Why it is built
the way it is.

Written 6 October 2026

Saving the world, faster

The server runs the world in ticks, one per second. Every thirtieth tick it saves the state of every character online: position, health, mana, experience, skills, the explored map, what has been discovered, active effects and more. That save is called the flush.

Until now the flush sent one database command per character. With 300 test characters online, the tick that saved them used about half of its second.

Now all rows go together, up to a thousand characters per command. A second change puts everything that must be saved before anyone is told about it (gold, loot, deaths, level-ups) into one transaction per tick: all of it is saved, or none of it.

It matters because the design target is about a thousand players at once in one undivided world, and that scale goal is the risk that steers the architecture hardest.

The figures below compare before and after on the same machine, in a benchmark; they do not say what a live server will see. Sending each tick’s updates to the players in parallel is not done yet.

• Saving 300 characters: 489 ms and 300 commands before, 69 ms and one command after.
• Saving 3,000 characters: 4,917 ms before, 550 ms after.
• Wear on 300 items: 387 ms before, 6.5 ms after.
• One tick with 100 deaths: 245 ms and 209 commands before, 54 ms and 11 commands after.

Written 4 October 2026

A mystery, not a war

The world now has a premise. It was one continent, alone; then other continents arrived and joined it, each a whole mythology. Since then the world is out of balance: the whole world, not only its magic. Everything falls into disorder when whole continents suddenly appear.

The newcomers do not attack. That is the point: the story is a mystery to investigate, not a war to win. Why are they here, what happened, is it an invasion? Something in the order of the world has shifted, and what was known has to be written anew.

The mythologies are one layer of the game, not its centre. They are also meant to teach: area names, the people who give quests, quest texts and the pictures of the beings should show how a mythology is built. The Nordic one comes first, north of the first city.

The oath every character swears concerns the task, not obedience. How a character meets the imbalance is its own matter. That is also where player-against-player combat gets its meaning: the dispositions are different answers to the same oath.

Written 3 October 2026

Why progression is slow

Glinaur is meant to be played over years: hundreds of levels, long chains of schools, equipment that scales high. The slow pace is decided, and deliberate.

There are two reasons. A long road is what keeps a living population online; the design is dimensioned for 400 to 600 players around the clock. And system depth that keeps scaling is the only kind of content one developer can afford to make.

The gap between newcomers and veterans is to be closed with mechanics (mentorship, perhaps a way to catch up), not by shortening the road. Neither is built, and their form is still open.

Written 3 October 2026

Engines before content

The game is built engine first. Content and balance numbers are data, edited in a panel, never code: a new monster, shop, school or region needs no programming.

A mechanic is proven once. One iron vein proves mining; more metals are just more rows of data. That is why the world looks small today: what you see understates what the engines can carry.

It is also why almost everything you see (names, items, monsters, prices and every number) is a placeholder. The content and the balance come in their own passes, once the systems exist.

Written 3 October 2026

No dates, only dependencies

You will not find a date for anything on this site. Work is ordered by what depends on what (what has to exist before the next thing can be built), never by a calendar or by urgency.

Design is different: it is always open. Any design question can be settled at any time; only the building has an order. That is why suggestions are useful at every stage.

Written 3 October 2026

The risks, named honestly

The design documents name their own risks:

• The scale goal (about a thousand players at once in one undivided world) steers the architecture, and has to be stress-tested early.
• A world that changes and remembers: plots, buildings, shop stock and corpses on the ground all persist.
• Plots, raids and city control are the largest new design surface, and the largest risk to the scope for one developer.
• The economy needs continuous tuning, and measurement built into the game.
• Producing content needs tools, now largely answered by the editing panel.
• The design depends on its population: below roughly a hundred players online, the core mechanics lose their meaning. NPC demand with dynamic prices, decided but not built, is meant to soften that.
• A living NPC simulation is a large and growing surface.