following upstream is overrated since we have good package managers and version control.
it's completely feasible to stick to something that works for you, and only update/port/rewrite when it makes sense.
what matters is the overall cost.
following upstream is overrated since we have good package managers and version control.
it's completely feasible to stick to something that works for you, and only update/port/rewrite when it makes sense.
what matters is the overall cost.
Hmm, if one writes a library Zetalib for the language Frob v0.14 and then Frob v0.15 introduces breaking changes that everyone else is going to adapt to, then well, package managers and version control is going to help indeed - they will help in staying in a void as no one will use Zetalib anymore because of the older Frob.
For libs, yes, for applications dev, no.
I would expect fixing an application to an older version would be just fine, so long as you don't need newer language features. If newer language features are a requirement, I would expect that would drive refactoring or selecting a different implementation language entirely if refactoring would prove to be too onerous.