Post #1

We want Henge to feel comfortable to play on a wide range of computers.

Henge already runs in a web browser and on old computers using Windows 3.1. Both versions share the same main game code. Working with older machines helps us spot waste that a newer computer might hide. If something takes too much work to draw, an old machine makes that very clear.

Our target on the Cyrix MediaGX, one of the old computers we test with, is 15 frames per second. That target is still a goal. Some scenes run much better than others, and busy or demanding scenes can fall well short of it. We want to be clear about that while explaining what we’re doing to improve things.

A useful first step is finding out where the time goes. Showing a scene involves several jobs: working out where people and objects belong, drawing the landscape, drawing characters, adding the interface, and putting the finished picture on your screen. We measure those jobs separately so we can tell which one is slowing a particular scene down.

We also look for pauses. A game can seem reasonably quick most of the time and still feel uncomfortable if it suddenly stops for a moment whenever you enter a busy area. Those interruptions matter just as much as the usual speed.

One improvement comes from leaving finished work alone.

Imagine standing in town with your backpack open. The world is moving, but most of the backpack window may be exactly as it was a moment ago. Its border, labels, and item pictures do not all need to be drawn again just because someone walked past you.

Henge now keeps the unchanged parts of the interface and redraws the areas that need attention. If a number changes, a message arrives, or a window moves, it updates those areas. The view of the world continues to be drawn as needed.

That sounds like a small saving, but drawing the same unchanged interface over and over was taking a noticeable amount of time on our older test hardware. Remembering what is already there gives the game more time for everything happening around your character.

We use a similar idea for the people in the world.

A character is made from several parts, including their body, face, hair, and equipment. Building those shapes takes work. When something has not changed, the game can reuse what it has already made. Your character’s hair does not need to be built from scratch every time the game shows another picture.

Movement makes this more complicated because a walking or working character keeps changing position. The motion settings let us reduce how often those poses are updated. Small characters in the distance can update less frequently than someone standing beside you.

There is a visible trade-off: lower motion settings make animation less fluid. They also give a slower computer less work to do. The game spreads those updates out so a whole group of characters does not always need to be rebuilt at the same moment.

Distance gives us other opportunities to save work.

The ground near your feet needs enough detail to make its slopes and edges look right. Ground far away takes up much less space on your screen. The land-detail settings allow Henge to use simpler shapes for that distant ground.

The joins still matter. We do not want a performance setting to leave gaps between nearby ground and distant ground. The simpler pieces keep matching edges so they meet properly.

Memory is another part of the work. You can think of it as the computer’s working space: somewhere to keep the things it needs close at hand. If that space becomes too crowded, the computer may have to keep putting things away and fetching them again. On an old machine, that can cause long pauses.

Henge has controls for its memory allowance and how many nearby players it shows. It also keeps character models available while those players remain in view, so it can reuse them. We want busy areas to use that limited working space sensibly.

Some improvements turn up in unexpected places. In one Windows test, a window was just a few pixels larger than the game’s picture. That small difference made Windows stretch the picture every time it displayed it. Giving the picture an exact fit removed a surprisingly expensive extra job.

Changes like that are useful because they save work without asking us to remove something from the world.

We also check that improvements draw things correctly. Making a scene faster is not helpful if it leaves part of a window missing or draws text in the wrong place. Our tests can compare the pictures the game produces with saved examples, helping us catch unintended changes.

Tablets are part of this work too. Henge’s browser client already supports tapping, holding a finger down to open action menus, dragging, pinching to zoom, and entering text with the on-screen keyboard.

Having those actions available is a foundation. We still want to improve how they feel: making controls comfortable to touch, keeping text readable, and making sure opening the keyboard leaves the game usable. We also want to check performance on real devices, including how consistently the game responds during play.

The aim is to make Henge’s world comfortable to spend time in. Each improvement gives us a little more room for the things that make it feel alive: people gathering in town, characters going about their work, changing scenery, and places worth exploring.

There is more to do, especially in the heavier scenes. We’ll keep measuring, testing, and improving the parts that need it, while keeping the game’s hardware requirements modest.