This makes me want to cry. I see myself going down the rabbit holes recently. I had this great idea that I was investigating. Then I saw one of the functionality can become a product by itself. Hey, let me do that quickly. While doing it, I needed to build a library. Hey, let me generalized it quickly and released it as open source.
Good question. I guess it would require at least doing some due diligence before dismissing an idea. For me the OS library is generic and I can always reuse for other projects. The other two ideas require much more extensive tryout.
And that's how JSODA came into existence, https://github.com/williamw520/jsoda
But I'm still longing at popping the stack two levels up back to the original idea.