I was re-reading your comment, and I think this is an intriguing & provocative statement. But is it true? When you say “most people” do this, it sounds plausible, but I wonder what you base it upon.
If it is true, it seems to point to something problematic. I love Minecraft, but I’m not sure that sitting down with Java & LWJGL is necessarily a great way to approach building a game. Silly example, but hopefully, it illustrates the point.
If we acknowledge that there is a lot of room for different approaches to the problem, perhaps a decision tree to lead potential authors to the best tool for them is warranted. For example, a prospective author with little affinity for programming will likely have a rough time with Rez, no matter how much they liked a game built upon it.
On that note, I recently implemented behaviour trees as an alternative to scripting behaviour. I wonder if a writer with a bit of logical thinking could build behaviour with a more technically experienced partner creating condition & query tasks that they need (these are implemented in Javascript).
Anyway, thanks again for your thoughts.