Design of an immediate graphical user interface api with managed layouts

Terms:

1. Canvas - a drawing area

User stories

1. I would like to call UI-rendering functions without specifying dimensions and locations

2. I would like to call custom UI-rendering functions while specifying dimensions and locations

/----------------------------------
Story 2 needs a way to get a proper starting location or a bounding box If 1 is in place.

Consider the single method Layout interface:

type Layout interface {
	// Advances the layout and returns a starting point
	// for an element based on the last ending point.
	Next(last Point) Point
}

Story 1 necessitates that library-aware UI-rendering functions acquire their starting locations
on their own. This allows for the following two approaches (more are possible of course)

* Store the ending point of the last element. New calls can get it and pass it to the layout
  + does not advance the layout unless required, saving cpu
  + does not imply a dependency on the layout, the UI drawing needs just a starting point

* Store the next starting point by calling Next() at the end of the element drawing routine.
The next drawing can continue from it.
  - advancing the layout costs cpu
  - the user might want an out of layer render
  - puts a dependency on the layout

The first strategy is definatelly more sound and neatly allows for a layout-aware tracker
to handle the locations.
\----------------------------------

3. I would like to enhance drawing functions with additional behavior (e.g. layout trackers, styling, etc)