I'm not conflating them, but clearly software can be both researchware and open source. I'm not going to start distrusting tensorflow or scikit-learn, but I'm going to be wary before I invest time in a promising niche software package, regardless of whether it was actually from academia or not.