You must be 18+ to visit this website
The content on this website is AGE RESTRICTED
Please confirm you are at least 18 years old of age. Otherwise leave the website.
Aridecan profile
Aridecan
18+
Aridecan
I write both erotic and mundane science fiction and fantasy. I also stream and develop video games.
Message

Subscription Tiers

$3
USD monthly
Supporter Tier

Enough to make me want to keep making more content and to keep you maintained. This is basically a tip jar for my work. Come join the discussion on Discord.

0 subscribers bronze supporter
Unlock
$4
USD monthly
Limited Essential Parts

This tier is an experiment. Set to 10 people only. Basically, this tier is to see if I open up the $5 tier. With the $5 tier open I'll be adding an index to the unpublished chapters that is only available to Subscribers of that level. Chapters will always be published publicly. This just gives you early access to the chapters.

This is a limited tier, so get it early.

Limited (0 out of 10) subscriptions
Silver Giants
Unlock

Features

The subscription gives you:
  • Access to Creator's profile content.
  • Ability to support your Creator by pledging – one-time or recurring.
  • Means to reaching out to the Creator directly via Instant Messenger.
Aridecan

February Index

Comments
Like(0)
Dislike(0)
Posted for $3, $4 tiers
Unlock
Aridecan
Public post

MGA Dev Report — October 2026

Ordinary Girl Meghan / Magical Girl Ava


Billing: Still Paused

Patreon billing is still paused, and it stays paused until further notice. Nothing about the reason has changed since August. It turns back on when there's a build worth putting on a hard drive, and there isn't one yet.

This month's report also comes in two posts on each platform. The text is the same in both. This is the members post, so it adds work-in-progress pictures of the character; the free post is text only. That split exists because of the pictures, not because of the billing. Nobody is being asked to pay to see them.


The Honest Number, Third Time

Source submissions in October: zero.

That makes three months running. Last month's report ended with a week off the day job "going straight into CoreX-M2." It did, but not into the engine. The critical path for M2 runs through the character, and the character is in Blender. So that's where the week went, and most of the month after it.

The pattern is the same as September: nothing changed in Perforce, and the mesh moved forward a lot.


The Prediction Held: 7.29

Last month's report found that the base body's foot was modelled in a 25-degree high-heel pose, which made the figure measure taller than it was. The fix was to flatten the foot before rigging and re-measure, and the prediction was about 7.30 heads.

The foot was flattened on 2 October. Re-measured from the side, heel to crown: 7.29.

That's mid-band for an adult anime figure (the target range is 7.0 to 7.5), and it's now written into the public character pipeline document along with the measurement protocol, so the number can be reproduced instead of trusted. Two other numbers came out of the same pass:

  • Crotch height is 0.518 of stature. Real adult women measure 0.47 to 0.48; stylised anime runs 0.50 to 0.53. So the legs are long on purpose, and still inside the style's normal range.
  • The foot is 23.75 cm at Meghan's height, about 8% small against stature. Its internal proportions (width against length, heel against length) all sit inside real adult ranges, so it was accepted as-is. One lesson worth keeping: a foot that measures about one head long looks too long in the viewport. The eye is the unreliable instrument here, not the mesh.
The whole cast now has scale factors. Every female character is this one mesh at a different uniform scale, from Tierney at 147 cm to Jennifer at about 192 cm, with Meghan at 170 cm as the canonical height. The factors get recomputed once the mesh is final, because every edit to the sole of the foot shifts them slightly. The current values are good to under 2 mm.


The Body Is Done, in Both Tiers

The body is finished and scaled to Meghan's 170 cm. The bigger news is how the two content tiers now live inside it.

MGA ships a Core game (the Steam build, with Twitch-safe coverage) and an optional Spicy tier that's never on Steam. Both use the same body. The way that works:

  • The full body is the base mesh. A shape key removes the parts that don't belong in Core.
  • One export script produces both. It exports the Spicy body first, then applies the removal key and exports the Core body. Same topology and vertex order either way, so skin weights, animation and clothing fit carry across without any extra work.
  • The textures are split the same way. There are two texture sets. The Spicy one lives only inside the Spicy content, and the Core body's material never refers to it. The Spicy body gets its own material that swaps in only the two textures that differ.
That last part matters because of a gap it exposed in the build's safety net. The CI leak-audit gate (the check that stops Spicy content ending up in the Steam build) works by looking for Spicy asset paths inside the Core package. A shape key isn't a separate asset. It exports as a morph target, and its data sits inside the body mesh. If the Core export ever carried the removal key by mistake, it would pass the gate cleanly while still containing exactly the geometry it's meant to exclude. So the export script has a rule (the Core file carries no morph that encodes Spicy geometry), and the gate has a new candidate extension: a cook-time check on the body mesh itself. Better to find that hole before production starts than after.


Hair, and a Test of the Magic

The first hair mesh is done, built with the skull-cap method from the documented hair pipeline. Its UVs are about three-quarters laid out.

The layout is deliberate. Every strand is a straight vertical strip in UV space, root at one end and tip at the other. That gives the material a root-to-tip gradient for free, and the gradient drives two things:

  • Colour: the hair blends between two colours along its length.
  • Glow: the same gradient boosts emission toward the tips.
The glow is a test, not final art. When magic is in play, the tips should glow brighter, and this is the cheapest way to find out early whether the emission path through the material system works the way the art direction document says it should.

The Face Is Being Textured Now

That's the work in progress as this report goes out. It's also where the art direction document has its two newest open questions, both written down this month:

  • Does Unreal's new Toon shading model accept a normal map at all? If not, normal maps get skipped on organic meshes like skin and kept for cloth and hard surfaces.
  • Can a face shadow map feed its light/shadow decision? Anime faces usually shade from a hand-painted threshold map rather than true lighting, so that shadow shapes like the triangle under one eye fall where an artist put them. If the Toon model works out the shading band internally with no way to override it, faces may need their own material.
Both get answered in the editor, once there's a face to put on screen.



A Second Test Mode: Haunted House

Delivery Girl, the first test mode, exists to build movement and traversal before the story needs them. This month it got a sibling with a different job: Haunted House, a Halloween horror mode, designed today. Like Delivery Girl it's non-canon, and like Delivery Girl it's a place to build systems the main game needs, somewhere they can be tried without a story depending on them.

The inspiration is the board game Betrayal at House on the Hill: a haunted house that's built as you explore it.

  • The house assembles itself behind you. Doors to unexplored rooms stay shut. Open one, and a room is drawn and placed behind it for the rest of the run. In the fiction, the rooms are pocket spaces stitched together by doorways, an aspect of hell that surfaces at Halloween. That makes the house a graph of doors rather than a floor plan, so stairs and floor changes are just transitions between rooms.
  • The door hides the load. Rooms are loaded when you decide to open the door, and the wait is covered by the door itself: most doors creak open in a second or two, and locked ones are a lock-picking, hacking or trap-disarming minigame. Which room you get is decided the moment you start, so failing the lock doesn't re-roll it.
  • Four from a cast of six. You pick a party of four from Meghan, Celeste, Daisy, Jennifer, Trevor and one more still to be decided. Nobody is magical in this mode: these are the characters and their personalities, not their powers. You play one at a time and switch freely; the others run on AI.
  • The party doesn't stay together. Companions wander, search, find things and get into trouble on their own, and their personalities decide how far they stray. One sticks close; another goes off alone.
  • Capture and rescue. Lose a fight or an event, and that character ends up in a capture room in the basement or upper floor, possibly one nobody has found yet. The direction to a captured character shows in the AR display, and you can switch to them to see where they are. Captives can work on their own escape. At least one capture room has several cells, so getting two captives out means switching back and forth between them. Capture all four and a ten-minute countdown starts.
  • Every run is a mission. The mission sets the house's mystery: appraise the place for a real-estate deal, investigate it for a school newspaper, extract data from a specific room, find a missing schoolmate, find out who sent out party invitations to the place. The twist is a mystery of the house that depends on the mission, not a traitor in the party.
It's solo with an AI party. Co-op would be wonderful and is out of scope, but there's a rule attached: nothing gets built in a way that locks multiplayer out later.

Two practical consequences. First, the character pipeline needs a male base body, which it doesn't have yet; Trevor is a big man and isn't a uniform rescale of the female base. Second, the mode comes after Delivery Girl and before work on the prologue, because it tests systems the prologue needs: combat, the inventory, party switching, capture and rescue, searching, stealth and room hazards. Ready for Halloween 2027 would be lovely. That's a wish, not a date.


Clothing damage runs in every tier

The haunted house has room hazards that damage clothing, and designing them exposed a mistake. One of the design documents had placed clothing damage in the Spicy tier only. That broke a rule the tier system is built on: game mechanics never change between tiers; only content does. A fight shouldn't play differently depending on which content pack you installed.

So clothing damage now runs in every tier, Core included. Core stays stream-safe with a floor at the underwear layer: underwear can't be damaged in Core, and it already meets the Twitch-safe standard. Spicy removes the floor. Same system, same balance, different content. Clothing also gives item-specific bonuses (a gas mask in a room full of gas, gloves, shoes), so what's worn matters to the mechanics in every tier.

The mode doubles as a test of the tier split for content: mission lists differ by tier, so the list has to accept missions that a tier pack adds when it activates.


What each mode tests

The two modes were designed to cover different ground. Roughly: Delivery Girl tests the open world, and Haunted House tests what happens once you get somewhere.

Both modes (each pushes these in a different direction):

  • Third-person camera. Delivery Girl's is a plain follow camera in an open city. Haunted House's has to work in tight rooms: the camera bumps into walls, and doors swing shut so you see at most two rooms at once.
  • Searching. Delivery Girl searches a large district for the crypto keys that unlock night jobs. Haunted House searches a small, dense house room by room.
  • Stealth. Delivery Girl has stealth and exposure on night runs. Haunted House has stealth against the house's enemies.
  • Clothing. In Delivery Girl, what she's wearing affects exposure. In Haunted House, room hazards damage clothing, and some items give bonuses (a gas mask, gloves, shoes).
  • The tier rule. Delivery Girl tests it on bodies and outfits. Haunted House tests it on content: each tier pack adds its own missions.
  • The AR display. Delivery Girl's highlight widget marks keys and containers, and costs widget points and drone battery. Haunted House's points toward captured characters.
Haunted House only:

  • Combat, for the first time. Losing a fight leads to capture, and companions fight on their own.
  • The inventory, for the first time: the reach-based physical inventory from last month's report, where what you can use depends on what's worn, which pockets you can reach and whether your hands are free.
  • A party, and switching between its members.
  • Companion AI that wanders, searches and gets into trouble on its own.
  • Capture and rescue, including captives escaping on their own and rescues that take two characters.
  • Fights you aren't watching, resolved in summary with an alert so you can switch in.
  • Rooms placed as you explore, from a seed, with navigation built at runtime.
  • Loads hidden behind doors, with minigames covering the wait.
  • Room hazards.
  • Missions that set up the run's mystery.
  • More than one playable character, including the male base body the pipeline doesn't have yet.
  • Groundwork for multiplayer, without co-op itself.
Delivery Girl only:

  • A large streamed city that survives the cook. That's CoreX-M2's main job.
  • Day and night as gameplay, with different work for each.
  • Deliveries: the phone app and routes.
  • A debt and an economy where honest work never quite covers it.
  • Drone battery as something the player actively spends.
  • The HUD widget budget as a real cost.
  • City sound: ambience and footsteps that change with the surface.
  • Vehicles, eventually, from a bicycle up to a hoverbike.

Lore: Meghan's Cover Story, and Life on the Station

Most of the month's lore work is spoiler territory and stays private. Some of it doesn't need to.

Meghan's cover story. Meghan arrives on Terridyn needing local ID to enrol, rent and work, and the ID she gets says she was born on the orbital colony, the station that took in the Exodus evacuees. Specifically, it says she's a pre-Exodus native: one of the tens of thousands already living there during the century it ran half-empty, before nearly a million evacuees arrived in a matter of weeks.

It's a good cover, because natives are a small minority most evacuees never met, so "I don't remember you" is normal. It's also under constant pressure, because about nine in ten of her university-age peers grew up on that station, and the dangerous question is the ordinary one: which school did you go to?

Her technique is true feelings, false facts. Six months of drifting between stations before she arrived gives her the texture of station life honestly; only the specific station is false. "Are your parents still up there?" "We don't talk." That one's entirely true, and it closes the subject.

What she can't fake are the physical tells. Someone really raised on the station squints in Terridyn's sunlight (the station gets about a quarter of Earth's), feels heavy for weeks, misjudges catches without the spin, and is startled by a sun that rises in one place and sets in another. Meghan does none of that, and an observant station kid could notice. That's a candidate player mechanic: when someone asks about her past, you choose to deflect, answer with true feelings, or commit to a specific fact, and specifics can be remembered and checked.

Life on the station. Last month's orbit fix led to a proper write-up of the station itself, and the details that reach a player got done carefully:

  • Station slang has its own directions: turnwise and widdershins (with and against the spin), hingeway and sunway (along the axis). Station kids say them by reflex. Meghan has to translate, and can slip.
  • Direction changes your weight. Moving turnwise adds to the spin, so you get heavier; widdershins, lighter. A sprinter feels about 9% either way. A fast tram feels like 1.5g one way and 0.6g the other.
  • Dawn is the mirrors opening. The light comes from mirrors that swing open through the day, so morning is dim and the sun's image sits low toward one end of the station.
  • The trip home is date-dependent. The station orbits the star, not the planet, so its distance changes all the time. At the moment the game starts it's about 323 light-seconds away, a trip of a little over two days. At other times it's on the far side of the star.
  • Why TMC keeps it running. The Terridyn System Defense Force, usually called the TMC Space Force, is based aboard. The system belongs to no interstellar nation and pirates are a standing threat; the station gives the Space Force a home port without Terridyn's gravity well.
The magical girls' colours follow a scheme. The first three are the additive primaries: Ava blue, Nisa green, Nagi red. After that the scheme moves to the subtractive set, starting with Arri yellow.

The rest (how hell is organised, where its demons came from, a good deal of history) stays private. As before, that part will come out in the game.


Housekeeping

The owned-asset inventory is at 1,046 products on Fab, up 5 from a Free Asset Tuesday. 3 of 1,046 have been reviewed for AI-generated artwork. Fab listing pages now show a "Generated with AI" field, which makes the review faster than it was. Blank still means unchecked, not clean.

WorldBLD announced MCP support for its city and road tools mid-month. The road bake that turns the procedural roads into static meshes is still a manual, click-through step with no scripting hook, and it's part of why the packaged game has no city. If the new support exposes the bake, that changes. It hasn't been evaluated yet. The engine upgrade to 5.8.1 stays on hold until WorldBLD ships a compatible build.

The dev machine has been hard-resetting about once a week, about a second after it's unlocked from idle. No crash dump, no hardware error logged. The CPU diagnostic passed, and a GPU power cap is being tested now to rule the power supply in or out. Separately, the main monitor turned out to have a stale 2023 colour calibration loaded into the graphics card, which made it read dim and blue. That's fixed, and the monitor is being kept as the HDR and lighting review screen.


By the Numbers

  • 0 source submissions in October, third month running
  • 7.30: the predicted head count after flattening the foot; 7.29: the measured one
  • 0.518: crotch height against stature, long by design and inside the style's range
  • 23.75 cm: Meghan's foot, accepted as-is
  • 147–192 cm: the cast's height range, all one mesh
  • 1 body finished, 2 tiers exported from it by 1 script
  • 1 gap found in the leak-audit gate, before it could matter
  • 1 hair mesh done, UVs ¾ laid out, 2 things driven by one gradient
  • 2 open questions about Toon shading, both waiting on the face
  • 2 test modes now: Delivery Girl and Haunted House, both non-canon
  • 6 characters, 4 in a party, 0 of them magical, 7 missions so far
  • 1 male base body the pipeline now needs and doesn't have
  • 9 in 10: Meghan's classmates who grew up on the station she claims as home
  • 323 light-seconds: the station's distance when the game starts, a two-day trip
  • 1,046 owned Fab products, 3 reviewed for AI artwork
  • 3 unexplained hard resets on the dev machine, 1 stale colour profile found

What's Next

  • Finish texturing the face, then answer the two Toon shading questions against it in the editor.
  • Finish the hair UVs and build the gradient and glow material.
  • Join the head to the body, rig, skin, and write the two-tier export script with its Core check.
  • Import her. Upgrade the engine at the same time, once WorldBLD allows it.
  • Get the city through the cook. The packaged game still has no city in it.
  • Write up Haunted House as a creative brief, the same way Delivery Girl got one. Real-time capture is the hard open problem: how a fight resolves when you're playing a different character two floors away.
September ended with "she isn't in the engine yet." That's still true. What's changed is that the body she'll arrive in is finished, measured, and already split into both tiers. That was the part most likely to go wrong after import, and it's been dealt with before import instead.

MGA is a solo-developed adult action RPG currently in pre-production. Follow development on the public dev board, on Patreon, or at the Project Tangent SubscribeStar. Every tier, free included, gets the full text of every report. This month's work-in-progress pictures are in this members post only.


Like(0)
Dislike(0)
Comments  loading...
Sign Up or Log In to comment on this post
Aridecan

MGA Dev Report — October 2026

Comments
Like(0)
Dislike(0)
Posted for $3, $4 tiers
Unlock
Aridecan
Public post

MGA Dev Report — September 2026

Ordinary Girl Meghan / Magical Girl Ava


The Honest Number, Again

Source submissions in September: zero.

That's two months in a row, and it's worth saying why this one is different from August.

August's zero had burnout in it. September's doesn't. It was a quiet month in the engine because the work was somewhere else. Last month's report said the critical path for CoreX-M2 now runs through Blender, and this month it did. The character course is at lesson 12. The saved files in the art depot are numbered well past that, but only because Blender's incremental save bumps the number every time, so the filenames overstate the progress. None of this work becomes a changelist until there's a mesh finished enough to import, and there isn't one yet.

So nothing changed in the engine, and the mesh moved forward a lot. Here's the mesh.


The Body Is Done; the Head Is Next

At the end of August the body was a blockout with hands and feet. At the end of September:

  • The full body is done. The hands and feet from August are now part of it rather than separate pieces.
  • The head exists but isn't attached yet. Face, eyes, teeth and tongue are built as separate pieces. Joining the head to the body waits on the foot fix described below, because the order of those steps matters.
  • The hair is half done. The front is finished and the back is in progress, following the course's hair method, which was already the documented hair pipeline for this project.
The face is the part that matters most. It's where the old mesh's worst defect was (eyelid loops that didn't run clean), and it's what every facial expression will be built on. So it's the part being built slowly.


Measuring the proportions, and a trap in the foot

Partway through the month the figure got measured properly for the first time. Stylised anime characters are specified by head count: total standing height divided by the height of the head. It's the single number that most decides whether a figure reads as childlike, heroic, or somewhere in between.

MGA's base body measures 7.25 heads. That sits mid-band for an adult anime figure. It's also a cast-wide constant, because every female character in the game is this one mesh at a different uniform scale. No mechanism exists to vary it per character, and building one would break the shared-animation and shared-clothing design the whole character pipeline depends on.

The measuring turned up a trap that's worth recording.

The foot is modelled in a high-heel pose. It's plantar-flexed 25.2 degrees, so the heel sits 1.72 units above the ball. The lowest point of the mesh is therefore the ball of the foot, not the heel, and you can't see that from the front. From the front the foot looks flat. Measure the figure in front view and you get 7.55 heads, not 7.25, and the character looks noticeably leggier on paper than she is. Reference charts all use barefoot stature, so the correct measurement is heel to crown, taken from the side.

The fix is to flatten the foot in the mesh before rigging, so the bind pose is flat by construction and matches the MetaHuman reference pose the skeleton now targets. The order of the steps matters. Flatten, re-measure (the prediction is about 7.30), apply scale, apply the mirror, and only then join head to body and skin. All of that is now written into the character pipeline document as a numbered procedure rather than left to memory. The alternative, rigging in heels and fixing the rest pose afterwards, is a well-known way to end up with a skeleton that believes the foot is flat while the mesh doesn't.


Hair Under Hats

A design problem that has to be solved before any hair is final: what happens to a character's hair when she puts on a hat?

The easy answer is to hide the hair. It's also wrong. A beanie that makes the whole head of hair vanish looks bald, and a character whose ponytail disappears under a baseball cap looks broken. The rule MGA has adopted is: hide only the hair the headwear actually covers, and reshape the rest.

The hair is cut into four regions along a shared line: cap, crown, fall and fringe. Headwear is sorted into five classes by how it meets the head:

  • Band: sits on a ring around the head (caps, beanies, most hats)
  • Arc: goes over the top from ear to ear (headphones)
  • Strap: a thin band around the head (goggles, respirator straps)
  • Shell: a hard enclosure (helmets)
  • Hood: soft fabric that drapes over everything
Arc and Strap exist because this is a cyberpunk setting. Headphones, goggles and respirators are everyday items, and a system with only hats in mind would treat them badly. Each class hides the regions it covers and applies a per-class compression shape to the rest, so the hair visibly presses in where the headwear sits. The class also gives the hair author and the hat author a shared outline to fit against.

One thing is deliberately out of scope: magical girl hair ornaments. Transformation hair is signature hair and gets handled as its own case.


The Inventory Moved Out of the Phone

A UI decision that fixed two contradictions in the design documents.

The in-game UI now has three categories, and each one is gated by something different:

  • Phone: digital things, available if Meghan has her phone
  • Physical: the inventory, limited by what she can actually reach
  • Out-of-character: player tools that exist outside the fiction, always available
The inventory used to be a section of the phone app, which made no sense. Losing your phone shouldn't stop you checking your pockets, and the prologue taught the inventory before the phone even existed in the story. Now it has its own input, runs in real time, and is reach-based. What it offers depends on what's worn, how it's layered, which pockets are usable, whether her hands are free, and whether she's restrained.

Configuring the action bar moved the other way, out of the phone and into the out-of-character menu. It's a player aid that Meghan herself never sees, so it shouldn't be locked because she lost her phone or transformed.


A Bug in the Sky

The project has a small C++ library for the setting's astronomy: sun and moon positions, twilight, the calendar. This month it turned out to have been putting the orbital colony in the wrong place for over a year.

The colony orbits the star, not the planet, with an 800-day period. The library derived its orbital radius from Kepler's third law with the wrong exponent, which gave 1.34 AU. The correct figure is 1.692 AU. The old number describes a 568-day orbit, and that error spread into the lore as a light-travel distance roughly half of what it should be. Fixed at the source, and every document that quoted the old number now points to one place.

That fix led to a proper write-up of the colony itself. It's a classic three-island O'Neill cylinder, 10 km by 10 km, spun for 1g, with long mirrors bringing in starlight. That produced one principle that's now a working rule:

Background lore that never reaches the player stays deliberately vague. Mirror masses, cap structures and flywheel sizing have each had one plausibility pass and then stopped. Someone born on the station doesn't know every spec by heart. They know it works. The physics gets done properly only where it reaches the player: travel times, light levels, and gravity and spin that someone could notice. A thrown ball curves slightly on a station spinning at 0.42 rpm, and people raised there learn to allow for it. That kind of detail earns its keep. The tonnage of the mirror hinge doesn't.


Character Backgrounds

Some of the month went into the characters' backstories. That's private, spoiler-heavy work, so all this report will say is that it happened. What it produced isn't coming out in a dev report. It'll come out in the game.


Housekeeping

The owned-asset inventory is at 1,041 products on Fab, up 2 from Free Asset Tuesday pickups.

The generative-AI column is live. Last month's report promised it; it now exists as a hand-maintained flag file that's merged into the inventory each time it's rebuilt, so it survives a full re-parse of the library. It has three states: flagged, reviewed and clean, or blank. 1 of 1,041 products has been reviewed so far, and it's the one that prompted the column. Blank means nobody has checked yet. That's not the same as clean, and the inventory says so on its first page.

A week off the day job started yesterday, and it's going straight into CoreX-M2. The engine upgrade to 5.8.1 is still held until the mesh is ready to import, so the toolchain moves in one step.


By the Numbers

  • 0 source submissions in September, second month running, this time with no burnout in it
  • Lesson 12: where the course actually is, whatever the filenames say
  • 1 full body finished; 1 head built but not yet attached
  • ½ a head of hair: the front done, the back in progress
  • 7.25 head count: the cast-wide proportion, measured, not assumed
  • 7.55: what you get if you measure it from the front by mistake
  • 25.2°: the plantar flexion hidden in the foot, to be flattened before rigging
  • 5 headwear classes, 4 hair regions, 0 hats that make a character bald
  • 3 UI categories: phone, physical, out-of-character
  • 1.34 → 1.692 AU: the colony's orbit, after fixing the Kepler exponent
  • 0.42 rpm: the colony's spin, which is why a thrown ball curves slightly there
  • 1,041 owned Fab products, 1 reviewed for AI artwork

What's Next

  • Finish the back of the hair, flatten the foot, re-measure, and attach the head. Then retopology and UVs.
  • Break lockstep and take the female body to an engine-ready export: facial shapes, skeleton, weights, the merged cover shell, and both tier exports.
  • Import it. Upgrade the engine at the same time.
  • Get the city through the cook. The packaged game still has no city in it, and that hasn't changed.
Last month ended with "next month there should be a character." There's more of one than there was: her body is finished, her head is built, and her proportions are measured and written down. She isn't in the engine yet, and that's the job for this week.

MGA is a solo-developed adult action RPG currently in pre-production. Follow development on the public dev board, on Patreon, or at the Project Tangent SubscribeStar. Every tier, free included, gets every update in full.


Like(0)
Dislike(0)
Comments  loading...
Sign Up or Log In to comment on this post
Aridecan
Public post

MGA Dev Report — August 2026

Ordinary Girl Meghan / Magical Girl Ava


The Honest Number

Source submissions in August: zero.

July's report ended with fifty-one. The last changelist in this project's history is dated the 30th of July and nothing has landed since. That isn't a rounding artifact or a reporting-window quirk — no code changed hands this month.

There are two reasons, and they're worth separating because only one of them is a problem.

The first is burnout. July was a hard month by design — a build pipeline stood up from nothing, six settings tabs, a rebindable control scheme, a content-tier system, a leak found and gated. It closed on the 28th with CoreX-M1 declared done and version 0.1.0-cxm1.57 published and verified. Then the tank was empty, and the mesh work that was supposed to start immediately slipped a few weeks. That is what it looks like when one person does this.

The second is a deliberate change of shift. CoreX-M2 is an art milestone. The whole point of it is a character standing in a city, and the character does not exist yet — which means the critical path this month ran through Blender, not through the engine. Blender work doesn't produce changelists until a mesh is finished enough to check in.

So: a quiet month with a real dip in it. Here is what's in it anyway.


First, the Tail of July

Last month's report was written on the 26th, and the four days after it were busy. Folding them in here rather than losing them:

CoreX-M1 closed. The remaining items landed — localization packs published as plugin targets, gamepad glyphs per key, the UI-scale slider fixed, the project's C++ and CI scripts swept to pure ASCII to kill the corrupted-character bug on the splash screen, and a boot-order fix moving content-tier activation to a point in the sequence that isn't forty milliseconds ahead of the age gate. Verified in a packaged build, published, audited.

CoreX-M2 was defined: Delivery Girl: First Run. One sentence — Nessa stands up in a hand-built block of the city, in a packaged build, walkable, with sound. The milestone exists to prove a specific list of unknowns, and the honest headline from planning it was that the packaged game currently has no city in it at all. The city exists in the editor. Getting it through the cook is a real piece of work, not a formality.

The mode got a premise. Nessa is a student working delivery to make rent. One of her packages is a prototype drone; something goes wrong; the drone locks to her and the recipient wants it back. She now has a debt and a piece of equipment she can't return. It's non-canon and deliberately light — the premise exists to give the test mode a direction, not to be a story. The interesting mechanical piece is that she doesn't receive jobs directly: she has to find crypto keys hidden around the city, and each key unlocks the object holding the next batch of work. A day-side app and a night-side key hunt, using the same city.

The third technical unknown was proven. The movement plugin the project is standardising on now builds, cooks, stages and boots in a packaged build. Getting there cost a genuinely surprising discovery — enabling it silently force-enables the Steam online subsystem, which an empty-looking project file gives no hint of.

The first megablock of roads was baked. The city's roads are generated procedurally, which is wonderful in the editor and useless in a shipped build. Sixteen static meshes now exist covering nine full superblocks and seven edge superblocks, with the procedural originals kept alongside so the whole thing can be regenerated after changes. Nanite enabled on all sixteen. This is expected to be re-authored later — it's the first quick pass, not the final geometry — and saying so now is cheaper than pretending otherwise later.


The Month's Real Decision: Start the Body Over

MGA has had a female base mesh since last autumn. It was built between August and October 2025, it's rigged, it's textured, and it has the only working facial shape set in the project.

It's being replaced.

Not because it's broken. Because five specific things about it are wrong, and every one of them gets multiplied by the fact that this is the most-shipped asset in the game — every female character in MGA is this mesh, scaled, with a shape key. A defect here is a defect several hundred times over.

  • Back arch too pronounced — The lumbar curve is overstated; it reads in every idle pose
  • Glutes sitting too high — Wrong mass placement relative to the sacrum
  • Hands and feet — Copied from open-source Reallusion meshes — a third-party dependency in the most-used asset in the project
  • Eyelids not straight — Loops that don't run clean, which fights every facial expression built on top
  • Crotch geometry doesn't flow — Too much geometry, badly routed
There's a third reason on top of the artistic one, and it's the honest one: relearning Blender. A year away from a tool costs you, and there's a course — 3D Anime Character Creation, by the YouTuber 2AM — whose target style is genuinely the style MGA is aiming at, and whose hair-modelling method is already the documented hair pipeline in this project. Same author, same head shape, same assumptions. Following it properly rather than skimming it is a defensible use of a month.


Three sculpts, six exports, one skeleton

The scope grew, deliberately, while the decision was being made. The plan isn't one body — it's three sculpts (female, male average, male heavy), six exports, and a single shared skeleton. The marginal cost of the two male bodies is small, because the course is being worked through anyway and the sculpts will never again be this cheap to make. Three bodies sculpted at the same skill level in the same stretch of weeks look like one art direction; three bodies sculpted a year apart look like three.

The sequencing splits the difference between learning and shipping: run the early chapters — blockout, sculpt, retopo, UV — in lockstep across all three, because that's where repetition actually teaches and where cross-body consistency matters. Then break lockstep, drive the female body all the way to a finished engine-ready export, and come back for the males. Authoring fifty-two facial blendshapes three times in lockstep teaches almost nothing the second and third time and costs a great deal of calendar.

The critical path only needs one of the six. Nessa is the only character CoreX-M2 renders.


Coverage that can't be turned off

The most useful thing to come out of designing this was a compliance decision, and it's a better one than the project had before.

MGA's base game is stream-safe by construction — that was July's commitment, using the Twitch VTuber minimum-cover guideline as an actual spec. The obvious way to implement it is to spawn a cover garment onto the body. The problem with the obvious way is that anything spawned can fail to spawn: a bug, a bad config, a mod, and the base build has a nude character in it.

So the cover is merged into the base mesh itself — a half-millimetre shell with its own material slot, joined into the geometry, shipped as one asset. Coverage becomes inseparable from the thing that ships. That's compliance by construction rather than compliance by spawn logic, which is the standard the content-tier boundary was always supposed to meet and previously met only by intention.

The same reasoning drove the tier split. Rather than hand-editing a second copy of the body — which drifts the first time the sculpt changes and never tells you it has — the master mesh carries a Featureless shape key. The adult export ships with it at zero; the base export bakes it in at full and strips it. Two genuinely separate meshes still ship, with identical vertex order by construction rather than by discipline, and changing the sculpt regenerates both.

One more correction fell out of it. The automated check was going to assert an exact morph-target count on the base export. That's a proxy, and a fragile one — it breaks the moment the facial shape set isn't exactly fifty-two. The robust form is a denylist: assert that no base export carries the four adult morph targets by name. It says what the rule actually says, it survives any future decision about the facial rig, and it fails just as loudly.


The Skeleton Decision, Measured

A separate thread, from thinking about toes.

Standard game skeletons give each foot a single toe bone. MGA wants per-toe articulation — but not at the cost of throwing away compatibility with the enormous library of animation that assumes the Unreal mannequin. The question was whether you can have both.

You can, and it required actually counting rather than assuming. Bone-name sets pulled from the shipping engine assets:

  • Manny / Quinn (UE5 mannequin) — 159
  • MetaHuman body — 342
All 150 of Manny's real deforming joints appear in MetaHuman under identical names — including the whole corrective set. Epic aligned the mannequin to MetaHuman. So retargeting between them is name-identical on every driven chain, and adopting MetaHuman costs nothing in animation compatibility while adding twenty toe joints, a large finger corrective set, and access to MetaHuman-targeted animation and Control Rig assets.

The decision changed: MGA's characters rig to a MetaHuman-compatible skeleton, not Manny.

There is exactly one trap, and it's the kind that would have been found the hard way six months from now. MetaHuman has no IK bones at all. ik_foot_root, ik_hand_gun, weapon_l/r and the rest exist on Manny and are simply absent — and a large amount of third-party animation and most standard foot-IK setups reference them. They have to be added back by hand. That's now a checklist in a technical brief rather than a discovery.

The reasoning here is the same as the hair-bone reasoning from July, and it's becoming a house rule: author the bones now, decide later what drives them. The skeleton is shared by every body and every garment, so adding joints later means re-skinning everything made in between.


What Actually Got Built

Hands, feet, and the body blockout with arms and legs — the geometry, live in the file as this is written:

  • Body block — Vertices 1,273 · Faces 1,228 · All quads, mirrored
  • Hand — Vertices 1,380 · Faces 1,442 · hand-authored from primitives
  • Foot — Vertices 1,238 · Faces 1,310 · hand-authored from primitives
Target for a finished body is roughly 17k triangles, which is Genshin-class and correct for a game that needs crowds. It also settles something worth writing down before someone standardises it away: characters are deliberately not Nanite. Seventeen thousand triangles is below the point where Nanite earns anything, and environment assets in this project get Nanite as a default import step — so characters need to be the stated exception rather than an accident.

Two things about that table are more interesting than the numbers.

Hands and feet came first. They're the hardest geometry on a human body, they're visible constantly, and they were the one part of the old mesh that wasn't MGA's own work. Doing them first means the third-party dependency is gone from the project's most-shipped asset before anything is built on top of it.

All quads on the body. Quads throughout, triangulated only at export. That isn't fastidiousness — it's what makes the cel-shaded outline read cleanly and what makes deformation behave when the character moves. The hands and feet carry a small number of triangles at the fingertips and toe tips, which is normal and fine.


Three Control Modes, Not Two

Early in the month the control scheme got a revision that should have happened earlier.

The design had two on-foot modes: first person and third person. There are three, and the axis that actually matters isn't first-versus-third person at all. It's whether the mouse cursor is captured or free.

  • First Person — Cursor captured · Move WASD, character-relative · Camera anchored to the character's eyes
  • Third Person Direct — Cursor captured · Move WASD, character-relative · Camera anchored to the character — park it over her shoulder and it stays there
  • Third Person Pointer — Cursor free · Move click a destination, or WASD camera-relative · Camera anchored to the world — point it north and it holds north while she turns beneath it
First Person and Third Person Direct share their controls completely; the second is just the first plus a momentary camera-adjust layer. Third Person Pointer is the genuinely different one — a point-and-click paradigm where the entire game is playable with the mouse alone, but where WASD stays fully live, because mouse-only should be an option and never a requirement.

Cursor capture is a hard boundary in the engine as well as in the design, which is why it's the right axis to build the state machine around. Everything else — heads-up display profile, camera rig, animation set — layers on top of it.

CoreX-M2 will ship first person and third-person direct. Pointer mode comes later.


Where Two Stars Are

A worldbuilding session that produced something more durable than expected.

Terridyn and Calibran are now anchored to real catalogued stars. Not as in-fiction naming — nobody in 2352 calls Terridyn's sun "HD 143436" — but as an authoring backbone, so that distances, transit times, orbital parameters and night skies stay consistent and derivable rather than being invented one scene at a time.

  • Star — Terridyn HD 143436 · Calibran 18 Scorpii
  • Constellation, from Earth — Terridyn Serpens · Calibran Scorpius
  • Distance from Sol — Terridyn 145.6 ly · Calibran 46.1 ly
  • Type — Terridyn G3V · Calibran G2Va
Both are documented solar twins with no catalogued planets, which leaves their systems entirely free to invent. The selection had to satisfy three things already fixed in the lore: Terridyn's star is Sol-like and slightly younger than Sol, Terridyn's year is 400 days, and Calibran is roughly six months away by Curve Drive.

The satisfying part is how well the real data cooperated. HD 143436 is a near-exact match to Sol in temperature, mass and iron abundance, and its one anomaly is about six times solar lithium — independent evidence of youth. Its published age carries a ±2.9 billion year uncertainty, which is exactly the lore position: Terridyn's star is younger than Sol and we do not say by how much.

And the geometry landed better than it had any right to. The three systems are nearly collinear: Sol to Calibran is 46.1 light years, Calibran to Terridyn 100.4, Sol to Terridyn 145.6 direct. Calibran sits 7.5 light years off the straight line — routing through it costs a 0.6% detour, which is nothing next to the fact that a jump needs a star at each end anyway. Calibran is on the road to Terridyn because the real sky puts it there.

Meghan's journey out is about a year, aimless for the first half and directed for the second. That now has a number attached to it rather than a vibe.


Diamond Shreddies

The other document written this month is about how this project talks to the people watching it, and it exists because of a YouTube video about breakfast cereal.

In 2008, Post's Canadian Shreddies brand rotated its square product 45 degrees and marketed the result as Diamond Shreddies. New shape. Completely deadpan, at full production values, eventually reaching a combo pack containing both squares and diamonds. Nobody was fooled. Nobody was meant to be. It ran for years and it worked.

Three properties of that campaign describe how MGA is made, and they're now written down as a commitment rather than a vibe:

Enormous sincere investment behind a premise that doesn't need it. The commitment is the point. Work done far past the level the joke requires stops being a joke and becomes an act of good faith.

The audience is a co-conspirator, not a mark. The pleasure is being on the inside of something performed with a straight face. The moment an audience is being worked, the whole thing collapses.

The affection is displaced into the construction. Nothing is stated directly. The effort is the statement.

That last one is the register the game is written in — reticence expressed as work, feeling routed through infrastructure because the direct channel is closed. It's also, not coincidentally, a description of the protagonist.

The reason this needed writing down now, in a month with nothing to show, is the rule that came out of it. A project this long has to answer a question a short one never faces: what is the relationship with people who are paying attention for years? The failure mode is promise debt — building anticipation for a payoff you keep deferring, until the audience stops being collaborators and starts being creditors. The rule is therefore: the rotation must be visible early; the reason can take years. Distributed payoff, never deferred payoff. Show the diamond now. Never sell the reveal.

Which is, in the end, why this report opens with a zero instead of burying it.


The bill that isn't being sent

MGA picked up its first paying patron this month. Eight other people follow along for free; one of them decided to pay.

Under everything written above, the correct response to that isn't a thank-you post. It's a pause on billing — no charges until there's something here worth charging for.

Because look honestly at what the money would currently be buying. Eight monthly reports and a public task board. That's a real thing and I'm glad people read it, but it is text about a game, and text about a game is not a game. Charging for it would mean accepting money against a promise, which is the precise mechanism the section above exists to refuse. Being on the inside of something built in public should cost nothing until the thing being built can be installed.

So the billing is paused. It turns back on when there's a build worth putting on a hard drive — and that is a description of a state, not a date.


What the free tiers get

Everything. There is no paywalled development content, and there won't be while the billing is paused. These reports go out in full to every tier on both platforms, and the public dev board — which is more current than any monthly report — has never required an account at all.

One thing about that will change shape later, and it's worth saying in advance so it doesn't look like a tier being invented.

MGA is an adult game, and the character work now underway means these reports will eventually want to show images. Some of those images won't be postable everywhere. When that starts, each update splits into two editions covering the same development — one with the adult images, one without. Not two tiers, and not a paid edition and a free one. The same report, twice, so that the version safe to post in a given place exists without the development being cut down to fit it.

Which is the same split the game itself is built around: a base that is safe by construction, and adult content as a separate thing you opt into. It would be strange to architect the game that way and then not run the dev updates the same way.


Housekeeping

The owned-asset inventory is now 1,039 products on Fab, up 26 across three merges this month. The tooling for this is boring and keeps earning its keep: a saved library page goes in, a delta comes out, and the "do we already own something for X?" digest regenerates.

Two audio bundles were checked and both were declined. A 38-product bundle turned out to be 28 products already owned — about C$1,470 of it — against 9 genuinely new. Separately, a vendor's AI-powered "build a bundle for your game" tool was given a detailed prompt about MGA and returned ten recommendations, three of which were packs already sitting in the account, and six of which were later volumes of series already owned. The tool matches a catalogue against your prompt; it doesn't check what you have. Useful to know, and cheaper to find out this way.

A new column is coming to the asset inventory: generative-AI artwork. Found on a pack whose store description disclosed AI content in one texture on a prop screen — trivially replaceable, but only if you know it's there. Two reasons it needs tracking: Steam's submission flow asks for an affirmative statement about AI content in the shipped build, and that answer needs a record behind it rather than a memory; and triage differs enormously between "a texture on a screen" and "the base meshes", which you want to tell apart without reopening the store page. It's designed as a tri-state — yes, reviewed-and-clean, or blank — rather than a checkbox, because a two-state box would silently manufacture a thousand unearned "clean" assertions on packs nobody has looked at. Blank is the honest default.

Unreal 5.9 exists. Announced on stage at Unreal Fest in Seoul and not yet in any published Epic material, which makes it one of those officially-unofficial situations. Noted; nothing to do about it yet. The engine upgrade to 5.8.1 is deliberately held until the body mesh is finished and ready to import, at which point everything moves together.


By the Numbers

  • 0 source submissions in August, against 51 in July — the honest headline
  • ~3 weeks of slippage on the mesh work, from burnout, said plainly
  • 1 base body mesh being rebuilt from scratch, for 5 specific defects
  • 3 sculpts, 6 exports, 1 shared skeleton — the scope it grew into
  • 159 vs 342 — mannequin bones vs MetaHuman bones, measured rather than assumed
  • 150 of the mannequin's deforming joints present in MetaHuman under identical names
  • 9 IK and weapon bones MetaHuman is missing, which have to be added back by hand
  • 3 hand-authored pieces finished: hands, feet, and the body blockout
  • 1,273 / 1,380 / 1,238 vertices — body block, hand, foot
  • ~17k triangle target per finished body; not Nanite, deliberately
  • 0.5mm — the thickness of a cover shell merged into the mesh so it cannot fail to spawn
  • 3 on-foot control modes, where the design previously had 2
  • 9 patrons, 1 of them paying — and billing now paused until there's a build worth paying for
  • 2 real stars now anchoring the setting, 145.6 and 46.1 light years from Sol
  • 0.6% — the detour cost of routing Sol → Calibran → Terridyn
  • 16 road static meshes baked, covering 9 full and 7 edge superblocks
  • 1,039 owned Fab products inventoried, +26 this month
  • 28 of 38 products in an audio bundle already owned — bundle declined
  • 3 of 10 AI-recommended asset packs already owned — recommender doesn't check entitlements

What's Next

  • Finish the body. Torso and head are the remaining sculpt work, then retopology and UVs across all three bodies in lockstep. That's the critical path and everything else waits behind it.
  • Then break lockstep and drive the female body through the facial shape set, skeleton, weights, cover shell and export — the first character asset this project has made entirely itself.
  • Get the city through the cook. The packaged game still has no city in it. The roads are baked; the buildings and the streaming setup aren't.
  • Upgrade the engine to 5.8.1 once the mesh is ready to import, so the toolchain moves in one deliberate step instead of piecemeal.
  • Implement the generative-AI column in the asset inventory, and start the review pass that fills it in.
July's report ended with a machine that turns this project into something a person can install. This month didn't add to that machine at all. What it did instead was decide — carefully, and with more measuring than usual — what the machine is going to be asked to carry.

Next month there should be a character.

MGA is a solo-developed adult action RPG currently in pre-production. Follow development on the public dev board, on Patreon, or at the Project Tangent SubscribeStar. Every tier, free included, gets every update in full.

Like(0)
Dislike(0)
Comments  loading...
Sign Up or Log In to comment on this post
Aridecan
Public post

MGA Dev Report — July 2026

Ordinary Girl Meghan / Magical Girl Ava


Two Pillars

June ended with a city that builds itself. July stopped building the world and started building the game.

The month got a name — CoreX-M1 — borrowed from the EA design flow: prove out the X. Before you make content, you de-risk the core technologies a viable game needs, using placeholders instead of art. Not "make the first level." Make the machinery the first level will one day sit on, and prove it works.

Two pillars. One you can see: the front end — the thing that happens between double-clicking the icon and standing in the world. One you can't: the pipeline that turns a source tree into something a player can install. July drove hard at both, and the honest way to tell the month is as two parallel stories.

By the end of it, MGA boots from a desktop shortcut, shows a splash, asks your age, tells you what you're about to see, opens a main menu, lets you set your resolution and your volume levels and your key bindings and your language, remembers all of it, and quits cleanly. None of that is gameplay. All of it is the difference between a project and a product.

To be clear about where that leaves the milestone: CoreX-M1 is not finished. Most of July was spent driving at it, and the two pillars below are the bulk of that work — but the milestone doesn't get declared done until the content packs hold actual placeholder content, the remaining control-screen pieces land, the main menu correctly greys out what isn't implemented yet, and the whole thing has been verified in a packaged build rather than in the editor. Those gaps are real and they're tracked. This is a progress report, not a finish line.


Pillar A — The Front End


The boot sequence, visible end to end

The boot flow is a state machine (B0 through B8), and the rule this month was that every stage of it gets a visible screen — a coloured panel with the stage's name on it — so you can always tell where you are and, more usefully, where it stopped. Placeholders on purpose: zero art files, engine-drawn, replaced later.

It runs: pre-engine splash → age gate → legal and content warnings → activation progress → content advisory → main menu.

That first step is stranger than it sounds. The splash has to draw before the engine exists — before UObjects, before Blueprints, before almost everything. It's raw Slate, registered from a C++ module, and it cost two genuinely obscure lessons. The module has to register at exactly the right loading phase (register too early, as everything sensible suggests, and the screen manager hasn't been created yet, so you get a null pointer). And the manager calls OnPlay() and then asks for your widget, but never calls Init() — so if you build your widget in Init like every example implies, you get a perfect, silent black screen.


Settings, six tabs deep

The settings menu went from nothing to six working tabs: Language, Game Play, Video, Audio, Controls, Sensitive Content.

Video does resolution, window mode, VSync, frame limit and quality scaling, backed by a subclass of Unreal's own settings object — plus the unglamorous correctness work that separates "it has a resolution dropdown" from "it works": preselecting the current resolution, clamping windowed sizes to the actual monitor work area so the title bar can't end up off-screen, and making the packaged build DPI-aware.

Audio is eight channels, and the split is a design statement rather than a technical one: Master, out-of-fiction music and in-fiction music (the bar jukebox, the boombox on the street corner — the music that exists inside the story), character dialogue, ambient dialogue, the player's own sound effects (footsteps, breathing), ambient effects, and interface. Anyone who has tried to hear a conversation over a nightclub will understand why those are separate sliders.


Controls: rebind everything

A full Enhanced Input rebinding screen, which turned out to be the single biggest chunk of the month: 52 canonical actions in a fixed, human-sensible order, click-to-capture ("press any key…"), separate keyboard and gamepad slots, per-cell clear buttons, a settings-wide reset to defaults, and live conflict detection that refuses a binding already in use.

Conflict detection is where the interesting bug lived. First-person bindings were blocking third-person bindings — the validator had only one axis of exclusion, so it couldn't express "these two contexts never run at the same time, let them share a key." Fixing it meant teaching the system about mutually exclusive control contexts, which is a better system than the one that was there before the bug showed up.


Language, and a template for someone else's work

Language got a per-module table: one row per content module, a Text column and a Voice column, each a dropdown, each able to be set to None. Source text is English (Canada); everything else is a swappable content pack.

The naming convention is the point: LocText_<Tier>_<Language> / LocVoice_<Tier>_<Language>. A community translator clones one of those, fills it in, and ships it — no engine access, no source, no permission needed. That's the north star for localization here, and this month is when the plumbing to support it got built rather than just described. (The packs themselves are still empty skeletons; filling them is next month's problem.)


Streamer safety, engineered rather than promised

MGA is an adult game, and adult content is a content tier — separately installed, separately cooked, not present unless you add it. That structure made a related decision easy to state precisely:

The base game's outfits meet the Twitch VTuber minimum-cover standard. Not "we'll be careful" — an actual published third-party guideline, used as the spec. The base game is stream-safe by construction, which means Streamer Mode isn't a filter bolted on afterwards; it's the default state of the shipped product. On top of that: a Sensitive Content settings tab where the adult toggles live, and a boot-time advisory that points recording and streaming users at it before they've broadcast anything they didn't intend to.


Pillar B — The Pipeline


From zero build jobs to a packaged game

On the 1st of July there were no build jobs at all. The project built on one machine, by hand.

By the 11th, a Jenkins pipeline (mga-weekly, running on a dedicated Windows build agent) was doing the whole thing unattended: sync from version control, build the C++, cook the content, package it, audit it, archive it. First green end-to-end build: ~23 minutes, producing an executable and pak files that installed and launched.

Getting there meant confronting an awkward truth: three of the project's most valuable plugins are editor-only city-building tools, and one of them ships with source that doesn't compile. They can't go in the shipped game and they don't need to. The pipeline now switches them off transiently — editing its own working copy of the project file, building, then restoring it — so the depot and the live editor are never touched. It's a small trick that unblocked everything downstream.


Content tiers as real, separate downloads

The adult tiers aren't a flag in a config file. Each is a GameFeature plugin cooked into its own pak, which the game discovers, mounts and activates at runtime — and which simply isn't there if you didn't install it.

The test for whether that works is deliberately the smallest possible surface: the content advisory screen. The base game supplies a warning. Installing a tier registers an override that chains onto the base one. Installing the next tier chains again. So the number of lines on that one screen tells you, unambiguously, which tiers loaded and in what order — and it exercises exactly the mechanism the whole game will eventually need for tiered content, on a surface small enough to debug.

It works. Base alone gives one line, base plus one tier gives two, all three give three, and removing the middle tier correctly refuses to load the one that depends on it rather than half-loading it.


Versioning and publishing

Builds are now stamped 0.1.0-cxm1.<build> and published as separate 7z archives — one for the base game, one per content tier — to a release share, with a manifest recording sizes, hashes and the exact source revision.

The one-archive-per-component split is doing real work: it makes "did an adult tier ship in the Steam build?" a question about which file was uploaded, rather than a question about the contents of a 900 MB blob.


The audit gate

Every release build now runs a gate that asserts no content-tier data is present in the base game's containers, and fails the build if any is. It was written after a leak that put two tier files into a base build — caught before release, fixed at the root, and now the sort of thing that can't recur silently.

Two details from that worth keeping: the leak came from the cooker, not the game (the cook process was loading tier data as a side effect of its own startup, which quietly pulled it into the build); and the gate has to inspect the modern container format directly, because the leaked files lived in the 870 MB data container while the legacy pak the gate originally read held only 11 MB of descriptors. A gate that checks the wrong file passes the build that ships the problem.


Act Three — Story and Design

The first week of July, before the engine work took over, was spent on the script.

A full-arc critique. The whole story got read end to end against its own structure, and the most useful finding was an uncomfortable one: the protagonist is too passive. Things happen to Meghan and she reacts. Two independent read-throughs landed on it separately, which is usually a sign it's real.

The fix isn't to rewrite her into someone louder. It's a principle now written down: solve it on the player-agency axis, not the protagonist-agency axis. A quiet, reactive character is a legitimate and interesting protagonist — the problem is a player who has nothing to decide. Give the player meaningful choices inside Meghan's situation and the passivity reads as characterisation instead of a hole. Two related rules came out of the same pass: every chapter needs its own arc, not just a position in the larger one, and stakes described as cosmic have to be made concrete and put on a clock or they stop mattering.

The finale also got split into two chapters. It was doing too much work as one.

The Chapter 1 economy. The month's best design work went into the ordinary-life half of the game: a time-management loop with a long-term stamina budget, study time toward an entrance exam, and five jobs — three legal, two decidedly not.

The legal jobs are deliberately tuned to cover only about 75–90% of your monthly expenses.

That number is the whole design. It isn't a difficulty knob; it's the coupling between the game's two worlds. A player who does everything correctly and legally still comes up a little short each month — and the shortfall, not a cutscene, is what makes the riskier work look reasonable. The normal world's demands push you toward the secret one. There's a satirical edge to it as well: the legal jobs enforce a hard 40-hour weekly cap, which is presented as a worker protection and functions in practice as a funnel toward the unregulated market.


Housekeeping: knowing what we already own

One unglamorous but genuinely valuable exercise: a full audit of the asset libraries this project has accumulated. Store pages and launcher caches were parsed into a searchable inventory — 1,013 products on Fab, 269 from Cosmos, 537 archives on disk, 247 audio packs — and then triaged into a "do we already own something for X?" digest organised by what MGA actually needs.

For a solo project the expensive mistake isn't buying an asset. It's spending a week building something you already had.


Following Along Between Reports

These reports come out once a month, which is a long time to wait to find out whether anything moved. So the work is now tracked in the open, on a public board:

It's a Kanban board — a column per state, cards moving left to right as work progresses. The columns are Backlog → Todo → In Progress → In Review → Done, and each card is a real task with the actual detail attached: what it is, why it's being done, what's blocking it, and what was decided. Clicking a card gets you the discussion, not just a title.

Everything in this report exists on that board somewhere, and so does everything in the "What's Next" list below. If you want to know what's being worked on today rather than a month from now, that's the place — including the unglamorous half that never makes it into a monthly write-up.

Two honest caveats. It's a development board, so it's dense with engine and pipeline work and won't always be readable if you're here for the story. And there's a second, private board for story and lore, for the obvious reason that a public list of plot tasks would spoil the game for everyone reading it.


By the Numbers

  • 2 pillars in the milestone: front-end framework, build pipeline
  • 6 distinct boot screens built — pre-engine splash, age gate, legal/warnings, activation progress, content advisory, main menu
  • 6 settings tabs — Language, Game Play, Video, Audio, Controls, Sensitive Content
  • 8 independent audio channels, including a diegetic/non-diegetic music split
  • 52 rebindable actions, keyboard and gamepad, with conflict detection
  • 51 source submissions in July (against 31 in June) — almost all engine and pipeline
  • 0 → 1 continuous build pipeline: sync, build, cook, package, audit, archive
  • ~23 min for the first green end-to-end packaged build
  • 3 separately cooked, separately installed content paks (base + two tiers)
  • 4 install configurations verified automatically, including the negative test that a tier refuses to load without its dependency
  • 1 release-versioning and publishing scheme, one archive per component
  • 1 audit gate now standing between a content tier and the base build
  • 1 protagonist-agency problem found, and a principle written down for fixing it
  • 75–90% — the share of monthly expenses the legal jobs cover, on purpose
  • 1,013 / 269 / 537 / 247 — owned assets inventoried across Fab, Cosmos, disk and audio

What's Next

  • Fill the loadables. The content packs are proven as plumbing and empty as content. Next is a machine-translation pass over the string tables so the localization packs have something in them, and real placeholder content in the tiers.
  • Finish the front end for the main menu. Gamepad glyphs in the rebind rows, the analog sensitivity and dead-zone controls, and the main-menu buttons that should be visibly disabled until there's a save system behind them. Worth being precise about the scope of that word: everything above is the out-of-game front end. The in-game front end — HUD, pause and inventory menus, the whole interface you use while actually playing — hasn't been started, and isn't part of this milestone.
  • Two known bugs, still open. The pre-engine splash renders a corrupted character at the end of each line — root-caused this week to a version-control encoding setting that silently mangles non-ASCII text only on the build machine, so it never reproduced locally. And the UI scale slider does nothing: also root-caused (it was driving the wrong scaling hook — the game viewport recomputes its own scale every frame and overrode it), fixed in source, not yet verified in a packaged build.
  • Close out CoreX-M1. The milestone is a foundation, not a feature, and it exits when the boring load-bearing things are actually finished and proven in a packaged build — not when they mostly work on the dev machine. The remaining items are on the board.
  • Then: gameplay. Which is the entire point of building a foundation.
June's report ended with "the first city was built by hand; the second one is learning to build itself." July's ending is less romantic and possibly more important: there is now a machine that turns this project into something a person can install and run.

Like(0)
Dislike(0)
Comments  loading...
Sign Up or Log In to comment on this post
View next posts (6 / 47)
WE USE COOKIES

SubscribeStar and its trusted third parties collect browsing information as specified in the Privacy Policy and use cookies or similar technologies for analysis and technical purposes and, with your consent, for functionality, experience, and measurement as specified in the Cookies Policy.

Your Privacy Choices

We understand and respect your privacy concerns. However, some cookies are strictly necessary for proper website's functionality and cannon be denied.

Optional cookies are configurable. Disabling some of those may make related features unavailable.

We do NOT sell any information obtained through cookies to third-party marketing services.