My game Never Gives Up Her Dead wasn’t really placed in any competitions, so scheduling a post mortem was kind of difficult. So I’m doing it now since it’s been out long enough that many people have finished it!
The game itself contains extensive notes on its creation, so this article may be unnecessary, but time will tell.
Also,
This post mortem has a lot of spoilers, which will reveal everything about the game starting in the next few paragraphs
Origins
Like several of my games, this one started as an idea that I sent in an email to an IF friend that I hadn’t been in contact with for a while. I like to message people who I trust and who are also ‘out of the loop’ to get their reaction. In this case, it was Mike Spivey. Here was my original message on July 14, 2022:
Hey, can I bounce a game idea of you? It’s been stewing in my brain a bit, but you’re the first person I’ve talked about details with (major spoilers below):
I’d like to make a game either [should be ‘with’] these ideas:
set on a broken, empty and derelict spaceship that’s encountered a singularity which has caused little tears in space everywhere that lead to very different dimensions
Each dimension has mysterious helpful messages and items from an unknown benefactor that seems to know you.
You eventually realize the other dimensions are actually an alternate version of the spaceship you are on, but one filled with tons of museums, nature preserves, and other stuff; it’s a seed ship with frozen humans, and is designed to give them a good life when they wake up
The final realization is that the other ship is actually in the future. You recorded everything in the past and passed your journal down to your descendants, one of which is the captain of the new ship. All the messages were left by your last self and you helped create the future utopia.
Each sub dimension would be about as big as Impossible Stairs. I’d plan on taking a couple of years to work on it, aiming for Hadean Lands or Muldoon Legacy size gameplay.
Do you think the basic concept could work? Or is “abandoned spaceship” and “portal travel” overplayed? Because I have other ideas!
With his positive response, I decided to start building the game.
I went for a modular approach, inspired by @McTavish’s game Dr. Horrors House of Terrors, which was split into little studios, each with their own significant puzzles. I wanted to go for a big game because I had won Parsercomp and Ectocomp and was on the committee for IFComp so I was running out of places to submit games. And, of course, while writing I started running Spring Thing, so I no longer had many outlets. So making a modular game, I could really make 10 separate IFComp-size games and that would keep me occupied for a while.
My other impetus was that my ‘10+ hour’ category on IFDB reviews is almost entirely full of old games from the 80s and 90s, and I wanted to make a game specifically to recreate that feel I had when I first played Mulldoon Legacy or Curses or Not An Ordinary Ballerina: being lost in a giant game that felt like it could go on forever, sometimes not even finishing. I figured that the audience might just be 20 hardcore parser players, but that would be fine with me.
Originally, I was going to use a Viking idea, with Norse names, with earning an honorable death being the concept (you encourage each crewmate to overcome their fear of death by showing they’re honored in the future). It later became Welsh-based, because Tragwyddol is a bit easier to say than Eifilgir or whatever the original name was.
Overall Design
My main design idea was to have 10 separate dimensions, each corresponding to one crewmate. The dimensions would be disparate in gameplay and setting. There would also be a ‘hub’ world in the present.
My original brainstorming included a wax museum, a forest, a ‘toolroom’ where you got a sonic screwdriver-type tool and upgraded it over the game (including unlocking doors), a western saloon with robots, a ‘hall of squids’ with music, and a big gravitational puzzle set in a starry sky.
Eventually, though, I settled down on 10 ideas: a haunted house with classic gameplay, a murder mystery, a wax museum escape room, a spells-based area like Enchanter, a combat area like Gun Mute, a garden with plant-based puzzles like Starry Seeksorrow, a zoo with animal-based puzzles, an area where copies of world monuments floated in a dodecahedron in a starry sky, a tool room as described above, and a ‘wildcard’.
I’ll describe these in the order I made them.
Haunted House
I wanted this area to be classic parser gameplay, set in an area with classic movie monsters. One of the first things I wrote in the actual game, in September 2022, was this joke:
The TargetResponse of Franken-name is “The great creature sighs in response. ‘That’s a common misconception. In fact, [italic type]I[roman type] am Frankenstein, and my creator is technically known as Frankenstein’s Scientist.’”
My inspiration here was @Warrigal. Garry Francis’s games are what I consider the platonic ideal of standard IF gameplay, so I set out to make a game like his. It ended up not really resembling his oeuvre, but that was the original concept.
I find that I am more creative the more constraints I have, so I also gave myself the goal of associating each monster with a different emotional issue. Dracula, pestered by bats, is anxiety; the creature of the black lagoon can’t see the stars and remains lethargic from depresssion; the yeti is anger; Frankenstein is greedy; and so on.
This was my first world so a lot of work went into fixing standard responses, setting up the basic game structure, and customizing the dialogue system. The working name was ‘Tears in Space’, a pun on the rips in the fabric of space and the tears shed by people dying in space.
I wrote out a chunk of gameplay, and sent it to testers, who reported it took about an hour to play. I left in blank connections to other areas, like a locked trapdoor that goes nowhere, so I could hook it up later.
I was aiming for a two-hour gameplay per section, so I made a whole new second area. I wanted everything to be explained by science, eventually; there’s no magic in this game. So I made the monsters all holograms, and explained it as a program made by the past ship’s psychologist.
My idea for the second area of the Haunted House was an unfinished adventure game made by the psychologist; you’d have to pick up items and set up puzzles and traps for an adventurer who would come and solve the puzzles, like Adam Cadre’s Lock and Key. This eventually morphed into a ruined castle where a clone follow you around, copying your moves from two actions ago (like Matt Weiner’s Faithful Companion), so you solve puzzles together. This took a ton of work rewriting all of the NPC code to report every action the NPC does; I made an extension to help others with that but never finished it or released it.
I separated the two halves of the haunted house with the river Styx and Charon, who has to be paid with an obol, and I put a machine to give you obols in a (blank) other dimension. This was the start of an idea–so many times in modular games I’ve seen players just bulldoze one area at a time without exploring everything. By making each dimensions depend on the others, it would force them to take a look around first. I also put a dark room at the end that you needed the flashlight for (making it have two dependencies, one of them being the tool
dimension), but eventually discarded that idea.
Player response was moderate, indicating that I was doing okay but not wowing anyone. I moved on to the next dimension.
Murder Mystery
This was originally going to be in a Victorian house, but since the haunted house already had that distinction, I had to think of a new place to put it. Why not, I thought, a train?
Now, the future was supposed to be mildly utopian, so an actual murder wouldn’t really fit once the big reveal happened. So I made it in the end a ‘dinner party’ murder mystery with a fake corpse, a fact that you don’t realize until the end.
My most played game of all time is Color the Truth, a murder mystery where you enter suspect’s memories with flashbacks, but, like Rashomon, everyone is lying/telling their own version. By combining clues, you can force people to admit facts, which changes the flashbacks. I later went on to implement the same ideas in Sherlock Indomitable.
I wanted to make this even bigger. If I had a giant murder mystery as just one tenth of a big game, that would make it a really really big game! So where Color the Truth had two versions of each flashback, this would have three, with the third version of each flashback being gated by an item from another dimension.
I based my characters on different experiences in my life; Maeve, an author, starts lying with the stereotypical view of authors having a glamorous life before devolving into the truth of authors often being very poor. Arthur, a professor who is desperate to keep his funding, is based on my many experiences with egos in academia. Elen and Mari were designed as red herrings, each being clearly culpable of something (outrageous lying and theft), but not the actual murder, with Mari being modeled on my experiences in New York and Philly (where the bike-riding teens can be pretty overwhelming).
This is by far the biggest chunk of the game code-wise, with around 40K lines of code; not huge for an Inform game, but big since the others are 20-30K each.
Wax Museum
The idea here was for a combination wax museum and escape room. The wax figures would be based on historical people and also on the seven deadly sins. I wanted to include a variety of people, but ended up having quite a few British monarchs (Bloody Mary as part of Wrath, Henry VIII as part of Gluttony, and King Arthur as part of envy, with reference to Mordred). I had a token ‘we need a historical figure from the future since the game is set in the future’ called Fishblade, based on a short Twitter microcraze where someone made fun of micro-TTRPGs by pretending there was one called ‘Fishblade’ and people made rules for it.
I was inspired in this section by Ryan Veeder’s big escape room puzzle in one of the Little Match Girl games. I tried to think of a variety of puzzles and wanted to have it different from the first area, so I included a lot of codes, riddles, etc. I learned a lot of Inform 7 from this, but it was difficult, especially making new kinds of values to create lots of climbing wall rocks.
Spells dimension
It was around this point that I solidified two main ideas.
The first is that the game needed a human NPC that would play opposite the player, a kind of antagonist/deuteragonist. I imagined having a flash of white and the player walking down a hallway, turning to the side and seeing Death, as a woman in white (as death is often a woman in Spanish stories like La Dama del Alba). This woman would be the one orchestrating everything, and would eventually ask you to die for the future. Thinking more (and having some religious influence) I thought the woman offering herself in your place could also work, because what if the player doesn’t feel like dying?
The second is that I needed way better Inform skills. That’s when I started my ‘Let’s play’ of the Inform 7 manuals. It was fun, and helped a lot, and over time the impetus for many areas was to try out the cool techniques I learned.
For this area, I tried to think of how spells could work, but I couldn’t get a good idea. Then I hit on two concepts:
-This could be a horror area, with areas themed around my favorite horror podcast, The Magnus Archives
-Since there is no magic in the game, everything could be nanobots with the player controlling an avatar, and they could die a lot with no harm.
This could let me make an area unlike my others. I don’t like player death/gameovers and I don’t like things like diagonal directions, so I included both in this area to make it unsettling.
I originally named each spell after the corresponding Welsh word, but I got feedback that it sounded goofy, so I changed to be file/dos names, with extra x’s added (like .exe). So ‘TELOX’ was ‘teleport.exe’, ‘mallox’ was ‘malleable.exe’, ‘virox’ was ‘virus.exe’, and so on.
Each area is based on one of the 13 or so Magnus Archives fears, including being alone, being buried alive, not being able to trust your senses, and so on. Originally, it was even more intense, but I asked @CMG and @AmandaB, two of the best horror writers I know, for advice, and both side to give the players some breathers and less scary areas so it would contrast more with the fear.
Combat dimension
I had always wanted a combat dimension, and had imagined maybe having an army or something. But in doing my ‘Let’s play’ of the Inform manuals, I realized that there was tons of code in Inform for directing NPCs to do stuff, the majority of which I had never seen used. I took two of Emily Short’s examples and made some robots to order around.
I realized that to do complex army combat, the player would need to learn both how combat works and how to command NPCs, so I split this area into a ‘funnel’ kind of thing where you can do two tutorials (one on solo combat and one on cooking with a robot) before being sent to a set of increasingly difficult combat pieces.
I like this part, because I have to think hard every time I replay, but it has had a mixed reception. As @JoshGrams has said, ‘[T]hat’s tedious and boring’, and the puzzles ‘were so boring I definitely didn’t want to have to do all that typing over for one mistake’. I realized that would be an issue, and gave them short names (originally Sam and Max, but now Dan and Max, given that Sam and Max is a different IP); however, this concept was a tedious one, and putting it in was a gamble. Having a giant tedious puzzle does fit in to my original idea, since Mulldoon Legacy and Original Zork also have some puzzles that require tedious work, and Curses has the robot mouse puzzle, which is similar.
Garden dimension
Originally, I wanted plant-based spells here, like Caleb Wilson’s Starry Seeksorrow, but I realized that I had been doing more and more outrageous areas as time went on, and wanted to dial it back.
This become a completely supernatural-free area. It’s a cabin in the woods that needs remodeling, and you gather natural plants and herbs and make recipes for things. I looked up a ton of details and got advice from others on what would make sense here. I learned quite a bit about oils and paper and such.
I implemented a lot of fancy background effects here; there is a time of day system that keeps track of the hours you spend painting and such, and triggers messages as time changes. In the night, the moon is out and shadows are cast over everything; during the day, there is a sprinkle of rain and some brighter times. Every room has different descriptions at every time of day. I knew that most players wouldn’t see that, since no one reads room descriptions twice, but I wanted to put in some extra effort for very observant players.
Zoo dimension
Honestly, this dimension is my weakest one. I finished it very quickly, it’s the smallest, and the puzzles are very basic: take animal A that does behavior B to area C to resolve problem D. A lot of it is based on classic puzzles, like the sheep/wolf/cabbage puzzle.
I based the area on this image:
and that’s really that. I had the most fun learning about different animals and their behavior, like deer shedding their velvet. I used the area to do a bit more storytelling about the ship itself.
Monuments area
It was at this time that I started my dev diary, which is amusing because I was 70% done by now.
One of my testers, Max, had said that a lot of my areas felt too similar: fetch a bunch of items to get one big item that solves a puzzle. So I decided to make this area completely different from the others.
I had also finished my Let’s Play of the Inform manual, and wanted to fit in all the material that no one uses from that. Specifically, the parts with advanced math like inverse hyperbolic tangent, the physics problems like spreading gas, the money stuff, and so on.
To continue my ‘this world is different’ idea, I organized it vertically, so you must descend, and made it linear instead of open like my other worlds. I added books for infodumps as well.
Instead of gating other puzzles directly by needing an object from another world, I made the math and logic puzzles hard and put all the guides and answer in the Library of Alexandria, which needed an object from another world to solve.
Unfortunately, almost every player instead decided to brute force those puzzles, only going to the library at the end. It worked out in the end, but it’s a good example of how ‘what the author intends’ and ‘what the player does’ are different.
This area uses the second most code, much of it borrowed from Emily Short’s examples.
The tool area
I had always imagined that the tool area would have one big puzzle area with limited parser capabilities that you returned to to upgrade a tool and use it on other areas. I contacted @dibianca to get ideas on limited parsers, but I had some trouble here.
I also had trouble coming up with what a tool could do. A flashlight, a lockpick, and…what?
I eventually realized two things:
1-I had never explained the rifts in the game that led between dimensions, and
2-There was no overarching storyline in the game
Since the overal game is completely nonlinear (any dimension besides horror can be started right away once the game opens up), I couldn’t feed the storyline in the main areas, but if each area had some interactions with the tool, I could put the storyline there.
So I made the tool have rift-related powers and set up little ‘outposts’ in each other dimension. I had intended something like this from the start, so it was genuinely thrilling to finally hook up areas that I had left blank a year earlier. I cried a bit when I did so.
The one-room limited parser stuff showed up by having one machine in a room that had levers and dials with the commands on it. Instead of simplifying my code, the limited command set forced me to use complex scene-based scenarios which were much more difficult to debug and write than my other areas, giving me newfound awe for Arthur DiBianca. This was really nasty stuff.
These puzzles were very complex, and testers reacted very negatively to them, because they’re so much at once. The puzzles also revealed lore explaining things that players would have seen while playing other areas, so beta testers that tried just this area (including my father) were getting hard puzzle after hard puzzle and tons of infodumps.
But I had faith that it would work in-game, where the mystery of the locked doors would intrigue players, and the distant locations would serve to dilute the heavy puzzles and lore dumps. And it worked! Players of the full game have almost universally expressed appreciation for this area and enjoyment in the lore dumps (although the time travel caramel apple puzzle is still causing problems to this day).
Realizing the puzzles couldn’t continue getting worse, I ended up making them disappear entirely in the last few sections and focusing on pure storyline.
Knowing that I couldn’t anticipate the order players would find the outposts, I made each one have two parts: a static part that was ‘behind the scenes’ of its host dimension, dropping bits of lore, and some movable parts containing the actual puzzles that were dynamic, appearing in each area exactly as soon as you found it.
Finale and beginning
My goal was to write a 300K word, 20 hr game with around 30K and two hours of gameplay words in ten dimensions, plus an ending. One of those dimensions was a ‘wildcard’ I had left to the end, but I decided to make the ending itself large.
I split this last 2-hour segment into 3 parts: an intro (which I had never created), a puzzleless party area based on some Emily Short code, and a finale.
Beginning
I always leave the beginning of a game to the end, because by then I can best write the characters and my experience is greater.
I wanted to put the player in a situation where there was clear social pressure on what to do, and which served as a way to introduce new parser players to the kind of puzzles in the game, and to drop some hints on what the game was about. I also wanted to use the line ‘This is the story of how you die’, which I stole directly from Dr Who, where Rose’s last episode starts with the line ‘This is the story of how I died’, and which I consider one of my favorite lines in fiction.
I created the social pressure by having you late for a party and forced to make a presentation on ship history. This let me infodump by having you wing the presentation (the way the protagonist does in The Art of Misdirection) which would teach players about the ship. I had you start in a dark closet to introduce light sources, keys, and containers. Each item you find is directly based on one of the areas of the ship (so the flashlight is the tool, the rubber spider is horror, etc.)
I also went back and added psychic flashes after completing each section of the game, connecting you to Arawn, your future counterpart, and advancing the story.
Party section
I wanted to reward players for playing this long, so I made a museum where you could infodump literally everything about this future world (I did this after a beta tester reached the end with wildly different ideas about things than I had intended). I then included a party with multiple areas in one room where people interacted with each other and you organically, based on Emily Short’s Emma example. This also gave me a chance to make your own death sadder, as you discover it right after connecting with everyone you know.
Finale
I always planned on having two variants of the finale, one where you sacrifice yourself and another where Arawn took your place. I realized that I could go even further, make a truly big game, and have a third, unrelated, ‘bad’ ending where you refuse to go through with the plan and undo all the good you did in the game.
The first two endings, I wanted to be a ‘showing what you learned’ part of the game, where every puzzle was based on a puzzle from an earlier dimension. Programming a robot, panning for gold, it all came together here. This was a bit confusing for play testers who only tried this area.
I set it in a nuclear reactor, and did tons of research. I found out that explosions were not common, while meltdowns were more realistic, and watched a lot of videos touring nuclear stations (MIT has a cool one).
The very ending is kind of anticlimactic. A lot of feedback has said that it wasn’t as satisfying as it could be. I’m not too inclined to change it; I barely remember any of the endings of big games I’ve played, like Curses! or Mulldoon legacy. I knew the big reveals earlier or just the dimensions themselves were what people would remember. So I didn’t change it. Also, I just don’t know what I would change it too, to be honest; you already get a chance to say goodbye to everyone you love, which I liked.
The other ending was always meant to be the ‘bad ending’, but I realized after testing that people don’t really like ‘bad endings’, which I should have known, because in Choicescript games I’ve learned that people don’t pick ‘mean’ choices very often and will restart if they get a bad ending.
So I changed it from ‘this is what happens if you mess up’ to ‘this is you defying fate and being free’. You replay through your original game map but everything’s messed up. I included a very complex rope example from Emily Short, and boy was that a mistake. Everyone says ‘stay away from implementing rope, it’s a nightmare’ and it absolutely was, slowing my game down to a crawl and causing dozens of bugs. I hope it works now.
Playtesters
So many, many people helped test this game. I can’t thank them all enough!
Three specifically provided much more testing than any others, playing every area multiple times (except one or two near the end): @johnnywz00, @SomeOne2, and @Jade. Several other people spent dozens of hours testing many areas, but these three helped me over and over and were basically in constant contact with me for a year. I couldn’t have done it without them!
Each area was designed and tested to work independently from the others. Only at the end were they all assembled, where @AERobert was the first tester to finish the whole game (and I named the captain after him in the end). Full credits are here:
This game was written by Brian Rushton/Mathbrush using Inform 7
(by Graham Nelson). Beta testers include Amanda Walker, John
Ziegler, Jade, Christopher Merriner, Patrick Mooney, Brett Witty,
Rovarsson, E. Joyce, Max Fog, Dee Cooke, Ian Greener, Chandler
Groover, Lance Cirone, Zed Lopez, Cody Gaisser, Alex Proudfoot,
Radioactive Crow, Doug Egan, Mike Spivey, Larry Horsfield, Dirk
Spivey, Arthur DiBianca, Hal Rushton, Hidnook, Charm Cochran,
Grueslayer, Lynnea Glasser, Wade Clarke, Onno Brouwer, Mel Jason,
Tabitha, Daniel Worm, Mike Russo, and Robert Eggleston, who was
also the first to beat the full game.Bugs were found in subsequent versions by Max Fog, Garrett O., and S
B Wiegner.
Reception
Like I said, I hoped at the beginning to make a game for the 20 or so hardcore parser players on IFDB that review everything. Marco Innocenti was one person I had in mind; Sean Shore, who also was the first person to mentor me and give me advice when I made my first game, was another. I also hoped to be nominated for a Best Game XYZZY award, something I’ve never achieved (actually winning is pretty random; in past years, the nominees all seemed equally good, so just being nominated was my goal).
And that’s basically what happened. It has 14 reviews on IFDB, and is one of the highest rated games of its year. It wasn’t a breakout hit, but it was never intended to be. I’ve thought about putting it on Steam, but ran into some hiccups.
It was (a distant) third place in the Best Game of the Year IFDB Awards, which feels to me equivalent to being nominated for a Best Game XYZZY, so I was happy about that!
It has more source code than Blue Lacuna, making it one of the largest Inform games of all time (behind only the erotic furry game Flexible Survival, that I know of) and one of the largest parser games in general, although John Ziegler, one of my most valiant testers, managed to make an even larger (quite a bit larger!) TADS game the same year, How Prince Quisborne the Feckless Shook His Title.
Surprisingly, the people who have played it and loved it have not been longtime IF community participants. The 14 people who have rated it on IFDB are all either newcomers outside of Jade and Zape. I know quite a few older players who have started the game, but either never finished it or were underwhelmed enough not to rate it. The first is just fine for me, as I wanted to recreate those games that were too big for me to ever get to finishing, but the second is a little more worrisome. I wonder if the fact that I borrowed so many elements from other games means that the game is only fun for newcomers, as it’s all exciting and fresh, whereas old hands find themselves retreading old paths. Or maybe people who play lots of IF simply don’t have time for big games. I do think this has kept up its high rating though–only those who enjoy it finish it, and only those that finish it rate it. I know that Club Floyd was playing it at one point; I’m not sure if they finished it.
One review said, “Never Gives Up Her Dead captures the feeling of playing through massive adventures like Curses and Mulldoon Legacy while also skipping some of the potential frustrations that can come with revisiting classics now over twenty years old,” and that thrilled me, as that was my original goal.
Making a game ten times as big did not bring 10 times the attention. I never intended it to; while writing I always looked on my list of big games that were ignored and think about the travails of other authors of days gone by who made great work that has lain unregarded for too long. Any big work would put off potential players. Why did I do it? Because I appreciated the work of others that had done so for me before, especially Graham Nelson with Curses.
Either way, it was a long journey. The game became an obsession for the year I wrote it. Every day I worked on it for hours, interspersing it with talking with my son, playing games, or tutoring. I’ve never achieved such a good workflow before or after. It was a thrilling experience and greatly deepened my skills as a writer and programmer, and brought me into conversation with many good friends. It led me to read the manuals and to make the Bisquixe interpreter, and it was the point where I stopped relying on crutches for my writing and came into my own as an author; not that I’ve become great, but that I feel confident moving forward.
In any case, thanks for reading!
These are my design documents, one piece of paper for the game as a whole and then one piece for each sublevel. It unfortunately became blurry while scanning and I threw away the originals (c’est la vie).
designDoc-compressed.pdf (776.4 KB)
(note: there are a few random life-related notes in here, so if something makes no sense, that’s why!)
