Since there are several other sites that partially overlap with this one (Stack Overflow in particular), what are the criteria for keeping a question here vs. migrating it to Stack Overflow?
In particular, I have a lot of Java and SVG graphics questions. Sometimes a question might on the border of performance vs. API vs. alternative methods of drawing.
Full disclaimer: I'm not a part of the original Area51 proposal past committing to it.
There has been a lot of discussion (as there should be) on the exact scope of this stack. Here is how I am personally angling myself:
First off, the original description for CG.SE was
[A stack for] computer graphics researchers and programmers.
which
[i]ncludes discussion of image processing, computational geometry and topology, ray tracing technologies, graphics libraries, gpgpu, specific graphics cards questions, and performance related questions.
For completeness, here are the top 5 questions from the definition phase:
When assessing the on-topic-ness of a question, I've been asking myself a few simple questions:
To me, that is the rigid, undeniable scope of this stack.
That being said, there has been some discussion about questions relating to graphics computation, though not specifically talking about it.
The linked question asks about perlin noise. As pointed out by Tom Cody, perlin noise is predominately used in computer graphics (though, since it is a mathematical algorithm, it can be used in any field that requires multi-dimensional randomness).
These quasi-off-topic questions are tricky, because it comes down to determining whether or not such topics are prevalent in the graphics community. As well, it must be determined if these topics will go out of scope over time.
Such outliers should be handled on a per-case basis.
There were a number of discussions back on Area 51 of what the scope of our site should be, and what should be on-topic or off-topic. I'll quote part of an answer I made there as it seems relevant:
This is fundamentally a programming site, as opposed to a site about art or design. We want questions that you'd encounter in the course of coding a renderer, basically. (Either a 2D or 3D renderer. Either real-time or offline. Either running on a GPU or not. Creating either photorealistic images or non-photorealistic ones, etc.)
What wouldn't be on-topic, for instance, would be questions about how to use 3D modeling tools, or how to create a skybox in Photoshop, or how to select a color palette for lighting a scene. Added to that, I guess we probably would not want questions about how to use specific 3D engines, e.g. "how can I increase shadow quality in Unity/UE4/whatever". (Maybe I'm wrong about that, but I don't consider such questions very interesting, and I think they belong on an engine-specific community site.)
The appropriateness of engine-specific or tool-specific questions was probably the issue people disagreed about the most. I argued against such questions above, but some others argued that questions specifically about Unity, UE4, Maya, Houdini, etc. were fine. I think we'll have to give it a try and refine our criteria as we go. The main concern for me about such questions is that they have to do with programming with those tools, as opposed to just using the tools. For instance, I can see questions about how to build a custom exporter plugin for Maya, or how to implement rendering technique X in UE4, as potentially on-topic.
Also, I want to expand a bit to say that not just rendering, but other related areas such as image processing, particles/fluids/cloth simulation, geometry processing (e.g. CSG, voxelization, remeshing), procedural geometry/texture generation and such things like that are probably on topic here as well. Maybe parts of computer vision too. Again, we'll probably have to refine the "edges" as we go.
