LWN.net's Journal

Cannon: Where are we in the Python 3 transition?
Brett Cannon continues his series of posts on Python 3 with a blog post likening the path of its adoption to the Kübler-Ross model (i.e. the five stages of grief). "Unfortunately people are running up against the classic problem of lacking buy-in from management. I regularly hear from people that they would switch if they could, but their manager(s) don't see any reason to switch and so they can't (or that they would do per-file porting, but they don't think they can convince their teammates to maintain the porting work). This can be especially frustrating if you use Python 3 in personal projects but are stuck on Python 2 at work. Hopefully Python 3 will continue to offer new features that will eventually entice reluctant managers to switch. Otherwise financial arguments might be necessary in the form of pointing out that porting to Python 3 is a one-time cost while staying on Python 2 past 2020 will be a perpetual cost for support to some enterprise provider of Python and will cost more in the long-term (e.g., paying for RHEL so that someone supports your Python 2 install past 2020). Have hope, though, that you can get buy-in from management for porting to Python 3 since others have and thus reached the "acceptance" stage."