Architecture Notes

Some notes to help me decide on the best format for functionality

draw vs step vs render

I choice to avoid a render method to emphasize a more adaptable, lower-level approach. This approach keeps it open for the end-user to integrate their own render method.

For all items with drawing features, we maintain a step for integral changes, and a draw for full pen work.

Notably at the root of a Point, the draw object provides many methods. This is coupled with the pen object to provide the two scoped integrations. This leads to a discontinuity in the naming convention, as the draw method is less complex items is ambiguous.

Therefore I'm proposing the following:

• step: for integral updates, such as the motor for gears.
• update: as a method for entities with full changes, and called when required.
• render: a method combining draw and pen on a context across many sub items - such as a Ribbon

Step

The step method should be purely logical, a avoid performing context drawing when required. A step can be called to update a point values, such as rotations and positions.

In the future we may move the step method off the primary loop (e.g. in a thread).

draw

A draw performs context updates for a frame (noting immediate mode drawing). This performs the context update actions (such as ctx.moveTo), but attempts to avoid colouring, strokes or other final visuals.

On a Point (and potentially other complex drawable items), the draw is an object with many methods. In primary cases, the draw is followed by a pen method call to render graphics.

render

Reserved for end user products a render method is more formal, utilising many aspects of the reference object (self) to perform many draw and pen actions, with conditions.

A Point, Stage, or generally most components supplied with the Polypoint library don't have a render method, allowing a developer to write their own where required.

Info