1:dev/lab/hexmap.md
" ==========================================
" rcepre / portfolio
" last release: 2026
" ==========================================
▾ audio/
·· 
README.md
README.md·· 
reterraform.antres
reterraform.antres·· space_dust.antres
space_dust.antres·· the_junkyard_of_the_jettisoned.antres
the_junkyard_of_the_jettisoned.antres▾ blog/
·· apte.md
apte.md▾ dev/
·· ▸ lab/
·· ▸ projects/
·· 42-school.md
42-school.md·· experience.md
experience.md·· skills.md
skills.md
hello-world.html0:raw1:render*
hexmap
│
Geophysical cellular automaton on a hexagonal grid. Local propagation rules, and rivers, microclimates and seasons emerging.
The thing
Rust. ~500 hexagons, double-buffering, deterministic from a seed. The world is a terrarium: closed, water and energy conserved. If a total drifts, it's a bug, and a test fails.
The confession
I don't have the physics skills to write this, nor the meteorology, nor the erosion. The code is largely written by Claude, steered toward the right models. So the project's question isn't "can I write it" but "can I tell when it's wrong".
Verification
▸conservation invariants tested continuously (water, energy)
▸5 diagnostic tools: cloud clusters, rain attractors, water flows,
cycles, drainage basins. Numbers, not visual impressions
▸hypotheses tested against data: the "lake-cell" hypothesis fell to
direct measurement (0 lakes out of the 6 suspect cells). The same
tooling showed a near-periodic simulation year over year (0.998
correlation), open issue, fix specified
Status
Paused since May 2026: v0.6.1-pause tag, issues triaged, documented
re-entry point with its validation metric.
NORMALdev/lab/hexmap.md[+]
^E explorer│^R render│^; theme
markdown35L1:1
FRcatppuccin