Lifetime of the product. Given such constraints I would seriously consider just starting different projects and let that app die a slow, painful death by having

Lifetime of the product. Given such constraints I would seriously consider just starting different projects and let that app die a slow, painful death by having the ecosystem around it move forward. Yes, the app will be functional on Android 8 in 2039 - if you can still find a functional Android 8 device then (after the epoch roll-over). Basic maintenance helps the app move forward to Android 30 (or whatever is current then), letting you enjoy precisely the features you've bought back in 2017, so to me that seems to be strictly better. (I exaggerated version numbers and dates to take in the epoch roll-over issue as an outside-anybody's-responsibility issue with an ancient base OS: https://www.quora.com/What-will-happen-to-a-64-bit-Android-device-in-2038-because-of-the-Y2K38-problem https://www.quora.com/What-will-happen-to-a-64-bit-Android-d..., but the same scenario also applies when talking about supporting whatever base OS of 2022 and the ways it will or will not be compatible to today's apps)