Kind of, it's just that the approach Google takes is a lot more palatable than Apple's. As someone who has written a PWA (albeit one that almost entirely relies on SSR), Google's PWA approach is definitely better than Apple, but there's some marked issues.
For one, the actual PWA packaging process gets shunted off to a Google server; I think you can make a "thin client" APK from a manifest using a tool they wrote some time ago[0] (Twitter Lite is one of these), but I've not really looked into it. It's not quite the extension to Chrome you'd really want it to be; if you use a non-Chrome browser on Android, it means you can't really ditch the Chrome dependency if you want to use a PWA. (Further not really helped by the fact that Google is basically the only PWA implementer on Android, since Firefox does not consider PWAs a priority whatsoever.) Similarly, Google's servers need to be able to read out the manifest declaration, which makes them unfeasible for intranet software unless you want to punch a temporary hole and expose it to the internet for a bit.
The other kinda annoying thing Google does is really aggressive degradation between PWA and homescreen shortcut. If the manifest isn't entirely up to snuff in terms of what's listed, there's no attempt at trying to resolve the issue, it just instantly degrades to a homescreen shortcut. A basic example of this is the requirement to use a service worker (even if the service workers entire job is to do nothing); it's not really stated in the manifest spec that it's required, but if you don't have one, the PWA straight up refuses to install as a PWA.
Google's strength with the play store really mostly comes from their bundling advantage; Play Services and the attached Store and Google Apps are required for OEMs to add to their devices (might change with the DMA?). That's the kinda odd reality that makes Apples desire for control seem so extreme - we know what an open platform looks like on Android. It works pretty well for the most part and the incumbents advantage for a store is large enough that almost every app developer submits to the Play Store regardless.
Tend to agree with all of this. Manifest is way too finicky.
Would be interesting to see how the play store changes in the event of Android honoring code signing for side loading like windows.
Eg..no scare screen on side loaded apk’s as long as they’re code signed.
I suspect the App Store would live on as a consumer focused App Store and the enterprise apps would direct distribute which makes sense anyhow cause IAP does t understand account based pricing.
Android doesn't do scare screens actually. The only real difference between installing an APK from the play store and an APK you found on the internet is that the application calling the installer has a one time "OK" to make sure you are the one who wants to install an APK; Play Store has this as well, but the default distribution has it turned on, by going in the settings you can fiddle with it and turn it off if you want to.
The only thing actually needed for feature parity with the Play Store is mostly just that F-Droid can't auto-update; the Play Store can skip the update/install prompt screen, F-Droid can't. They added install origins to APK files last year iirc, so there's a likely chance they're allowing it though.
Yeah, still have what equates to a scare screen.
Tells you file may be harmful upon download, then you need to change a setting which is streamlined to what it was before, but still a scare screen.
Now you can allow from source…but the source isn’t the web address, it’s the initiating application eg. Chrome or Files so there’s a huge security hole with this implementation presumably on purpose to manufacture the incident they need to justify their behavior.
I’ll have to test this tomorrow. Last time I tried sideloading direct from our website I had to flip a switch in settings which came with a scare dialog.
If I remember correctly, it was a system wide setting too and didn’t allow for trusting specific sources.
If we can self distribute on Android, that will be 3 out of 4.
For one, the actual PWA packaging process gets shunted off to a Google server; I think you can make a "thin client" APK from a manifest using a tool they wrote some time ago[0] (Twitter Lite is one of these), but I've not really looked into it. It's not quite the extension to Chrome you'd really want it to be; if you use a non-Chrome browser on Android, it means you can't really ditch the Chrome dependency if you want to use a PWA. (Further not really helped by the fact that Google is basically the only PWA implementer on Android, since Firefox does not consider PWAs a priority whatsoever.) Similarly, Google's servers need to be able to read out the manifest declaration, which makes them unfeasible for intranet software unless you want to punch a temporary hole and expose it to the internet for a bit.
The other kinda annoying thing Google does is really aggressive degradation between PWA and homescreen shortcut. If the manifest isn't entirely up to snuff in terms of what's listed, there's no attempt at trying to resolve the issue, it just instantly degrades to a homescreen shortcut. A basic example of this is the requirement to use a service worker (even if the service workers entire job is to do nothing); it's not really stated in the manifest spec that it's required, but if you don't have one, the PWA straight up refuses to install as a PWA.
Google's strength with the play store really mostly comes from their bundling advantage; Play Services and the attached Store and Google Apps are required for OEMs to add to their devices (might change with the DMA?). That's the kinda odd reality that makes Apples desire for control seem so extreme - we know what an open platform looks like on Android. It works pretty well for the most part and the incumbents advantage for a store is large enough that almost every app developer submits to the Play Store regardless.
[0]: It's called Bubblewrap - https://github.com/GoogleChromeLabs/bubblewrap