apply-bridge-pattern

Apply Bridge Pattern

Decouple an abstraction from its implementation so that both can vary independently.

Why This Is Best Practice

Adopted by: Java AWT rendering pipeline (Shape abstraction + platform renderer implementation — ships in every JVM), JDBC (Connection abstraction + vendor-specific driver implementation — every Java database program uses this), Android's View system (View abstraction + Canvas implementation), and .NET's Graphics class hierarchy. Impact: GoF shows that without Bridge, combining M abstractions × N implementations requires M×N classes. With Bridge, it requires M+N classes. A real example: a graphics library with 3 shapes (Circle, Square, Triangle) × 3 renderers (SVG, Canvas, PDF) needs 9 classes without Bridge, 6 with it — and adding a 4th renderer requires 3 new classes vs. 1. Why best: The alternative — inheritance — permanently binds abstraction to implementation. SVGCircle extends Circle cannot use a PDF renderer without a new subclass. Bridge separates the dimensions, making each independently extensible.

Sources: Gamma et al. (1994) pp. 151–161; Java AWT documentation; JDBC specification

Steps

Step 1: Identify the two independent variation dimensions

Abstraction dimension: Shape — Circle, Square, Triangle
Implementation dimension: Renderer — SVGRenderer, CanvasRenderer, PDFRenderer

Any combination should work without a dedicated class for each.

Step 2: Define the implementation interface

from abc import ABC, abstractmethod

class Renderer(ABC):
    @abstractmethod
    def render_circle(self, x: float, y: float, radius: float) -> None: ...

    @abstractmethod
    def render_rect(self, x: float, y: float, w: float, h: float) -> None: ...

Step 3: Define the abstraction holding a reference to the implementation

class Shape(ABC):
    def __init__(self, renderer: Renderer):
        self._renderer = renderer   # the bridge

    @abstractmethod
    def draw(self) -> None: ...

Step 4: Extend both hierarchies independently

class Circle(Shape):
    def __init__(self, renderer: Renderer, x: float, y: float, radius: float):
        super().__init__(renderer)
        self._x, self._y, self._radius = x, y, radius

    def draw(self) -> None:
        self._renderer.render_circle(self._x, self._y, self._radius)

class SVGRenderer(Renderer):
    def render_circle(self, x, y, radius):
        print(f'<circle cx="{x}" cy="{y}" r="{radius}"/>')

    def render_rect(self, x, y, w, h):
        print(f'<rect x="{x}" y="{y}" width="{w}" height="{h}"/>')

Step 5: Compose at runtime — any abstraction with any implementation

svg = SVGRenderer()
pdf = PDFRenderer()

Circle(svg, 10, 10, 5).draw()   # SVG circle
Circle(pdf, 10, 10, 5).draw()   # PDF circle — same Circle class, different renderer

Adding a new renderer requires one new Renderer subclass. Adding a new shape requires one new Shape subclass. Neither side affects the other.

When NOT to Use

• When only one dimension varies — if the implementation never changes, composition or a simple strategy suffices. Bridge adds abstraction overhead without payoff.
• When both hierarchies are shallow (1–2 subclasses each) — the simplification only pays off at 3+ variants per dimension.

Common Mistakes

Confusing Bridge with Adapter. Adapter makes incompatible interfaces work together after the fact. Bridge separates abstraction from implementation from the start, by design. If you're retrofitting, it's Adapter.

Putting abstraction logic in the implementation. The implementation interface should only expose primitives (render_circle, render_rect). If it contains domain logic, the bridge leaks.

Over-engineering for a single renderer. If there's only one implementation and no plan to add more, Bridge is premature. Add it when the second implementation dimension appears.