vague” or “that vague room description reference to the room being messy wasn’t just scenery; I’m actually supposed to SEARCH MESS.” Highlighting is a way to address part of that.
So the problem is partly figuring out what is and what isn’t an object, and this is a piece of world model construction that most veteran IF players do automatically without thinking about it.
This is likely going to rub a lot of IF player and authors the wrong way, but I think there’s a tension between game writing and story writing that contributes to this a great deal. The tendency among a lot of IF authors is to draw in a lot of influence from the static prose that they admire.
Early IF tended to be much more laconic in telling the player what was available as part of the game world, which also tended to grow a bit stilted when read. More recently, authors have worked at upping the literary ante by making descriptions more expressive. Novices generally weren’t privy to that evolution, and that, I think, is where the confusion comes in.
One solution would be to go back to a more utilitarian mode of description (“You see a short-handled shovel and a long-handled shovel here”) but I don’t think very many of us are going to embrace that idea. The better (and consequently trickier) solution is to look for a functional middle-ground that allows IF prose to be expressive without sacrificing its utility.
This may simply be a stylistic preference, but I’d say a good way of approaching that middle-ground is by looking at IF prose style as something that accumulates, rather than something that must be established in the longer segments of text.
This is likely going to rub a lot of IF player and authors the wrong way…
Not at all. (Or perhaps I should say: not me.) I’ve written before about IF prose vs static prose, and hit many of the same points. Though my argument there is that this doesn’t have to be a game vs story thing, or a sign that you have to make all your prose highly utilitarian. Just that the aesthetic and functional requirements are a bit different than they are on the page, and compactness and clarity are very valuable.
Eric:
First the very strong downside to this approach is that it works against the natural feel of a seamless world that text is able to present.
This is something I’ve been thinking about a great deal in the wake of the Jonathan Blow-inspired discussions of the last several days, and my opinion at the moment is that we as a community may have put too high a premium on notions like immersion and seamlessness. A certain consistency of tone and style is generally a major plus in a work of IF, but I’m starting to think that IF as a whole will benefit if we start looking at the parser as the border that connects the game world to the more mundane reality from which the player approaches the game. That border can probably be a great deal more permeable than we’ve so far been inclined to treat it, and without significantly undermining the experience of playing.
As for your second point, I’m mostly in agreement. I think turning our focus to two different strategies will go a long way toward helping ease the frustration a lot of players feel in trying to determine the appropriate terms of interaction with a given game world. One is simply the development of more “best practices” heuristics. These wouldn’t be inviolable rules by which any given game could be judged, but rather conventions that an author would be wise to adhere to, to whatever extent they want their game to be accessible to a general audience. Most of the movies you see, for instance, adhere to a set of best practices that are mostly invisible to audiences because they make up a kind of lingua franca of cinematic communication. Movies that we think of as being more experimental are generally playing against that syntax.
The other is providing a rich backdrop of extensible authoring tools. Emily and others have talked about the immense amount of work that can go into trying to preload a game world with responses to every plausible interaction the player might try. And there are already some Inform extensions that attempt to ease that burden. But it seems to me that there could be a great many more packages to deal with specific sets that are only essential to preventing the break-down of the player’s commitment to a game that isn’t responsive enough.
“if only I’d be able to force the game to tell me something that the author had neglected to signal clearly. Like “what verb can I use on this explosive to mean I want to stick a fuse to it?” or “where are the exits? this room description is too vague” or “that vague room description reference to the room being messy wasn’t just scenery; ”
I think all 3 of those examples are/were/can be handled by the meta commands EXITS, VERBS, and OBJECTS, respectively, but it seems to me meta commands have been falling out of favor recently. EXITS is usually in the status line at all times, OBJECTS (or NOUNS) I saw in a few Inform 6 (?) games, VERBS, rarely.
Just an observation.
@Vertical Driver: TADS 3 is a good example of including everything at start. But authors have trouble turning bits of it off, and sometimes the feature will spoil a game. Who had the problem of the message, “The last time you saw the X, person Y had it.” which spoiled part of the story?
I’m lukewarm to your ideas, but perhaps auto-mapping isn’t the answer when ditching the map entirely is. Just for example.
@Eric: “I don’t think you need to do anything with the ___”
I second Em that new players don’t get much out of such messages, having been a new player not too long ago. But when she says that novices lack “the expectation that a thing described by a concrete noun is there to interact with while something whose presence is merely situationally implied, or described abstractly, is not,” then I make the observation that this may become less true in time, as conversation games frequently talk about abstract or absent things. This sets us back to square one, where even veteran players (including Eric) are unsure of the rules.
But mainly I wanted to respond to Eric when he says, “the very strong downside to [highlighting words] is that it works against the natural feel of a seamless world that text is able to present. […] to emphasize the seams by highlighting all the interactive bits explicitly weakens any atmosphere created by the prose.”
This I believe can be written around. I feel highlighted words belong in the last paragraph (or sentence, etc.) of a game response, because it’s both right next to the prompt where it’s needed, and because it allows the earlier writing to establish mood without the needs of interactivity interrupting.
Further, I feel the need to highlight words diminishes as the story progresses. I think a work might tend to follow a pattern of introducing the player vocabulary in Act I, and perhaps some of Act II, but for the latter half or so of the game the player should focus on using the vocabulary he’s got to effect what changes he’d like to see in events, rather than hoping a new vocabulary word will drop from heaven during the story’s climax like a deus ex machina.
A player must be able to plan ahead to feel like he has any control — this is true in videogames as well, such as examining a Braid level for a minute before moving — and I believe IF needs to recognize that. If it does, then that gives the highlighted-words feature an organizing principle.
I think all 3 of those examples are/were/can be handled by the meta commands EXITS, VERBS, and OBJECTS, respectively, but it seems to me meta commands have been falling out of favor recently. EXITS is usually in the status line at all times, OBJECTS (or NOUNS) I saw in a few Inform 6 (?) games, VERBS, rarely.
Just an observation.
EXITS I think we have better and better covered, and yeah, my personal preference if it’s going to matter a lot is to use a Bronze-like approach to the compass (or something graphical rather than ASCII, but with the same function). But I also do enjoy games that list exits in other approaches.
The OBJECTS verb standard in I6 never listed what was in the room with you; what it did was tell you which object you had handled, which explicitly leaves out stuff you haven’t noticed enough to interact with yet. Not so useful. But what I really want is not so much a magic meta-ability to know everything that’s in scope — that would break many game scenarios — but a way to compel the subtleties out of the text when I just wasn’t getting them.
VERBS is implemented by some games, more older ones than newer, though a number of modern games do include some help that tells you about important commands. I think it would be nice to have this a little more specifically targeted, though. For instance, I don’t always want a list of every object in the game; I want a list of verbs that are going to interact with this particular object. Which might be spoilery, but for many games it wouldn’t actually give away the puzzles unless they were guess the verb puzzles to start with. It’s the equivalent of clicking on an object in Sims and getting a radial menu of interactions.
This seems — maybe just to me — like an interesting related tangent:
Paradox is a company that, supposedly, should be dead. They’re a Swedish developer of deep, complicated grand strategy games like Europa Universalis and Victoria — games that do require you to RTFM, games with a deep connection to history, games that challenge the intellect rather than fine motor skills. Their games are neither casual — they are far too challenging and complex for that — nor “hardcore,” if hardcore means requiring l33t shooter sk1llz. Their games are of the type that Johnny Wilson would once have given lengthy ink in the lost, lamented Computer Gaming World, games that demonstrate that digital games need not be about joysticks.
Isn’t that something like what IF strives to be?
mattw:
Europe Universalis is a complicated game, but its GUI is a very conventional and easy to use interface. The challenge comes from the strategy, not the interface.
Well, I haven’t played it, and for some reason Costik linked to a post that doesn’t actually include any of the figures he mentions, but this doesn’t sound straightforward to me:
As a player, you spend your time monitoring your stability, displayed as a number from -3 to +3 at the top of the screen; checking how you are investing your tax revenues, and adjusting sliders to spend more or less on different things; pondering your domestic policy sliders and which to adjust next, and considering your relationship rating (a number from -200 to +200) with other powers, and whether it’s worth trying to change it. You’re always aware of your income, your rate of inflation, and your maximum manpower, and you’re juggling many different factors, trying to keep any one from spinning out of control…. These sliders determine how your monthly taxes are spent. You’ll see that all sliders are positioned at the same point, meaning they’re all set equally; you will most definitely not want this…. . So we slide “Naval” all the way to the left and fix it there (double-clicking to make sure it doesn’t change when other sliders move)…. Rolling over any of these gives us important information — including the fact that one level increase in Government will give us a “National Idea,” and one level increase in Production will allow us to build Workshops…. At the game start, we can adjust one slider one point in one direction; which you choose depends on your overall strategy, but mousing over anything tells you what a shift does…. Please note I’ve yet to actually play the game; it’s been paused all this time.
I mean, shifting sliders here, double-clicking on them there, mousing over that, all before we’ve started playing? Bluh. Compared to that, typing in words and hitting “enter” seems kind of straightforward. To me, anyway.
Which connects to something I was thinking about in your posts on the other thread. You said that you are a typical gamer, and that you find the IF interface frustrating. But I think IF ought to be trying to reach people who aren’t typical gamers, and that includes people who find typical gaming interfaces frustrating. There ought to be a potential audience in IF among people who have no interest in platformers, or real-time strategy games, or first-person shooters, or survival horror, or what have you. (Maybe there’s more intersection with RPGs.)
In those categories, I like platformers, but a lot of other IF folk don’t; I don’t know whether their issues count as interface problems, but I’m pretty sure it was an interface problem that had me howling with frustration at Lugaru during the tutorial. I think it takes a lot of training to be able to manipulate the mouse + four buttons when the camera is swinging around like crazy.
So, thinking about the IF interface is definitely worth doing; but I don’t think it’s necessarily a good idea to make it more like other games. Those have barriers that gamers can’t see. And (this is all I was really trying to get at) other kinds of game seem to be able to develop a niche even when they require a pretty big initial investment, because they reward the investment. Maybe IF should strive for that… which means something like the imprimatur of quality Emily discusses might be the way to go.
Emily, note that Inform already builds in a lot of defaults that are a lot of work to override. Hence, designers stick to those defaults, for better or worse.
It may seem that designer can shape everything to their will and therefore will do so, but in reality, that’s just not the case. Writing IF is a huge amount of work and even if you have mastered Inform, you will invariably end up sticking to the defaults in most places and focus on a few particular things that are out of the ordinary. If the defaults mean that the UI is less user friendly than it could be, then it will be. Since the main problem of IF is the UI, that’s a sad state of affairs.
An example from my own experience: I once wanted to write an IF that mostly progresses by scenes. But since the default in Inform are places and objects, I didn’t get very far. I can’t easily change the defaults either because Inform is one of the weirdest programming languages I have ever used. I can keep C++ and LISP in my head, but in Inform, it is simply impossible to do the most simplest of things without studying the manual for hours.
Emily, note that Inform already builds in a lot of defaults that are a lot of work to override. Hence, designers stick to those defaults, for better or worse.
Oh, heavens, I know! But the history of Inform 7’s development has been for the most part about removing some of those defaults in favor of greater flexibility. It’s still not a perfect tool for everything, or anything close to that, but it’s a better tool for more things within its basic domain. (It also defines is basic domain to be parser-based IF and not general-purpose programming, which is why Inform 6 is still very necessary if you want to write a game of Tetris for the z-Machine or something. So there’s room for debate. Nonetheless, many of the library-level changes are about removing assumptions that we were previously making about what the author should want to do.)
If the defaults mean that the UI is less user friendly than it could be, then it will be. Since the main problem of IF is the UI, that’s a sad state of affairs.
Okay; in that case, what UI changes do you think would make most games easier to play as a rule? Asking “what would make all games easier to play?” is too high a bar to set, but let’s say rather “most games, for most people”?
If there are clear answers to this, then it’s definitely worth talking about what we should be doing. Or, if the answers aren’t quite things that belong in Inform itself, whether it’s worthwhile for someone to create some template extensions — say a master extension that includes all the required elements to make your game conform to a particular standard, like having automapping, journaling, and the other things the original commenter brought up. I know that’s still not a default, but a complete, one-package solution would be more appealing to a lot of authors than picking and choosing elements for themselves and then dealing with any extension clashes.
Anyway, I didn’t mean to present this as a solved problem or write an ad for Inform — I know it interacts very differently with different people’s mental approaches, is opaque and frustrating for many, and so on. I meant rather that the process of working on Inform has made me sensitive to the problem of defaults that don’t do what the user wants, and that we’ve specifically adopted a policy of asking about new features whether they impose new restrictions on authors. So I’m wary of using the toolset as a means of enforcement.
(I’m also curious what problems you ran into with scenes — while it would be hard to get rid of rooms entirely, it’s possible a la Glass to use just one room and never mention its name or existence to the player, and then just not implement any scenery as objects. But possibly that wasn’t what you were looking for?)
Okay; in that case, what UI changes do you think would make most games easier to play as a rule? […] If there are clear answers to this, then it’s definitely worth talking about what we should be doing.
I take it you are implying that there are no clear answers to this at the moment, and I have to say: good point. :-) Of course, my counterpoint is that this doesn’t mean that the status quo is any good.
As an IF author, I would not mind the default so much, in particular Inform 7’s default, if it were easy to change for me. But despite extensive programming experience on my part, I have to say that this is extremely difficult, to the point of impossible. Unfortunately, this is by design: as a programming languages, Inform 7 is really awkward. Basically, it only works because it offers a lot of defaults and specialized concepts for common activities, but it has no underlying simple mental model that can be used to program arbitrary things. This is great for doing standard IF stuff, but a curse for anything modestly radical.
In short, it is neigh impossible for me to write an Inform 7 extension that implements things like journaling etc (and meet my expectations of clean code. It certainly is not too difficult to produce an extension like this that is awkward to use and fails to be sufficiently general).
Overall, my (exaggerated) argument is that because you are the only person who understands Inform 7, you are the only one who shapes what future UIs look like. It is you who sets the default and almost nobody will be able to change it because they don’t know how to do that. (I am of course making the ridiculous error of mistaking other people’s abilities for mine, but I do consider my programming experience solid enough to not dismiss this thought entirely.)
As for scenes, my main trouble was that starting and ending has to be encoded with objects. This makes sense if the manipulation of objects is the main trigger for scenes, but in my case, I wanted the flow of time and scenes to be the main trigger for the behavior of objects. Ah, in particular I had trouble with continuous actions, like a spider crawling up the player’s neck and (s)he has five turns to do something about it. It’s easy enough to program special cases (encode the time state as an extra property of the object), but I have so many of them that I long for a general programming pattern that can instantiated for each of my many examples.
As for scenes, my main trouble was that starting and ending has to be encoded with objects. This makes sense if the manipulation of objects is the main trigger for scenes, but in my case, I wanted the flow of time and scenes to be the main trigger for the behavior of objects. Ah, in particular I had trouble with continuous actions, like a spider crawling up the player’s neck and (s)he has five turns to do something about it.
Hrm, this should probably go to intfiction.org, but can’t you accomplish this with timed countdowns and/or named events? Something like:
When Spider Crawling begins: The spider bites in five turns from now.
Spider Crawling ends in success when the spider is killed.
Spider Crawling ends in failure at the time when the spider bites.
Death by Spider Bite begins when Spider Crawling ends in failure.
Cleaning Up begins when Spider Crawling ends in success.
Or:
The spider count is a number that varies. The spider count is usually 5.
Every turn during Spider Crawling: decrease the spider count by 1.
Spider Crawling ends in success when the player kills the spider. Spider Crawling ends in failure when the spider count is 0.
etc.
I’m sure that doesn’t even compile (I haven’t looked it up or checked it, and I’ve never used scenes), but if you took it over to the intfiction.org forums there’d be people who could help you with this.
Overall, my (exaggerated) argument is that because you are the only person who understands Inform 7, you are the only one who shapes what future UIs look like.
I understand what you’re getting at, but it is indeed exaggerated — not only because there’s a bunch of interesting stuff that’s been done by other people (and there has), but because I personally have less control over UI behavior than almost anyone else involved with the Inform project. :) When I said I “handled the feedback,” this is because a large majority of what I do with Inform is essentially author support and advocacy, trying to explain Inform to users and help draw Graham’s attention to especially pressing problems people are having. I don’t code the interpreters or set the standards for Glk or change the core code of Inform, all of which have more immediate bearing on this issue.
That said, I think a lot more could be done to make things flexible. And indeed Erik Temple has made enormous strides in this department for Glulx windowing; Andrew Plotkin and Dannii, and perhaps others, are doing work on browser presentation that might make it possible to do interesting things in CSS; and Dave Cornelson on Zifmia, which is also meant to let people essentially roll their own front ends, as I understand it.
My hope here is to see it get easier and easier for people to make what are effectively presentation templates for their IF, and for authors to plug these in.
But I also think that when we get to elements like graphical layout, making windows and building automaps, we really want to be drawing wherever possible on external ways to define these things, rather than building them into the core of Inform. I may disagree with you a bit about Inform’s expressiveness in general, but when it comes to layout-oriented tasks, it’s a bit like painting a miniature while wearing oven mitts.
As for scenes, my main trouble was that starting and ending has to be encoded with objects. This makes sense if the manipulation of objects is the main trigger for scenes, but in my case, I wanted the flow of time and scenes to be the main trigger for the behavior of objects.
Hmm, okay. It’s also not true that the starting and ending “has to be encoded with objects” — any condition you can express in code can change the state of things. As Matt says, a detailed discussion of this would belong on intfiction, but if there’s major functionality that you’d like available and you can’t get it to work, people tend to be fairly helpful there, and I’d be interested to know as well. Scenes have gotten progressively more powerful over time, but there’s probably more that could be done with them; it’s easiest to make a case to Graham about what we need if I have specific user feedback to point to.
Anyway, I think our points are pulling in two different directions here, and we actually agree philosophically about what I was originally saying (defaults should not be used to force new exotic features on the authors) while having different experience with the particular tool of Inform. That said, if you find Inform fundamentally awkward and annoying, then it’s probably not for you. I’m sorry you find it so, but fortunately there are other excellent IF languages out there.
To “Clown”: the Programmer’s Manual to Inform 7 may be more helpful to you in learning Inform. The PDF is at http://xyzzy.plover.net/~pscion/Inform%207%20for%20Programmers.pdf
Matt: I’m sure that doesn’t even compile (I haven’t looked it up or checked it, and I’ve never used scenes)
Ah, but this sentiment is my very problem. :) It is not difficult for me to come up with some encoding of the logic, the trouble is knowing whether Inform will accept it. This very much depends on idiosyncratic details and I usually spend hours trying to find the right way of expressing myself in Inform.
The devil is indeed in the details. For instance, Matt’s first version has a serious problem: “The spider bites in five turns from now.” Since “spider bites” is the name of an action, it will happen in five turns even if the spider is dead. Oops. :O The action “spider bites” will have to check for the presence of the spider, which is weird, because the name “spider bites” makes it sound like the action adheres to the invariant that it is only executed when the spider is alive.
A much nicer way to specify the behavior of the spider would be to code it as a sequence of actions:
spider live = wait; wait; wait; wait; kill player.
This way of specifying the behavior of the spider has the invariant built-in. If the spider is removed from the game, it’s life will not finish. Unfortunately, I can’t express it in Inform that way. (Perhaps I can, but I would just run into other devilish details.)
I’m certain that people on intfiction are most kind, but it is still a trade: I can choose to spend hours of searching myself, or I can wait hours for people to reply, hoping that they can indeed solve my problem. I have studied the “Inform for programmers” document, but I am afraid that it reads more like a list of idiosyncracies than an exposition of a few well-chosen general principles to me.
That is not to say that Inform is useless. It is indeed a marvelous tool for writing run-of-the-mill key-lock-puzzle IFs with the occasional twist or mechanical gimmick. And compared to previous tools, this is quite an achievement. What I want to say is that it strongly enforces a particular kind of game design and radical departures from that require extraordinary efforts. Some authors do have the stamina, but most do not. In this sense, Inform’s default will shape IF for years to come.
As for approaching Graham, I appreciate your offer, but my fundamental concern is this: I do not want to need to approach Graham, I want to be able to shape my imagination myself. In a sense, Graham has produced a tool with tremendous imagination, but it falls short of imagining the imagination of programmers.
*clacks LISP boots* I sure am picky. ;)
The phrasing of things like “The spider bites in five turns from now” is a bit of a peeve of mine — I think it encourages the programmer to think of “the spider bites” as a semantically complex phrase, when in fact it’s just the name of an event. I wish that these named events were set with the use of the verb occurs — “The spider bite occurs in five turns from now” — so it would be clearer that “the spider bite” was an indissoluble, unanalyzable name.
In this case, though, it’s probably right that setting “the spider bite” as an action isn’t the best way to do it; you might want to give the spider itself a countdown, which diminished every turn (or every turn when it’s alive), and have the event trigger when the countdown reaches zero and when other conditions are fulfilled (e.g., the spider is alive).
For the rest, though, it kind of seems as though you’re complaining that Inform 7 has a learning curve. I mean, I don’t think that I’ve ever learned a single programming language where I didn’t have some problem expressing myself in it before I learned my way around. Certainly I wouldn’t expect to be able to compile some stuff that a relative newb typed into a comment box off the top of his head, which is what my original code was.
For the rest, though, it kind of seems as though you’re complaining that Inform 7 has a learning curve. I mean, I don’t think that I’ve ever learned a single programming language where I didn’t have some problem expressing myself in it before I learned my way around.
It may sound like it, but I am complaining about the design that leads to the learning curve, not the learning curve itself.
For instance, is the following snippet legal?
Instead of opening the lid:
when Spider Crawling begins: the spider bites in five turns from now.
No, the expression “When Spider Crawling begins” is special notation for a rule associated to a scene, not a boolean value that is true or false. But there is no clean syntactic distinction between those to. The sentence “when Spider Crawling begins” is a proposition in english, but this is not the case in Inform. Instead, I have to memorize special verbs and syntax to be able to tell the syntactic role of an expression. At some point, there is no point in continuing to memorize idiosyncracies.
To cite Ron Newcomb at the end of his “Inform for Programmers”: It is a large language with many nooks and crannies to explore. Finding them is frequently a game in itself. I hope you find it as enjoyable as I have. My response would be: no, I do not find it at all enjoyable because it’s not conductive to programming.
From a game consumer’s point of view, I think what would help the most is curated collections and branding.
Definitely. One marketing obstacle for IF is that it simply can’t be marketed the same way that graphical games are typically marketed. Very few IF games will benefit from screen shots or video trailers (although, with the recent vogue in video trailers for books, maybe I ought to rethink that option.) From a consumer perspective, it can be difficult to decide which games are most likely to suit my tastes. The IF community has long had a mode of curation — mostly through reviews on SPAG or BAF’s Guide, and in various discussion groups — but I think it would help to start making such curation an element of how IF is packaged. There was some discussion of one possible approach to that over at Post Position, but there are obviously yet more avenues to be explored.
Obviously, if you read through the posts I’ve written on the matter (and thanks for the link!), the solution that’s most appealing to me is the creation of the IF equivalent to a literary journal, distributed through tablet and e-reader platforms like Kindle and iOS. But that’s clearly not the only possibility available to us. The important thing, I think, is that we start looking at publishing models other than the packaged game model that died with the old Infocom days. It may be that most IF is better off not being marketed as a stand-alone game at all.
Regrettably, there’s more to discuss than I can justify on a quick break at work, but I heartily concur regarding curation, the idea of a literary journal, and the need to repackage “finished” works of IF to make them attractive to modern users.
Presentation A. Briefly, and just to pick two recent examples, Andrew Plotkin’s “Set Sail…” and Jason McIntosh’s “The Warbler’s Nest” are essentially 15-minute folktales which would find a certain kinship in a collection, and would be valuable additions once “cleaned up.”
Presentation B. There’s been so much parser discussion recently, I don’t want to retread much. Grunk say parser not broken: Emily and Aaron teach Grunk use parser in game. Grunk happy. Yes, it can be tweaked, but it’s a great tool that didn’t bother my new-to-IF spouse, either, mainly because I started her on Bronze. (That should say enough for an unwritten paragraph or two.)
Presentation C. I like Courier New. I especially like it when it’s white on a blue background, or green on a black background. This makes zero sense, and it’s only the case as a nostalgic reaction to my childhood, playing on ancient CRTs or z-machines. We need to make IF look new and relevant. It doesn’t make new players nostalgic: it makes them think the game’s probably very fashionable amongst third-world Stalinist elites, who are definitely using the computer my Dad chucked in 1983.
Curation. Oh, “Games for Windows” magazine. However, there’s an awful lot to be said for picking something up and knowing it works, and having a brand associated with it to back it up. We do have the foundations in place, between Emily and Aaron’s (among many others’!) player-friendly extensions, and set-piece outlets for review as well as the IF blog-O-sphere, and voluntary beta-testers, as well as Emily’s trend I deeply sympathize with of not playing/reviewing games which have no testers to speak of. (That shouldn’t be everyone’s angle, all the time, but it makes sense for a competition.)
There are paid and unpaid avenues available, and it doesn’t necessarily have to be all one or the other, or necessarily start off the way it concludes.
My two cents worth in this fascinating dialogue is that the “killer app” for IF resides on handheld mobile devices, the vast majority of them not having a keyboard at all.
There is real money to make for the next skilled IF veteran who can implements a click-only IF gui that retains all the depth of a textual old fashioned typing-based one.
I’m not talking degenerating IF to a point-n-click adventure game, nor do I espouse dumbing it down to the level of a CYOA, but somewhere in between – a clickable dictionary of possible verbs in any given situation, while slightly spoiler-ish, would be a wonderful – and fun – feature that I’d like z-machine games to support by default without requiring autho coding to do so: some sort of state-dependent scan of all known verbs that would produce any form of response in the current room, and a means to use a touch screen interface.
Once IF can be played on the bus on your iphone with nothing more than your mind and a casual touch here and there, it will, IMHO, explode in popularity beyond any level seen before.
Keep up the amazing work Emily: you (and plotkin) are almost singlehandedly keeping the IF industry alive and I value all your blog posts and gamasutra articles.
Rather than trying to start up a “literary mag” for IF — getting into reader’s hands would still be a problem, and SPAG still needs something to focus on in the wake of IFDB — perhaps we could convince existing print mags to feature one IF game per issue? I think _Edge_ magazine has some staff that are warm to IF. (Or was it _Play_? One or the other.)
Ron, you’re not wrong about there being a bit of an existential revaluation going on at SPAG — a large part of SPAG’s mission statement has been providing IF reviews, which was in fact essential to the community when there was a higher barrier of entry on the net and fewer active sources. I love IFDB, but it is (and this is just an observation, not an emotional response) a competitor for exclusive content.
He or she who adapts lasts or least wins the dubious prize of obsolescence, and while I have no plans to discontinue seeking SPAG-specific reviews, the magazine’s distinctive nature and survival depend on wide-ranging community-generated, and community-relevant material. SPAG’s publication cycle, as just one element, makes it difficult to deal with timely, short-cycle issues — for example, last week I could have written an editorial on J. Blow and the evolution of adventure gaming, but not only would it not see the backlit glow of liquid crystals for another month and a half, but in no more than the next three days, there were already five or so well-written, cogent reactions which have spawned their own sub-threads. This post of Emily’s being a case-in-point.
So, SPAG at present is far better suited to a strategic rather than tactical approach. That doesn’t have to remain the case, but there are serious questions involved. Heck, one of the main points being discussed right now has to with IF’s aesthetic presentation — I think SPAG requires modernization without alienating or disrespecting its heritage.
This isn’t the right forum for a full discussion, but it is one I’ve been having with Jimmy for some time, and on which I’ve been spending considerable thought.
In terms of business model, I might argue the opposite. Instead of encouraging deeper, more expensive, more polished games that take a long time to create, it might make more sense for authors to write superficial, cheaper, and more bare-boned games that are released more often. Instead of marketing IF works as being part of a larger IF community, it might make more sense to market IF works as being the output of a certain author that just happen to be in IF form.
Although the gaming press likes talking about innovative, boundary-pushing works, the money-makers are sequels. The big studios make their money from endless sequels to the same games, and the casual games market survives on endless variations of the same resource-management/hidden-object games. People find some gameplay and a writing style that they like, and they want more. People like reading books from the same author, not from the same book publisher. People like watching movies from the same director but not from the same movie studio. If you are able to produce middling games with consistent tone, genre, and style, that are released often and cheaply, fewer people will come, but they’ll stay longer. To be honest, I thought Choice of Games was pursuing this business model because they had a style of game where an experienced writer could churn out a new game every few weeks, but they seemed to stop after two games.
“…and these things make a huge difference to user experience and perception. A lot of people, myself often included, would rather entirely skip adding certain features to their games than add ones that look painfully amateurish”
And therein lies at least one of the central problems of text adventures. To do a good game in various other venues, like first-person shooters or even graphically-based adventures (think Broken Sword), you have to have a certain degree of competence in drawing backgrounds and providing staged game play. Otherwise fans will immediately dislike your game and there are a slew of others to choose from. The barrier to entry for doing these kinds of games well is high, which means those who are willing to not look “painfully amateurish” are the ones that continue and the others are weeded out quickly.
Text adventure development doesn’t seem to have that impetus and the barrier to entry is pretty low. Just about anyone can do it. You don’t even have to be a good storyteller, which would seem to be one area where text adventures could shine over other game formats. It doesn’t necessarily matter if the interfaces are “painfully amateurish” because you’re still largely catering to an audience that is willing, first and foremost, to play a text adventure in the first place.
If you’re trying to win new audiences, then the bar has to be set quite a bit higher. There’s the interface. That’s one thing. But what about gameplay? How does a text adventure compete without one of the key elements that people respond to: visual stimuli? One reason games succeed is because they can provide a vibrant and dynamic world. They can combine sound, music, and artwork with gameplay to provide a total experience.
Think of a game like, say, Mass Effect — but done as a text adventure. You’d lose a ton of emotional and visceral impact without the audio and video. Or consider Bioshock, Half-Life 2, Fallout 3, Dragon Age, or whatever you like or respond to. You could even lump World of Warcraft in there or the upcoming Old Republic. I would maintain there’s zero chance for text adventures to ever again compete successfully in this arena. Unless the mechanics and interface change so much that you’re no longer really talking about text adventures any more. (Didn’t Infocom sort of start going down that path? Remember Mines of Titan? Didn’t Legend Entertainment start going down that path? Remember Companions of Xanth or Mission Critical?)
Text adventures, a far as I can tell, basically had to evolve themselves away from being text adventures. Even the Sierra games that relied on a parser eventually went away and then they just became adventure games. Text adventures have shown historically that they have to evolve so much they eventually are no longer themselves. This has happened even with the addition of sound, static pictures, animated pictures, even video.
I think that text adventures have proven they can survive, but they’re sort of an extremophile: they require very specific, and limited, conditions in which to exist.
No, I don’t think a parsed, text-output game is likely to compete for summer blockbuster credibility with Mass Effect or Half-Life 2. Nonetheless, there is strong and repeated evidence from a number of directions that some people are interested in the idea of a readable, changeable story that you can manipulate, perhaps even one with some puzzle challenges in it, but find that the parser or some other aspect of the experience keeps them from really getting into it. Those are the people I want to try to reach. “We can’t do the *same* effects as X” doesn’t mean we have nothing else of our own to offer.
“….there is strong and repeated evidence from a number of directions that some people are interested in the idea of a readable, changeable story that you can manipulate, perhaps even one with some puzzle challenges in it, but find that the parser or some other aspect of the experience keeps them from really getting into it.”
Right. So they like a changeable story with some puzzles. That describes most games. What they don’t want is a text adventure that has a parser. That’s a pretty major component of a text adventure, unless I don’t know what a text adventure is. That’s like saying they want a text adventure that isn’t a text adventure.
Which gets back into mechanics and the form of the game being another kind of game entirely. It seems like when text adventures moved too much away from text adventures you had a move from Zork –> King’s Quest 1 (Parser) –> King’s Quest V (point-and-click).
The one key distinguishing word in your text there is “readable.” That assumes people want to play a game that they almost entirely have to read. I’m not sure what the strong and repeated evidence you mention is. But let’s assume it exists. So people want something they read but without a parser mechanic. That’s a book. But then you said they want it to be changeable and with puzzles. Okay. That’s a game like many of those that exist out there right now that aren’t text adventures.
Mass Effect and Dragon Age have subtitles that you can turn on. Does that count as “readable” in this context? What about games like Tunguska, Syberia, Broken Sword, and so on. Let’s assume all of the dialogue and textual description is done via text on the screen and not via spoken audio. Does that count as “readable” in this context?
Again, it seems like text adventures that try to be relevant seem to be moving away from being text adventures and being something else entirely. Which is a path the game industry seems to reflect as well.
So when you say “readable” how much of the gaming experience with these text adventures is purely readable, versus other aspects (video, pictures, audio)? For example, where does Ren’Py fall on this? (Those are described as “visual novels.”) It sounds like you have to settle on at least some levels on the continuum because otherwise it’s unclear what kind of experience is being discussed.
It seems to me gamers these days want something that can be experienced with various stimuli; not just something that’s readable.
You also mention that “some other aspect of the experience” keeps people from enjoying text adventures. It seems like it would be good to find out what that aspect is. You might find it’s the fact that they have to read reams of text. For me, playing text adventures gets me too much into the mode of reading a book but then forcing me to interact with the book as if it’s a game. A book doesn’t force that on me. And a game in the mold of Mass Effect (or any other ones) doesn’t try to overlay the experience as if it was a book. Instead the games provide an aural and visual component. Even Pong and Space Invaders way back in the day gave us that much.
To be clear, what I’m interested in is, in fact, keeping parsed input. I agree that without parsed input, IF is no longer the same form entirely. Similarly, I am interested in the possibility of illustrations, etc., but my own preference would run to using those as addenda to the text — specifically, that graphics do not become the primary way in which information about the world model is communicated to the player. Once graphics become the primary way to communicate to the player, they should also become the primary context for input (by letting the player click).
Those are the things I think we need to keep in order to say that the results are still the medium people here generally refer to as interactive fiction. (Calling them “text adventures” contains an implicit argument in itself. I’m not interested in defending the text adventure as such.)
The one key distinguishing word in your text there is “readable.” That assumes people want to play a game that they almost entirely have to read.
You’re assuming we’re talking about gamers here to start with. I’m not, or not exclusively: there are people who are primarily readers rather than gamers who have expressed an interest in the idea of interactive fiction, even including being okay with the idea of typing, but who founder on the current implementation of the parser. (People who have done some user research into the interests and problems of prospective IF players include Jeff Nyman, Aaron Reed, and Peter Nepstad.) There are also indie game players interested in the storytelling possibilities of IF, e.g. commenters on Jay Is Games and Play This Thing!, who have expressed a need for more accessibility in the works they would otherwise be interested in playing.
So I’m not talking about an audience of millions that we might gain by completely replacing our front end with something glittery. I’m talking about an audience of perhaps thousands or tens of thousands who would like IF if it didn’t present so many barriers to entry, and who might be reached if they could try in their browser a smoothly-loading, attractively typeset IF work which gave them a lot more guidance about how the world model and parser function. In fact, a number of them are already interested in IF in concept, but find it too frustrating in practice.
I’m sure there are people, including indeed a large proportion of hardcore gamers, who are not interested in reading reams of text. Fine. That’s not the audience I’m looking to reach.