Yeah the happy medium I have found is to just see if the authors have submitted the package to anaconda and use that instead. Brew when that is no longer availa

Yeah the happy medium I have found is to just see if the authors have submitted the package to anaconda and use that instead. Brew when that is no longer available. I will usually have one conda env per package or at the most per project. It isn't just that things "work together" but in some cases authors change how their tools work between versions and I may have stuff installed that depends on these outdated paradigms. In other words, the stuff I have built is not sure to work together with updated brew packages. I have a few scripts for example where numpy complains how I call some function will no longer merely throw a warning but actually not work at all in future versions. I can either spend my free time to rewrite the script to do it how the numpy authors now want me to do it, or I can just not do that and live on the old numpy and not be any worse for wear.