We are seeing some questions which receive answers that deal with magic. There isn't anything wrong in building worlds where magic exists, but there are two main issues that I can see:
- Magic systems differ from world to world. No two writers are likely to come up with exactly the same magic system, so anything proposed would need (possibly fairly extensive) tailoring to the world where it will be used. This isn't necessarily a big problem, but it is something worth keeping in mind. Nothing kills suspension of disbelief like a good repeated whacking over the head with inconsistency.
- More importantly, not everyone is building a world where magic exists. In fact, lots of worldbuilding involves worlds where magic doesn't exist, or where it exists but is not understood by the protagonist's race.
If I were to be world-building for a detective story set in some town on Earth, and get stuck on some particular point, I want to come here and be able to ask about it (within the limits of what we ultimately decide is on topic for the site in the first place) without having to explicitly state that I don't want answers involving magic wands, shapeshifters and highly specialized magic systems.
On the flip side, if I'm building a magic-based world, I want to be able to come here and ask questions and receive answers that are likely to be useful to me in that endeavour.
The question thus becomes: should we assume that answers invoking some form of magic are acceptable, by default?
I would argue that we should not; if an asker does not explicitly ask for answers that invoke some sort of magic, then I feel our default stance should be that any answers given should be accurate to within known sciences. Also consider the Area51 proposal blurb:
Q&A site for writers/artists using science, geography and culture to construct imaginary worlds and settings.
This goes in hand with the suggestion that answers should be supportable by facts and references. Answers based (in whole or part) around magic cannot be supported by references to known sciences, for which there appears to be a fairly wide consensus even though the specifics differ between various proposals for how and how strictly to implement it.