It has a certain 80/20 feel but I do think the amount of time we need to invest in finding why people hang on to python2 is worth the investment, but the returns to investment will be diminishing constantly. For me it has been mostly finding the py2 scripts before I find the py3 uplifts, and I have been too lazy to remove the local refs to the py2 so just re-hack the print () in, and install the pip dependencies. I think I have hit two languageisms which made me wince, in that time<p>At some level, you just accept. I have a favourite game (zork) which is a fortran translation of the original DDL, uplifted into BSD 3 or more decades ago, and since converted to C, but the build instructions do suggest keeping the fortran around and there are fragments of DDL wisdom in the codebase. I kind of miss not playing the fortran code, but even that made me (at one remove) an interloper given that was a translation from older code.