It got there by doing short cycles
2) it doesn't mean they are still benefitial to the users. At some point an app does "all" and new features are more likely to lead to bloat than improvements for 99% of the users. Thus "they got there due to that" isn't an argument.
Increased patching to address this reality seems unavoidable, at least until we understand how to produce truly secure software on the first try far better than we currently do.
Today with SaaS/etc offerings you have very different difficulty spikes. Imagine only rolling out an update once a year to 1B customers. Especially with all the big changes that might entail.
I can't imagine the fear. Releasing often irons out small problems before they become massive problems. Or, at least, is a strategy to attempt that.
And honestly, I don't get the idea of making an app for every service where a simple web app suffices. Why making a native app that requires installation when you can do it through the browser?
Depends on how many libraries there are. I have to deal with at least 2-3 different CVEs existing in our libraries or runtimes a month. And I’m dealing with a pretty small surface area compared to some of my peers.
We rebuild our environment from scratch every week to reduce the manual work involved in resolving those CVEs.
There is no question that the current app reality is one or more orders of magnitude more complex than in "the old days"; but even so, I think the race to MVP (and perhaps the VC pressures that drive some of that?) have led to layers of rushed out, risky codebases.
We probably get the unnecessary updates because Google (and others) have a boat-load of engineers that need something to do and they aren't getting assigned to something useful.
Now we suffer from update fatigue, where half the time we fire up an app or turn on a device, it must spend a few minutes updating apps.
I'm not convinced that we always get value from these updates. (Security updates are indeed important, but sometimes their need is evidence of pushing out features too soon, without proper testing and review.)
Google Calendar seems to do everything I need and more; same with Gmail. I don't want new versions of them unless there's a really compelling reason. I sure don't want < 2 month release cycles.