> Upgrade pricing also doesn’t solve this. [...] versioned upgrades incentivize developers to chase shiny features that people might pay for rather than improving their app and building for the long haul. It makes the product worse.

In my mind this is missing the point. I am very pro-upgrade pricing because upgrade pricing actually forces the developer to think about what their users actually will pay for, or what improvements will make my workflow better, instead of just adding whatever fancy side-quest thing that you've decided I need today.

If users are asking for shiny new features and are willing to pay for them then fine. If we want refinements and are willing to pay for them then fine. If you want to optionally provide backend storage or backups or whatever then fine.

However, if my continued use of a specific version of your software has no continued cost to you, then be absolutely assured that I do not feel a moral obligation to keep on paying for it endlessly, nor should you assume that I do.

> However, if my continued use of a specific version of your software has no continued cost to you

That's just it, it does carry a cost, at the least for security updates and allowing the application to continue to be installed on later OSe or key libraries.

That's why even open-source libraries that in actively being used will get forked if unmaintained too long, because there's always something to do to keep bitrot at bay.

Can you list any examples of companies that have done this well? The products I have used that tried this either stopped trying it (Ulysses articulated it really well) or devolved into shinier objects. In practice I think it's pretty tough to pull off.