OpenFl and Starling

Hi,
I see that a lot of people are using OpenFl with Starling, would like to know the reasons behind that :slight_smile:
Is that just because of as3 experience, legacy projects, plugins or there are some features/performance improvements compared to pure OpenFl?

Thanks for feedback

1 Like

As I have used Starling I can share my experience , first of all starling was awesome tool when it came out for as3 , so the api is very similar to flash api so that’s a big point for me , its easy and make more sense and I am used to it.

But most important part is even though openfl is hardware accelerated like starling , starling provide more features that openfl dose not have. Like spriteSheet support , assets management and it got many plugins like partial,custom sharers etc.

OpenFL supports hardware and software rendering. This is a big bonus in OpenFL’s favor – canvas and DOM support, for example, are not supported for Stage3D. However, since Starling is hardware only, there are some features it has which may not be available in OpenFL’s standard renderer (yet). One example is how OpenFL added Tilemap, which works using copyPixels on software renderers, and the equivalent of a QuadBatch or Mesh style renderer when using hardware.

OpenFL includes asset management, but Starling’s asset management system works differently, and perhaps is more familiar or better suited to certain use-cases.

We support both use-cases, because it is good to have choices, especially when porting code.

Thank you for input :slight_smile:
As I have starling + featherui project written in haxe but using as3 swc lib files, trying to understand what would be best investment in future :slight_smile: as there are 3 ways:

  1. port featherui to haxe and use same setup in native targets (openfl + starling + featherui)
  2. drop featherui but still use starling but for ui to use some have ui lib (openfl + starling + haxeui for example)
  3. drop both starling and featherui and go with (openfl + haxeui)

I am more for 3. solution that mean least amount of layers for rendering and dependencies. :slight_smile: but at the same time would not like to lose perf/features and reinvent them from scratch :smile: Maybe starling 2.0 asset manager could be refactored out of starling and could be used separately :thinking:
@Galib good point about spritesheet as I rely on them.