People are lazy and will wait for absolutely the last moment: "why should I spend time configuring my network when I still can purchase IPv4?" This was especial

People are lazy and will wait for absolutely the last moment: "why should I spend time configuring my network when I still can purchase IPv4?" This was especially bad because vendors and ISPs had the same attitude, so even when there was someone willing to do the work, those parties stayed in the way. I actually see great parallels about Python 2 vs 3. The changes between these versions aren't really big, and converting is not that hard. Other languages already did it many times (Ruby actually did it when switching from 1.8 to 1.9, but people moved on quickly, because if they didn't, their apps and libraries no longer would work and no one would use them). IMHO the real cause was that Python gave a lot of time to do it (nearly 15 years!), but there was no real interest in Py3 until 2015 when Python 2.7 went into maintenance mode (no new features backported, BTW all features of Python 2.7 were backports of Python 3, fueling "Python 3 doesn't have anything new, what should I switch?"), and organizations most likey will wait until 2020 (EOL) to port their applications.