SILVERSTAR MAZZY
Janae Barbra Petersen · creator mark Silverstar Mazzy

Systems with a pulse.

I build playable worlds where ecology, weather, lineage, and player choice share one ruleset. This portfolio documents the Age of Eons work visible in this repository: native Unreal C++, deterministic simulation code, authored data, a local web portal, and the tools that connect them.

C++ / Unreal EngineData-driven simulationWorldbuilding toolsEvidence-led workflow
Selected work

A world built as a system.

Age of Eons is an open-world prehistoric prototype. The strongest parts are the rules underneath the spectacle: server-authoritative creature state, deterministic ecology, and a toolchain that makes iteration inspectable.

Built and checked here

Selected implementation records.

The artifacts below are local proof. They describe implementation and recorded verification, not employment history, awards, or a shipped commercial game.

21focused beta client tests passing
17beta/config security checks recorded
5.7Unreal Engine target in build logs
1deterministic open-world prototype
Recorded Frontier world preview showing a prehistoric landscape
Case study · Frontier world pass

From authored seeds to a readable landscape.

The Frontier pass combines procedural terrain, biome dressing, climate and creature systems into an Unreal world. The recorded verification says the ordinary game continued beyond map loading with 40 AI and regular autosave.

Built/tested: local Unreal capture and verification artifacts. Planned: production multiplayer release, public server operations, and final art polish.

Case study · Deterministic rules

Simulation that can be reasoned about.

Weather and ecology are seeded and advanced through bounded data models. Headless tests cover creature, ecology, registry, and world math paths so a visual pass does not become the only source of truth.

// same seed → same world inputs
FAoEWeatherSystem Weather;
Weather.Bind(Config, WorldSeed);
Weather.Advance(DeltaHours);
const auto& State = Weather.Current();
Recorded Age of Eons world capture
Creature hall · September 2026

Every animal was two animals.

The sea cow file was supposed to hold one animal. It held two, stacked inside a single mesh, sharing one name and one skeleton — and so did almost everything else on the shelf. What follows is the work of pulling them apart, giving each one a skeleton fitted to its own anatomy, and proving the deformation instead of hoping for it.

28models separated and filed
23animation clips on the pilot species
0unweighted vertices
16,829→84damaged edges after the fluke fix
Steller’s sea cow, neutral pose
Case study · Hydrodamalis gigas

The fluke was tearing itself in half.

Two bones, left and right, pointing opposite directions on a long lever arm. Every bend of the tail ripped the plate down its own midline. A sirenian fluke is a single stiff plate with no independent left and right articulation, so both bones came out and the chain now runs straight through it. Damaged edges fell from 16,829 to 84.

The flipper had a quieter version of the same fault. The first transition field measured only how far out along the limb a vertex sat — but the belly also sits below the shoulder, so the belly was being dragged along with it. Ten thousand of seventeen thousand damaged edges were nowhere near the flipper.

Built and tested: 25 static stress poses and 23 clips, every frame measured against rest, five camera angles. Planned: engine import, animation blueprint and playable wiring.

Case study · Anatomy over assumption

It never had teeth.

The reference notes listed continuous tooth replacement and tusks. That describes manatees and dugongs. Hydrodamalis gigas was completely toothless and ground kelp between rough keratinous plates — which is what Georg Steller wrote down in 1741, watching them off Bering Island. Five hundred and fifty-eight sculpted tooth vertices were relaxed back into the gum line and the plates were built in their place.

The mouth had no interior either — open the jaw and you saw straight through the skull. It now has an oral cavity, a tongue on three segments off the jaw, and an eye seated at the socket the skull actually has. The original placement sat forty centimetres too low, which is the entire reason it read as a crease.

Honest limitation: the blink is switched off. Five approaches to an eyelid each failed differently, and all for one reason — there is no socket recess and no lid edge loop for a lid to fold along. That is modelling work, and it is not finished.

Sea cow head, reseated eye and rebuilt mouth
The roster

Fourteen bodies, six body plans.

A skeleton renamed eleven times is not eleven rigs. Limb layout comes from where each body actually meets the ground, so four contact points build a quadruped, two build a biped, and none builds something finned. The dragonfly gets six legs and four wings. The ichthyosaur gets an axial chain and steering paddles.

Ankylosaurus
A. magniventris

Ankylosaurus

27 bones, 26 deform, four limbs from four ground contacts. Its tail club is wider than its skull, which broke the first method I tried for finding which end was the head.

Uintatherium
U. anceps

Uintatherium

27 bones. Short tail and a long projecting head — the combination that broke the second method, and sent me to rendering every model and looking instead.

Deinocheirus
D. mirificus

Deinocheirus

21 bones, bipedal. Two ground clusters instead of four, so the layout resolved to a biped without being told to.

Maiasaura
M. peeblesorum

Maiasaura

27 bones. Separated from a fused pair, oriented, and scaled from measured landmarks rather than a bounding box.

Ichthyosaurus
Ichthyosaurus sp.

Ichthyosaurus

19 bones. No ground contact at all, which is itself the signal — an axial chain with steering paddles rather than legs.

Mosasaurus
Mosasaurus sp.

Mosasaurus

19 bones. Its body runs along a different axis to everything else on the shelf, caught and corrected before the skeleton was placed.

Meganeura
Meganeura sp.

Meganeura

33 bones, six legs and four wings. Its longest axis is the wingspan, not the body — the first pass ran a spine straight down the wings.

Smilodon
Smilodon sp.

Smilodon

27 bones, feliform layout. Separated and filed with the rest of the hall.

Protoceratops
P. andrewsi

Protoceratops

27 bones, ceratopsian. Four ground contacts, quadrupedal chain.

Hipparion
Hipparion sp.

Hipparion

36 deform bones. Clears eighteen equine stress poses — graze, rear, gallop gather and extend, per-leg swings, neck turns, tail — with no failures.

Arthropleura
Arthropleura sp.

Arthropleura

Segmented arthropod body, separated from a fused pair.

Isopod
Isopoda

Isopod

Segmented arthropod. One of several isopod forms recovered and filed separately.

Ram attack, impact frame
Case study · Combat

The attack is the whole animal.

Several tonnes of animal does not attack by opening its mouth. The weapon is mass. The ram coils, charges, lands chest-first, takes the shock through the whole body and settles slowly across 132 frames — and the contact frame is written to a notify map, so the damage window lands on the frame you can see rather than on a guessed delay.

At impact, the worst-scoring frame in the clip, 0.128% of edges are strained. Nothing tears.

Case study · Proof

Measured, not eyeballed.

Every pose is measured against rest — edge stretch, triangle collapse, inverted faces, vertices flying away from their own neighbourhood — and rendered from five angles. A rig that reads clean in profile can be torn open from underneath, and the ventral view is where flipper-root failure shows.

Distance-based weighting has no concept of a midline, so a left shoulder will happily take hold of right-side skin through the chest. A smooth lateral gate, reading the rig’s own convention off the data rather than assuming it, removed 174,008 cross-midline influences on one animal and 179,856 on the other. Both now sit at zero.

Where it stands: the sea cow is asset-ready and not yet in engine. Twenty-eight models are separated, named and filed; most are rigged and waiting on weights.

Ram impact seen from below
Path of Titans · current context

Modding with a clear runway.

The Age of Eons work is an independent Unreal prototype and portfolio proof. It is not presented as an Alderon Games mod, approval, partnership, or released Path of Titans content.

Alderon’s current guidance

Private testing

Alderon’s current FAQ says Path of Titans modding is under private testing; people with prior modding experience can sign up through the official creator portal. The application guide says applicants must be at least 18, complete all application steps, and may need to accept an NDA. Access is not guaranteed.

Next honest step

Build the bridge

For an actual application, I would prepare a compact mod sample using the authorized Dev Kit, document original assets and permissions, and follow Alderon’s current portal and agreement requirements after access is granted.