I think the interesting thing to note here is that for most end users a standard protocol is not nearly as appealing as a standard user interface. Like how most iOS apps follow common navigation paradigms, so users don't need to spend time learning each new one. And how people tend to prefer buying things on Amazon instead of any random webshop. And how WeChat has almost become a little web of its own with custom chat UIs for e.g. ordering pizza. If open software for consumers is ever to stop being beaten by these walled gardens, it needs to get a ton more UX people involved and focus on the end user UX, not just the technical implementation.
I'm not saying the App Store is perfect, not by a country mile is it. I'm just saying if you compare it to the Play store, and the wider Internet, holy moly do walled gardens look nice.
That is very true too. I often wondered what would have happened had someone created say 10 years ago an email client with strict protocol compatibility but employing an higher abstraction layer where discussions/attaches are managed automatically just like in a Whatsapp chat (chat -> mail list) using a similar UI. Probably most users would have rejected the idea because they had no intention to learn a new thing nobody was using. So, back to square one.
There is more to it. Using apps instead of websites we're killing the whole concept of making information and services available using a standard protocol. That's like going back over 30 years, even before gopher. Then there is the cloud->mainframe similarity, with data and services being moved somewhere into the net where the user has no control over his/her own data, and the whole email/usenet->social media migration which puts private and public communications in the hands of corporations having no intention of guaranteeing interoperability with competing systems.
Basically the industry screwed 40+ years of IT development, not just in a technical way, but they did it behind shiny graphics and animated emoticons, so nobody really cares.