Question

I’m refactoring some code in project whilst reading Object-Oriented
Reengineering Patterns specifically the section on “Transforming Conditionals to Polymorphism”. So the current code base has constant that is referencing a factory class that which returns a bitmap based on displayWidth & displayHeight. To achieve this do I need to create two new classes, with each representing the differing screenWidth and screenHeight ? I’m a litte lost as to whats the best method of achieving polymorphism in this case.

public static final Bitmap TICKER_BACKGROUND_IMAGE = ImageFactory.getFooterBitmap();

Method in ImageFactory –

public class ImageFactory {

private static int displayWidth;
private static int displayHeight;

static {
    displayWidth = Display.getWidth();
    displayHeight = Display.getHeight();
}

public static Bitmap getFooterBitmap(){ 

    if(displayWidth == 360 && displayHeight == 480){
        return Bitmap.getBitmapResource("360x480/footer_bg.png");
    }
    else {
        return Bitmap.getBitmapResource("320x240/footer_bg.png");
    }

}

}

Answer 1

I would take all it arguments as parameters. Don’t use static variables as argument if you can.

public static Bitmap getFooterBitmap(int width, int height){ 
    String filename = width == 360 && height == 480 ? "360x480" : "320x240";
    return Bitmap.getBitmapResource(filename+"/footer_bp.png);
}

Using polymorphism is a good idea but not the best solution in every situation.

---

Another approach might be to see if the size is available and use a fall back position.

public static Bitmap getFooterBitmap(int width, int height){ 
    Bitmap bm = Bitmap.getBitmapResource(width+"x"+height+"/footer_bp.png);
    if (bm == null)
       bm = Bitmap.getBitmapResource("320x240/footer_bp.png);
    return bm;
}