Comments on: Invalidation Strategies for Flash Player

By: JesterXL

JesterXL — Sat, 10 Feb 2007 00:23:51 +0000

Programming by contract goes farther than ‘using components within yourself’ as Composition. It means there ARE no base classes. Everything is an interface, and you implement those. Even your children can be brought in via interfaces via Factories, so as long as a child implements a specific interface, say IUIComponent, he can be created within you via Composition.

Composition is merely the pattern, in Flash / Flex terms, of using components within components (Using Button + Label in a UIComponent to make CheckBox). There is no enforcement however unless you strongly type addChild (like it is via DisplayObject). Instead of using DisplayObject, Chris would argue to use IDisplayObject for example (which you can’t btw, but it would be cool if you could).

By: Steven Sacks

Steven Sacks — Fri, 09 Feb 2007 20:52:49 +0000

What Chris was talking to you about is a design pattern known as Composition. I know you hate design patterns (wink) but I think you should pick up Head First and give it a go.

By: JesterXL

JesterXL — Thu, 01 Feb 2007 19:09:23 +0000

No, because they have no clue how the children are doing the invalidating, they just know that the children have a method called ‘invalidate’ that takes no parameters and doesn’t return anything. That leaves the details of redraw up to the invidual components.

By: Bob

Bob — Thu, 01 Feb 2007 19:03:24 +0000

If parents are responsible for invalidating their children, does that break the notion of ‘black box’ encapsulation?

By: JesterXL

JesterXL — Wed, 31 Jan 2007 23:40:25 +0000

Yes, shorter variable names use less RAM, and are faster to execute (faster bytecode). Briefly mentioned that with other optimizations in a previous entry.