How do we approach not-quite-duplicates?

Question

There is a situation that occasionally arises in GDSE that is captured perfectly with two current questions:

• Detecting a ledge in Box2D 2d box2d 2d-physics platformer
• Detect Open Space in Farseer xna farseer-physics-engine

Here's the situation:

• Both are asking the same basic thing: "I have a ledge. I need to detect when my player has reached that ledge, so my player can hang onto it / pull themselves up."
• Both questions have the same basic answer: have a sensor upon the player to detect a collision with a wall, and have one sensor above to detect if there's empty space to ensure it's a ledge. This solution is engine-agnostic.
• For whatever reason, the more recent one isn't closed as an exact duplicate - maybe because they cover different engines.

In this case, we should only have one question attached to no particular physics engine:

• How do I detect ledges? physics

How should we make it so?

There's no precedent for this as far as I'm aware.

Do I edit the earlier Farseer question to be engine-agnostic? That would seem like it would be hijacking the question, but it would still be helpful to the original asker.

Do I create a new self-answered question and request to have both closed as a duplicate of the more recent one?

Answer 1

My gut feeling is that I'd probably edit one of the questions to divorce it from the irrelevant ties to specific engines or APIs and close the other one as a duplicate, possibly with a comment indicating that the subject is actually technology-agnostic to help curtail any backlash from the asker (who may not know any better).

However, I can see an argument for leaving both and creating a new question to close both as a duplicate of. if the original questions have good titles and keywords for search engines, it may be more useful to the health of the site to keep them untouched and just redirect via a duplicate-close to a technology-independent answer. This would be on the basis that neophytes probably won't realize how their problem is not specific to a piece of technology and will type in to search engines things like "box2d detect edge of platform" or whatnot.