When this site started we had a topic discussing a similar problem:
Are questions about neurobiology on or off topic? It seemed the 'general consensus' was: "They are on topic so long as they touch base with something related to cognition." I believe this is the general rule we should apply, whatever the topic.
I think everybody agrees that a line should be drawn. It's a similar problem to one which Stack Overflow had to address: "As a software developer, what is your favorite vegetable?" Does this mean it's still related to software development?
Granted that's an extreme example, but in the same way we need to be able to draw a line to prevent just any question from being asked which is only tangentially related. E.g. "I'm building this box for an experiment, how should I use the hammer?"
Since 'touch base' is kind of subjective, I propose the following more structured aid in order to determine whether it touches base enough.
- Are there any major problems with the question besides it only being partially on topic?
- Does the question asked also apply to different fields of expertise?
- Is there another SE site available where the question could be answered?
- If "Yes", would it be useful for the OP to create an account there to ask subsequent questions?
In case all answers are "No", the question should be more than welcome here.
In case answer 1 is "Yes", the questions should obviously be closed.
In case answer 2 and 3 is "Yes", the question should be migrated to the respective site. This causes the least overlap with other sites, and thus also the least duplication.
Any edge cases can still be discussed on a case to case basis.