The point is that there are apps that can pretty much be built entirely using web technologies as PWAs, but in doing so they are no longer "just web sites" and they need functionality of installed apps, like notifications. For example, most banking apps on Android could be entirely rewritten as PWAs, but they'd need to make use of things like notifications (I don't want a random website sending me notifications, but I DO want notifications of activity on my bank account) and camera APIs (e.g. for mobile check deposit).
It basically is just a difference in use case and how most people thing about "apps" vs. "websites". If I have a long term relationship with a business, and I access its functionality frequently, I'd rather have it as an app on my homescreen.
There are benefits to having something installed. You can give anything installed the ability to send notifications.
Having to manage that permission on a per-website basis is way above the tech ability of most users.
But this is apparently difficult for "tech literate" people to grasp.
The last thing being the worst as SMS 2FA is so insecure, but SMS of your bank deposits isn't much better.
Really? You don't understand the value in having the PWA appear as a native app icon alongside everything else on the device?
It's a bit confusing, isn't it? "Add to home screen" makes it seem like you're just adding a link, but it's installing the PWA, possibly enabling notifications, and etc.
This isn't true for all apps, but with the notable exception of games, I'd say it applies to most: banking/finance apps, social media apps, travel/airline/booking apps, etc.
So two of the three parts that are native are there to make it a worse user experience?
Analytics tell developers exactly where bugs and crashes occur.
And on which devices or versions of the OS the problem is.
Without analytics it would take weeks/months to figure out exactly what line of code is causing the issue. Heck, the developers might NEVER KNOW that the software has an issue.
Apps would just keep crashing on users for years. And developers would have no idea why users were abandoning their product.
Nobody that has any idea what they're talking about would say that analytics makes for a worse user experience.
Maps require complex gestures and advanced graphics and UIs that would never work well as a PWA. Try maps.google.com. Its nothing like the Google Maps app. Plus Android Auto integration.
Tile probably needs pretty deep Bluetooth integration and background processing that the web doesn't provide.
YouTube can do things like PiP that you can't do on the web.
edit: Sorry for the snark. I think that maps are probably doable with pointer events (https://caniuse.com/?search=pointer) and Android is doing pretty well in terms of Web Bluetooth https://github.com/WebBluetoothCG/web-bluetooth/blob/main/im...
https://developer.mozilla.org/en-US/docs/Web/API/Picture-in-...
If I request the Desktop version of YouTube, where Google doesn't hide their own Mini player button, the video continues to play in the corner.
Mapbox.js is also pretty capable.
Not really sure what you're on about TBH.
Right, because for a very long time (and maybe still) PWAs have been much closer to terrible websites than good apps. They generally don't have the same UX properties as real apps.
Remember when the iPhone first came out and web apps were the only option and they sucked balls? That never really changed. People still find mobile sites with maps that are impossible to use. They don't expect that from app store apps.
And if it's just a web site, why do you need to "install" it? A link is surely sufficient?