Hi, I've recently completed a set of FlexJS chart classes that "optimize" the chart graphics. You can read about it on the FlexJS Wiki [1] (at the end of the page). What this means is that rather than creating a nesting of objects to represent, say, a bar on a chart, I use the FlexJS core.graphics.GraphicsContainer's drawRect() function.
In JavaScript, this results in a single SVG <rect> element. In ActionScript, this results in a call to the Flash drawingAPI drawRect() function. This works well because there is an ActionScript and a JavaScript version of the core.graphics package. Most of the FlexJS chart code is in the FlexJSJX project which is cross-compiled into JavaScript so I don't have to write both versions. While this seemed ideal at first, I've run into a situation where I can get events from the JavaScript side, but not from the ActionScript side, on a per-element basis. For example, in JavaScript I can tell when the mouse enters the <rect> element, but in ActionScript, there is no object, just a drawing. What I propose to do is add to the core.graphics.GraphicsContainer object a set of functions that create elements in both JavaScript and ActionScript. Where we now have: graphicsContainer.drawRect(0, 0, 100, 200) (makes JavaScript): <rect x="0" y="0" width="100" height="200"> (makes ActionScript): graphics.drawRect(0,0,100,200) I would add: graphicsContainer.createRect(0, 0, 100, 200) which would do the same thing in JavaScript as .drawRect(), but the ActionScript version would yield a Shape. In this way we have parity between the JavaScript and ActionScript versions with a simple object tree for graphics that also can generate events. The current ActionScript "optimized" package would become the "NonInteractive" package and later we could have that match to HTML5 <canvas> and its drawing API and keep the parity going. Let me know what you think. Regards, Peter Ent Adobe Systems [1] https://cwiki.apache.org/confluence/display/FLEX/FlexJS+Charts
