Opening Remarks
At the time of this writing, I have released three games. The first, Repeat the Ending, made its debut just two years ago. My second and third games were released just over a month ago as “Back Garden” entries in Spring Thing 2025. The most remarked-upon difference between Main Festival and Back Garden entries is that the latter cannot win prizes or “Best in Show” ribbons. In the long term, though, the more impactful difference is that fewer people review them, which I suppose means that Back Garden entries usually get fewer IFDB ratings than Main Festival games. Meanwhile, aggregators favor games with a higher ratings count. If you have discovered Marbles, D, and the Sinister Spotlight or Portrait with Wolf at some future date, you have likely ignored their places in searches sorted by rating. Thank you for finding my work, and for possibly rebuking a bit of math along the way.
Some of you know that I do not explicate my own work. You will not catch me saying, for instance, “I’m disappointed that nobody picked up on Marbles, D, and the Sinister Spotlight’s theme of infant mortality in the global south.” That is not to say that I do not intend for my work to mean something. I write games, not nihilist riddles. Rather, I respect you—my audience—too much to look over your shoulder, hectoring and correcting. Your experience with my work belongs to you, not me.
This I do out of respect. I hope you will receive the respect I freely offer here. If I were to hand out answers, I would make students of you and a proctor of myself.
On Beginners and Beginnings
Repeat the Ending was my first game and, ever since its success, I’ve frequently thought of my fellow beginners. I was a beginning author, after all. In our conversations as a community, beginners come up often, usually in the very narrow sense of one question: “What is a good parser game for beginners?” I think that one common sort of answer is “well-made game that is a decade (or more) old.” I apologize for any perceived sacrilege, but I don’t think that’s a good answer, even if that game is “[game with unassailable place in IFDB ranking]” or else one highly rated on a “for beginner’s” poll. Here is a short list of factors that I consider while making recommendations:
-
Recency. Now I know, twenty years is recent to some of us, but I think anything older than a decade is too old. Why? Here’s an offer I’d like to make: “Interested in getting into parser games? Here’s what today’s authors are getting up to.” Better still, I’d like to recommend recent games by authors active in community spaces. Inviting people into a community first requires an active and available community.
-
Mindful. What are the problems or challenges facing players new to parser games? What’s keeping them out? The answer isn’t an absence of well-regarded games. We’ve been telling newcomers to play “[well-made game]” for years. Has doing so moved the needle? I’d like to consider the actual impediments. We can look at Emily Short’s “So, Do We Need this Parser Thing, Anyway?” as a start. More recently, the Interactive Fiction Technology Foundation’s “Accessibility Report (2019)” reinforces some of Short’s ideas while covering new ground.
-
Inviting. In the time since Short’s essay, the default terminal interface, customizable within an interpreter application, can now be experienced in a web browser. This is progress, but the truth is that these interfaces in their default states can be intimidating for players who, unlike so many of us, didn’t grow up with microcomputers and terminal screens. If I’m recommending a game to a player new to parser interfaces, it will be one with a styled web page. In fact, I would likely only recommend a game with images.
-
Transparent. I believe that the ideal beginner’s game also features public source code. Why? I would prefer to offer newcomers a comprehensive answer to the question “How should I begin?” Beginning authors can play along with the code, and beginning players can, should they wish to learn more about programming, refer to it. Obviously, the code should be well-commented or else be discussed elsewhere (blogs, wikis, etc).
Marbles, D, and the Sinister Spotlight and Portrait with Wolf were made with these goals in mind. In keeping with this evolving craft philosophy, I have recently updated Repeat the Ending to feature web styling and scaling images.
Marbles, D, and the Sinister Spotlight : The Cats of Our Lives
>examine me
Black fur, white locket.
Agile, brave.
Likes: fish, carpet.
Dislikes: loud noises, adult humans.
Adventurer and friend.
It was I, Marbles.
The game that eventually became Marbles, D, and the Sinister Spotlight started as a bit of six-hour speed IF to be entered as a La Petite Mort in the 2023 Ectocomp event. That started with a bit of curiosity about two Inform features: scenes and relations. I knew nothing about either and thought it might be fun to learn about both while making something short and wholesome. After Repeat the Ending, an intensely personal work, I needed a break!
I had recently started a blog about learning Inform. Since I had experienced trouble with the Inform documentation, I thought that other beginners might benefit from following my adventures as a beginning author. Sometimes, it’s better to learn from a peer, isn’t it? I remember, when I was just starting out, that I always worried about sounding stupid. There’s no risk of that with me! Researching scenes and relations would make for great material at the blog. They’re less-frequently discussed, and relations often come across as intimidating.
The La Petite Mort idea was to create a small space that changed as the story progressed. I had in mind a sort of lighthearted Shade (Andrew Plotkin 2000). There would be a scary (in a children’s literature sense) creature that the main characters would learn about and ultimately try to help. My ideas about scenes and relations went too well. I had a space that varied and evolved as the story progressed, and a sidekick character (a young version of D from Repeat the Ending) would remark on the current state of the adventure based on whatever interested him at the time. His quips printed according to a relation (“the doors interest D”). I say that progress went too well because the complexity of the world and of D’s behavior soon made a speed-IF entry impossible. In fact, the deadline for an Ectocomp entry of any kind came and went.
That was fine, I thought. I could do a sort of live blog of the development of Marbles, D, and the Sinister Spotlight. That never happened. It was too hard to sync up project progress with the blog. More insurmountably, my depression got in the way. And so, another deadline shrank down to nothing in the rearview mirror. Progress with the blog stopped. Progress with my Gold Machine project stopped. I achieved, for a time, very little. When I began to feel better, four months had passed. Another Ectocomp approached. Would I try again? The answer is yes! But not with this work. I wanted to take another swing at a six-hour game. How did that go, you might wonder. You won’t wait long for an answer! The story of Portrait with Wolf will be told in the very next post!
Sticking with Marbles, D, and the Sinister Spotlight: I needed a break while testers engaged with the wild monstrosity that was and is Portrait with Wolf. That work is dark, isn’t it? It would be nice, I thought, to return to the story of Marbles, a likeably optimistic cat. And so it was! At that time—September, I think—I had a game without an ending. In fact, I recall that the main throughline might have been 2/3 complete. All the other extras and goodies that, to me, separate “finished” and “well-crafted” were not yet even imagined. There was a lot to do!
The idea came, and I have rued it many times, that it would be interesting to release both Marbles, D, and the Sinister Spotlight and Portrait with Wolf at the same time! To pull it off, I’d need a very productive seven months. MDSS and PWW are both short works, but they are meant to be deep. There’s a lot for players to see and do. Realizing my goals for both games would require code I didn’t yet know how to write. More dauntingly, a dual release would require that most precious of resources, testers. I decided to proceed, and, from September on, Marbles, D, and the Sinister Spotlight was developed as part of a set.
There was a synergy. If something wasn’t working in one game, I looked at the other. After working on the wolf, I realized that I hadn’t adequately committed to the voice of the cat. What was the main attraction of MDSS ? Playing as a cat, of course. There wasn’t enough cat! Everything that printed, from parser messages to inconsequential descriptions, needed to reflect Marbles’s personality. EVERYTHING! At that point, I knew that I had to revise every single printed text in the game. This was a big job for a project in which room descriptions had as many as five unique iterations based on scene or other values.
Because I found that I could only write these varying descriptions of things in small chunks—my prose became rote while trying to revise things multiple times in one sitting—I I began to alternate between programming tasks and writing. On the programming side, I realized that Marbles didn’t have many opportunities to behave like a cat. Eventually, Marbles, D, and the Sinister Spotlight would categorize eleven actions as “cat behavior.” Up until the endgame, most of those actions had unique responses when applied to just about anything. This proved to be extremely high effort. But a cat is what she does.
I thought: there should be a score and unlockable content, which led to a scoring system. That started as a simple tally of unique cat actions (excluding those required for finishing the game), but a tester suggested rewarding experiments with multiple action types. That proved tricky, as I didn’t have a clear idea of how to keep count of multiple types of actions in an efficient way. My first attempt caused noticeable slowdowns during play and occasionally caused memory errors! That was a career first for me.
My second attempt worked better, and that’s the one in the finished game. I think it could be more efficiently-programmed, but it works and tests without errors. If you are a new author reading that code, I want you to know: If I can make a game, messes and all, you can, too!
I added more beginner-friendly features, including a status line that lists available destinations and items of interest. Despite the game taking place in multiple “rooms,” Marbles can traverse them without the need for player navigation. If the player types “meow at the ladder” when marbles is not close to the ladder, she will travel to the ladder before meowing at it. The story can be completed without directional inputs. Since I’ve often hear that new players struggle with the grid-and-compass geography of parser games, I thought that I could take it out of the picture altogether. The game is completely beatable just using examine, meow, jump to , go to , and smell commands. Players who wish to experiment can try the full range of cat behaviors in a low-stakes, rewarding way that will not hinder the story.
There’s a sweetness about this game that I love. Making a game about cats and friendship with a Zork flavor—a sentimental series from my youth—I think it’s nice. I enjoyed the subject matter, the mood, and, yes, the unambiguous joy of the ending, and I hope you do, too. Not that some darkness does not lurk. Some of us know D, after all. When I play this game and think about Marbles, I feel a lot of tenderness for its main characters, and for Gruecilla, too. It’s a good ache that I feel. Marbles has a lot of heart; she is a good cat and a good friend.
I would call this project a success! I’m glad that some people enjoyed it. I do have a request for future players or reviewers: if you enjoy the artwork or pictures, please say so. There hasn’t been much feedback about either. I’d like to pass on any positive comments to Callie (the artist), and, so far as the pictures go, who doesn’t like hearing that their cats are cute? Everybody wants that, it’s just a fact of the universe.
Portrait with Wolf: Ceci n’est pas un loup
>examine me
You look fine, I guess. In a good way!
Another year, another Ectocomp. I had been working on some posts about highly variable texts at Top Expert. I hoped to use these techniques in a kind of fortune-telling toy. A player could choose a few “cards” and receive a fortune based on the type and order of cards drawn. There were a few other conversations going on in Interactive Fiction groups at the time, and those began to color my ideas about the fortune telling gizmo. A short list:
-
A revival of the IF Art Show from the aughts. There were some discussions about bringing back the IF Art Show, a nontraditional event featuring artistic works. Those discussions became contentious, so I stopped following them. If I’m honest, though, when someone says “IF Art Show,” I hear Galatea. I don’t know if I’d fit in at an art show or not, but I would like to make something inspired by Galatea. I left those conversations wondering, “What would I do, and how would I do it?”
-
“Low Effort Games.” I recall a conversation about some minimum playtime threshold for rating IF Comp entries higher than five out of ten. The idea, I think, was that short works were “low effort.” Therefore, ergo, axiomatically, works with low play times warrant mediocre ratings at best. It’s science! You can’t argue with science! Whenever I hear things like that, I immediately assume that the speaker doesn’t know much about writing or making games. Or poetry. I also think, because I’m perverse, that I’d like to make some short, high-effort games. Why? I wasn’t out to court anyone’s IF Competition votes, that’s for sure. I suppose as a critic I reject attempts to formulate with simple math what is obviously qualitative. What would my design for a short, high-effort, good game be?
-
Following up Repeat the Ending. I had released a well-regarded game. How should I follow it up? I felt the sophomore’s dread, if I’m honest. I didn’t want to disappoint anyone.
-
Poetry. I saw some conversations about making poetry with Inform. I wondered: what would parser poetry look like? I never saw any models or examples for their poetry, but I was curious about the subject. My educational background is poetry. I have a graduate degree in writing it. I wondered if it would be interesting to come up with a parser-inspired poetic form. In terms of meaning and effect, what could such a form accomplish?
This all eventually came together as Portrait with Wolf, a project that uses the visual presentation of a parser game, complete with banner, room description, action/response, and “epitaph” as both poetic form and visual rhetoric.
There were a few iterations. Testers, as always, helped get me closer and closer to my goal. This project would have either never completed or else turned out miserably without them. In the creative writing world, we call them “readers”—an apt term, here. I spent so much early effort on the visual presentation and variations of the text that I had not developed the right voice. Everything was “talky” rather than poetic. It felt like a proof of concept. There were no real stakes. It was kind of a silly gotcha over and over again: “surprise, it’s a wolf.”
At this point, I knew that I had a strong form and design but the sentence-to-sentence writing was unacceptable. I went on to miss deadlines and thresholds for both Ectocomp events, and then IF Comp. The next attractive release option for me was Spring Thing. Could I release two games at the same time? I decided that I should make the attempt. They could provide contrast for one another, innocence (MDSS) and… whatever this was turning out to be. All signs pointed to vague, groping, associative.
I realized that the text of Portrait with Wolf, for all its poetic ambitions, contained mostly prose. I think I was worried about the “conventional” parser audience. I know that lots of people like Repeat the Ending, but would they embrace something more ambiguous and less linear? These concerns precipitated bland and unaffecting prose, and a lot of it, too! It was time for desperate measures. I exploded the text into multiple tables (150 in total), arranged them in thirteen extensions, and looked at everything all at once. It’s important to note that, in cases where multiple texts could print at one time (cat vs turnip), I worked to make every possibility fit with every other possibility. This is not a random work. There is no randomness in Portrait with Wolf. Everything has been chosen and placed deliberately.
The result felt right. Fragments, a stronger tone. Occasional references to or borrowing from canon poets. I got the “volta”—the last two portraits before the ending—working the way I wanted it in terms of both text and code. Next came postgame content. Testers wanted more but felt that the ending was the ending. There should not be, we agreed, a “secret” ending that elaborated on the fate of Protagonist and Narrator. While that was settled, two questions remained: what should the content be, and how should it become available?
I decided that a total number of unique portraits might be the best way to go. The problem was that I had no idea how to do that! If I did choose that method, I felt I would also need to provide a way for players to know what portraits had been rendered already. That became a rather Big Deal, with things going into lists, and texts of the things going to other lists, and so forth. The player output looks like this, with a new row for every unique portrait.
Portrait 1 (Series I, Default):
| C | C | T | B | C | A |
Once the count of unique portraits reached fifteen, two additional sections of the “guidebook” are unlocked: a collection of poetic scraps and Callie Smith’s drawings of all motifs. For the scraps, I tried to frame the speaker differently. Whether it provides a richer sense of who they are is for readers to say. I particularly enjoy the Goya one. I think Callie’s art is terrific. I especially appreciate its physicality. The cards are imperfectly proportioned and bent, and the hand-stitching at the borders adds a rustic craft flavor to their presentation. I hope you enjoy them as much as I do!
How do I, the author, feel about this work? I tried to express something hard to get at directly, and I hope that I succeeded. Once everything came together, I felt that I had done something new and possibly interesting with the parser form. I don’t think I’ve seen anything like Portrait with Wolf. That isn’t to say it’s a gimmick. There’s a lot at stake here. I am mainly interested in works where something is at stake, whether I am reading or writing. It is not easy to put oneself into something and to experience the vulnerability that follows. It is also not easy, as a reader, to attempt empathy.
I hope some of us have met or will meet at the far end of Portrait with Wolf with varied but valuable experiences: understanding or being understood; solidarity; worry in its many senses; recognition; relief at our mutual survival. Has there been a wolf in your life? I see you, and I hope that you see me, too.
Appendix: Parser Poetics
What would I call this new form? In a humanities or creative writing context, I would call it a “parser.” That is accurate, for one thing, and I also think the term might intrigue other readers or writers. Among an audience of parser game players, however, that name would likely confuse. Sadly, I have no better idea for the moment, so for the duration of this post any reference to a “parser” is, in fact, a reference to the poetic form used in Portrait with Wolf.
As I imagine it, a series is a long poem with six parts. Alternately, we could consider each part its own poem, and the series a collection. One framing emphasizes interrelation between the parts, while the other implies a looser and more associative relationship between elements. I will stick with “part” for now.
As I have said, the parser is intended to call to mind the visual representation of traditional adventure games. While this rather rigid structure is obeyed throughout Portrait with Wolf, I never found its constraints excessive. Let’s walk through the elements:
Still Life with Wolf ^-^
A good time by A Loving Friend
Release II / Serial number 12345 / Inform 7 v10.x / D
In this implementation, the title (“Still Life with Wolf ^_^”), headline (“A good time”), author (“A Loving Friend”), release number (“II”), serial number ("12345”), and compiler letter (“D”) are all dictated by the author. This construction matches the shape and content of an Inform game’s “banner text.”
This is a surprisingly rich format. In my own poetic style, I tend to favor repetition and echo, and I believe the banner text invites such tactics.
Next comes a brief prefatory statement, typically in the form of a prose sentence:
Press the “enter” key or choose a letter to make your “art.”
In my own use, prefatory statements tended to be abrupt and rather literal.
A room description and room name follow. Both vary according to the current series. It is important to note that the room name is printed in the “status line” at the top of the screen, which grants it a place of privilege.
Wolf Gallery (unoccupied)
It’s a cool morning in late March, and the air through the windows is sun-drunk and brisk. The floors smell like wood soap, shining dully in the indirect light. Here, the real vanishes into the possible, each day longer and warmer than the one before.
The room description is generally prose, though it need not be. In keeping with its inspirations, Portrait with Wolf uses it to characterize a setting. The author, of course, has a lot of freedom to decide what “setting” means within the greater work, but in my own project the room descriptions usually shrank as the game progressed.
The easel is next:
This easel certainly looks safe and stable.
In Inform terms, the “easel” is a supporter or container. It holds the motifs. In Portrait with Wolf, I treated it as a way to frame the player’s choices. By Series V, the text has moved on to abstractions like this:
Flat, stilted.
The choices themselves are a list. This is in keeping with traditional adventure games: the contents of supporters and containers are typically listed while printing room descriptions. I chose to arrange them vertically with clues as to which motif can be chosen with which keystroke:
( C ) calcium tide
( T ) turnip in decline
( B ) savory boot broth (times incorporated: 1)
( A ) cloud of space trash
These lists offer a lot of opportunities for poets. They can be as literal or as lyrical as the writer desires. I used them to develop the character of the motifs. I also enjoyed creating and removing semantic “space” between the choices and the easel. Sometimes they seem aligned and sometimes they don’t; tuning these relationships is a way to foster experiences of discord or unity.
We now have the command prompt, which is naturally important.
>
A poet must take advantage of every bit of real estate on the page, so we cannot neglect the prompt. In Portrait with Wolf, only four commands move the work forward. However, each of Inform’s built-in action and error responses have been customized. Non-productive commands can create a kind of “negative” agency. I found this a rewarding idea to pursue.
In my own poetic writing, I tend to enjoy dramatic shifts in content and presentation. I found the next section—the action response—a venue for indulging these interests. While different authors must decide for themselves, I felt that the “pause” created by the prompt created a sort of silence that action feedback could break dramatically. Here is a prose example from Series V:
They say that the closer you get to a black hole, the slower you move. You never get there, in fact, spending the life span of the universe falling into nowhere. However, you also get there immediately, the forces jerking you, helmet first, into the epistemological outlands.
The end of a journey is its own reward.
Here, a more lyric example from Series VI:
Push down
push
the dirt
down
While writing Portrait with Wolf, I felt freest writing action responses. Because of the rhetorical framework of a “game,” I often used them to respond directly to the protagonist’s actions. Sometimes these responses were immediate, and at other times they spoke to their past. I would say that they lean in to the “you” of the second-person voice.
At this point, this game “ends.” In classic parser games, this conclusion appears as an “epitaph.”
***** Who Is Winning the War on Hunger *****
I tended to use the epitaph as an abrupt turn, poetically. Since the message originates as a death notification, gallows humor often feels appropriate.
In the first five parts of the six-part series, the epitaph is followed by a prompt to “restart.”
Want to try again? Press any key to restart!
Initially, these are directly about restarting the game, but the instructions grow less explicit.
Let’s just finish this first.
Once the pattern is established, it’s possible to discuss other things.
Are you going to eat all that?
At the conclusion of the six-part series, a more fully developed ending based on the combination of parts (portraits) drafted. This consists of an ending-specific text that precedes the epitaph as well as a spoof of the “final question” dialog that appears when a parser game has ended. These two factors together contribute to a sense of punctuation missing between the initial five parts. These tend to be longer texts that synthesize the content of the part as well as the overall advancement of the work.
Mining bees hide in the cracked dirt, waiting for dawn. Somehow, you know they are friends.
It is better to see the stinger coming.
The turnip is sweet anaphylaxis, the mark of the fortunately stung.
In Portrait with Wolf there are, among all the possibilities, 32 endings to be found, and many more underlying combinations.
What does this form offer a writer? The parser form facilitates the creation of compelling patterns and rhythms by leveraging an established rhetorical structure: that of the classic 1980s text adventure game. I enjoyed the way in which the differently-structured stanzas afforded opportunities for irony, surprise, and modulation of voice. The practice of using deterministic output as opposed to randomly-generated text grants the player a sense of agency without necessitating that an author cede control of the work.
Since it is inherently cyclical, the parser form is also well suited to using a single concept, text, or thing as a poetic touchstone to which the author can revisit throughout the work.
While I can’t say whether fans of classic adventure games are interested in poetry, I can say that the formal structure of those games can readily accommodate it. Inform proved itself more than capable of managing the various texts and printing them as I required. While the absence of tab stops is a rather severe limitation, Inform otherwise proved itself a capable—if not affable—companion.
I’ve made short code templates for persons wishing to make a four-part version of the “parser.” A beginning author can add text by filling in tables. One template is for a game completely playable via hyperlinks!
Wandering Off
These projects, combined, were a lot of work.
- Marbles, D, and the Sinister Spotlight: 600 hours, 58k words.
- Portrait with Wolf: 450 hours, 40k words.
For comparison’s sake, Repeat the Ending is 150k words. I think it might have taken 1,000 hours to write. I’m not sure. Nevertheless, this was a big job. I think each game can be completed in 30 minutes, though engaging with optional content will take longer. I would not personally call these 30-minute games “low effort”, but who knows how easily someone else could have made them?
A recurring question I ask myself (and ask the readers of my blog) is “how far do I want to take this?” I’ll bet I spent 50 hours on disambiguation and scope for Marbles, D, and the Sinister Spotlight. That wasn’t necessary. I wouldn’t have hurt anyone to repeatedly see parser clarification messages. But why should they? If my goal is to make a beginner friendly game, I need to polish away unhelpful output and undesirable parser behavior. I need custom failure messages for unproductive actions so that I can point players in a better direction.
How far do I want to go? I can write custom messages for every built-in verb that is never used, or I can write one general message for every one of them. What is the difference? Will anyone even notice? Not everyone will, no, but I think someone will. The same goes for errors printed by the parser. Here is Inform’s default error when a “noun” is referred to that is either not present or nonexistent:
>get frob
You can’t see any such thing.
Here is the output from the same parser error in Portrait with Wolf:
>get frob
Oh, I’m sorry, did you hear me say “frob”? The fragments of our work are memory and thought, and you cannot act upon them. Their pangs are delicious, aren’t they?[to review a list of commonly-used commands, type HELP]
There is different output if the player uses an article in their command. Now, in this case, this solution feels warranted because Portrait with Wolf refers to many things that aren’t present. But again, how far should one take such things? This isn’t a low-effort thing to do, especially not for someone who, like me, doesn’t understand regular expressions.
Were so many unique descriptions required in Marbles, D, and the Sinister Spotlight? The answer is definitely no. It always depends upon how far one wants to take things. I feel that I grew a lot as a programmer while making these games, and I’m proud of that. I’m also proud that I seem to have met my own rather unfair standards. As a poetry MFA, I think: Portrait with Wolf is good poetry. I think Inform is a good tool for writing poetry, too. So far as Marbles, D, and the Sinister Spotlight goes, I feel I captured the youthful feeling of adventure that I experienced with Zork.
Both games were released at the same time, and that has been my intent for many months. This proved to be exhausting and grueling, and there was a fairly unpleasant amount of deadline crunch at the end. Still, I really do feel that these works belong together. I wonder if that will ever come up in critical conversation? Sometimes, I think our discourse is both siloed and atomized, since reviews of individual games are the most common form of discourse in this community. It’s always exciting to read essays or posts about multiple games or about an author’s work in total. I am not fishing, by the way! Don’t start with my work! Still, I think it would be interesting to see more critical or craft writing that cannot not fit within the confines of an IFDB review or an Intfiction review thread.
I do not except myself! I should write such things, I know. I probably could, If I wasn’t always busy with… parser errors.
What else? If you play and enjoy works by new authors, please consider rating them. Besides the “what do I recommend to new players” question, there is also the matter of “how do we retain new authors?”
Thanks to everyone who engaged thoughtfully with my work! It means a lot to me. I’ve released fixed-up versions of both games, complete with source code. You can find them below. I hope to see everyone next year with a main festival entry.
IF is for everyone 
Drew
Post-festival releases and source code:
thank you