For me, as long as the price is roughly what the annual upgrade cost is, and there are compelling new features, it's wash financially, and provides predictable revenue for the company to fund development.
I view OS compatibility as a compelling new feature (when apple changes the OS, it takes time and money for the dev team to update the app to work properly). Security fixes are something different. I do security for a living, so understand that they are critical, but since they can be viewed as a defect in the original product, it's not exactly a new feature. That's why I actually prefer a subscription model as it creates a stronger incentive to get it right up front (lower maintenance costs). Periodic licensing can actually create a perverse incentive to *not* get it right, since the security fixes can trigger upgrade spend.
In any case, I appreciate that they've moved to a .5 model again this year, but fully expect that next year will be a paid upgrade again. I have no expectation that a version from several years ago will continue to work. One of the companies that I work with (not VMware!) had a huge outcry from people who'd last bought software 6 or 7 years ago. They were outraged that an OS update was going to break their ancient versions, and were excoriating the company on social media. The company privately told me that they didn't view those people as customers since they weren't buying (and wouldn't buy) software anyway. That's a fair point, and that realization made it easier for them to transition to a subscription model - they weren't leaving customers behind, even if they were leaving users behind.
It hit home recently when I suffered through the 'game-pocalypse' on Catalina. A ton of my old mac games stopped working (and won't work in a VM because of the 3d graphics issue) - so all that spend is lost. I'd be happy to pay for an upgrade, but the publisher isn't going to do it because most people won't, so there's no business case to justify the development.
